key_map.json

{
  "encoder.layers.0.conv.batch_norm.num_batches_tracked": "encoder.layers.0.conv.batch_norm.num_batches_tracked",
  "encoder.layers.1.conv.batch_norm.num_batches_tracked": "encoder.layers.1.conv.batch_norm.num_batches_tracked",
  "encoder.layers.10.conv.batch_norm.num_batches_tracked": "encoder.layers.10.conv.batch_norm.num_batches_tracked",
  "encoder.layers.11.conv.batch_norm.num_batches_tracked": "encoder.layers.11.conv.batch_norm.num_batches_tracked",
  "encoder.layers.12.conv.batch_norm.num_batches_tracked": "encoder.layers.12.conv.batch_norm.num_batches_tracked",
  "encoder.layers.13.conv.batch_norm.num_batches_tracked": "encoder.layers.13.conv.batch_norm.num_batches_tracked",
  "encoder.layers.14.conv.batch_norm.num_batches_tracked": "encoder.layers.14.conv.batch_norm.num_batches_tracked",
  "encoder.layers.15.conv.batch_norm.num_batches_tracked": "encoder.layers.15.conv.batch_norm.num_batches_tracked",
  "encoder.layers.16.conv.batch_norm.num_batches_tracked": "encoder.layers.16.conv.batch_norm.num_batches_tracked",
  "encoder.layers.17.conv.batch_norm.num_batches_tracked": "encoder.layers.17.conv.batch_norm.num_batches_tracked",
  "encoder.layers.18.conv.batch_norm.num_batches_tracked": "encoder.layers.18.conv.batch_norm.num_batches_tracked",
  "encoder.layers.19.conv.batch_norm.num_batches_tracked": "encoder.layers.19.conv.batch_norm.num_batches_tracked",
  "encoder.layers.2.conv.batch_norm.num_batches_tracked": "encoder.layers.2.conv.batch_norm.num_batches_tracked",
  "encoder.layers.20.conv.batch_norm.num_batches_tracked": "encoder.layers.20.conv.batch_norm.num_batches_tracked",
  "encoder.layers.21.conv.batch_norm.num_batches_tracked": "encoder.layers.21.conv.batch_norm.num_batches_tracked",
  "encoder.layers.22.conv.batch_norm.num_batches_tracked": "encoder.layers.22.conv.batch_norm.num_batches_tracked",
  "encoder.layers.23.conv.batch_norm.num_batches_tracked": "encoder.layers.23.conv.batch_norm.num_batches_tracked",
  "encoder.layers.24.conv.batch_norm.num_batches_tracked": "encoder.layers.24.conv.batch_norm.num_batches_tracked",
  "encoder.layers.25.conv.batch_norm.num_batches_tracked": "encoder.layers.25.conv.batch_norm.num_batches_tracked",
  "encoder.layers.26.conv.batch_norm.num_batches_tracked": "encoder.layers.26.conv.batch_norm.num_batches_tracked",
  "encoder.layers.27.conv.batch_norm.num_batches_tracked": "encoder.layers.27.conv.batch_norm.num_batches_tracked",
  "encoder.layers.28.conv.batch_norm.num_batches_tracked": "encoder.layers.28.conv.batch_norm.num_batches_tracked",
  "encoder.layers.29.conv.batch_norm.num_batches_tracked": "encoder.layers.29.conv.batch_norm.num_batches_tracked",
  "encoder.layers.3.conv.batch_norm.num_batches_tracked": "encoder.layers.3.conv.batch_norm.num_batches_tracked",
  "encoder.layers.30.conv.batch_norm.num_batches_tracked": "encoder.layers.30.conv.batch_norm.num_batches_tracked",
  "encoder.layers.31.conv.batch_norm.num_batches_tracked": "encoder.layers.31.conv.batch_norm.num_batches_tracked",
  "encoder.layers.32.conv.batch_norm.num_batches_tracked": "encoder.layers.32.conv.batch_norm.num_batches_tracked",
  "encoder.layers.33.conv.batch_norm.num_batches_tracked": "encoder.layers.33.conv.batch_norm.num_batches_tracked",
  "encoder.layers.34.conv.batch_norm.num_batches_tracked": "encoder.layers.34.conv.batch_norm.num_batches_tracked",
  "encoder.layers.35.conv.batch_norm.num_batches_tracked": "encoder.layers.35.conv.batch_norm.num_batches_tracked",
  "encoder.layers.36.conv.batch_norm.num_batches_tracked": "encoder.layers.36.conv.batch_norm.num_batches_tracked",
  "encoder.layers.37.conv.batch_norm.num_batches_tracked": "encoder.layers.37.conv.batch_norm.num_batches_tracked",
  "encoder.layers.38.conv.batch_norm.num_batches_tracked": "encoder.layers.38.conv.batch_norm.num_batches_tracked",
  "encoder.layers.39.conv.batch_norm.num_batches_tracked": "encoder.layers.39.conv.batch_norm.num_batches_tracked",
  "encoder.layers.4.conv.batch_norm.num_batches_tracked": "encoder.layers.4.conv.batch_norm.num_batches_tracked",
  "encoder.layers.40.conv.batch_norm.num_batches_tracked": "encoder.layers.40.conv.batch_norm.num_batches_tracked",
  "encoder.layers.41.conv.batch_norm.num_batches_tracked": "encoder.layers.41.conv.batch_norm.num_batches_tracked",
  "encoder.layers.42.conv.batch_norm.num_batches_tracked": "encoder.layers.42.conv.batch_norm.num_batches_tracked",
  "encoder.layers.43.conv.batch_norm.num_batches_tracked": "encoder.layers.43.conv.batch_norm.num_batches_tracked",
  "encoder.layers.44.conv.batch_norm.num_batches_tracked": "encoder.layers.44.conv.batch_norm.num_batches_tracked",
  "encoder.layers.45.conv.batch_norm.num_batches_tracked": "encoder.layers.45.conv.batch_norm.num_batches_tracked",
  "encoder.layers.46.conv.batch_norm.num_batches_tracked": "encoder.layers.46.conv.batch_norm.num_batches_tracked",
  "encoder.layers.47.conv.batch_norm.num_batches_tracked": "encoder.layers.47.conv.batch_norm.num_batches_tracked",
  "encoder.layers.5.conv.batch_norm.num_batches_tracked": "encoder.layers.5.conv.batch_norm.num_batches_tracked",
  "encoder.layers.6.conv.batch_norm.num_batches_tracked": "encoder.layers.6.conv.batch_norm.num_batches_tracked",
  "encoder.layers.7.conv.batch_norm.num_batches_tracked": "encoder.layers.7.conv.batch_norm.num_batches_tracked",
  "encoder.layers.8.conv.batch_norm.num_batches_tracked": "encoder.layers.8.conv.batch_norm.num_batches_tracked",
  "encoder.layers.9.conv.batch_norm.num_batches_tracked": "encoder.layers.9.conv.batch_norm.num_batches_tracked",
  "encoder.layers.0.conv.batch_norm.bias": "encoder.layers.0.conv.batch_norm.bias",
  "encoder.layers.0.conv.batch_norm.running_mean": "encoder.layers.0.conv.batch_norm.running_mean",
  "encoder.layers.0.conv.batch_norm.running_var": "encoder.layers.0.conv.batch_norm.running_var",
  "encoder.layers.0.conv.batch_norm.weight": "encoder.layers.0.conv.batch_norm.weight",
  "encoder.layers.0.conv.depthwise_conv.bias": "encoder.layers.0.conv.depthwise_conv.bias",
  "encoder.layers.0.conv.depthwise_conv.weight": "encoder.layers.0.conv.depthwise_conv.weight",
  "encoder.layers.0.conv.pointwise_conv1.bias": "encoder.layers.0.conv.pointwise_conv1.bias",
  "encoder.layers.0.conv.pointwise_conv1.weight": "encoder.layers.0.conv.pointwise_conv1.weight",
  "encoder.layers.0.conv.pointwise_conv2.bias": "encoder.layers.0.conv.pointwise_conv2.bias",
  "encoder.layers.0.conv.pointwise_conv2.weight": "encoder.layers.0.conv.pointwise_conv2.weight",
  "encoder.layers.0.feed_forward1.linear1.bias": "encoder.layers.0.feed_forward1.linear1.bias",
  "encoder.layers.0.feed_forward1.linear1.weight": "encoder.layers.0.feed_forward1.linear1.weight",
  "encoder.layers.0.feed_forward1.linear2.bias": "encoder.layers.0.feed_forward1.linear2.bias",
  "encoder.layers.0.feed_forward1.linear2.weight": "encoder.layers.0.feed_forward1.linear2.weight",
  "encoder.layers.0.feed_forward2.linear1.bias": "encoder.layers.0.feed_forward2.linear1.bias",
  "encoder.layers.0.feed_forward2.linear1.weight": "encoder.layers.0.feed_forward2.linear1.weight",
  "encoder.layers.0.feed_forward2.linear2.bias": "encoder.layers.0.feed_forward2.linear2.bias",
  "encoder.layers.0.feed_forward2.linear2.weight": "encoder.layers.0.feed_forward2.linear2.weight",
  "encoder.layers.0.norm_conv.bias": "encoder.layers.0.norm_conv.bias",
  "encoder.layers.0.norm_conv.weight": "encoder.layers.0.norm_conv.weight",
  "encoder.layers.0.norm_feed_forward1.bias": "encoder.layers.0.norm_feed_forward1.bias",
  "encoder.layers.0.norm_feed_forward1.weight": "encoder.layers.0.norm_feed_forward1.weight",
  "encoder.layers.0.norm_feed_forward2.bias": "encoder.layers.0.norm_feed_forward2.bias",
  "encoder.layers.0.norm_feed_forward2.weight": "encoder.layers.0.norm_feed_forward2.weight",
  "encoder.layers.0.norm_out.bias": "encoder.layers.0.norm_out.bias",
  "encoder.layers.0.norm_out.weight": "encoder.layers.0.norm_out.weight",
  "encoder.layers.0.norm_self_att.bias": "encoder.layers.0.norm_self_att.bias",
  "encoder.layers.0.norm_self_att.weight": "encoder.layers.0.norm_self_att.weight",
  "encoder.layers.0.self_attn.linear_out.bias": "encoder.layers.0.self_attn.out_proj.bias",
  "encoder.layers.0.self_attn.linear_out.weight": "encoder.layers.0.self_attn.out_proj.weight",
  "encoder.layers.0.self_attn.linear_pos.weight": "encoder.layers.0.self_attn.pos_proj.weight",
  "encoder.layers.0.self_attn.<qkv>.bias": "encoder.layers.0.self_attn.qkv_proj.bias",
  "encoder.layers.0.self_attn.<qkv>.weight": "encoder.layers.0.self_attn.qkv_proj.weight",
  "encoder.layers.0.self_attn.pos_bias_u": "encoder.layers.0.self_attn.pos_bias_u",
  "encoder.layers.0.self_attn.pos_bias_v": "encoder.layers.0.self_attn.pos_bias_v",
  "encoder.layers.1.conv.batch_norm.bias": "encoder.layers.1.conv.batch_norm.bias",
  "encoder.layers.1.conv.batch_norm.running_mean": "encoder.layers.1.conv.batch_norm.running_mean",
  "encoder.layers.1.conv.batch_norm.running_var": "encoder.layers.1.conv.batch_norm.running_var",
  "encoder.layers.1.conv.batch_norm.weight": "encoder.layers.1.conv.batch_norm.weight",
  "encoder.layers.1.conv.depthwise_conv.bias": "encoder.layers.1.conv.depthwise_conv.bias",
  "encoder.layers.1.conv.depthwise_conv.weight": "encoder.layers.1.conv.depthwise_conv.weight",
  "encoder.layers.1.conv.pointwise_conv1.bias": "encoder.layers.1.conv.pointwise_conv1.bias",
  "encoder.layers.1.conv.pointwise_conv1.weight": "encoder.layers.1.conv.pointwise_conv1.weight",
  "encoder.layers.1.conv.pointwise_conv2.bias": "encoder.layers.1.conv.pointwise_conv2.bias",
  "encoder.layers.1.conv.pointwise_conv2.weight": "encoder.layers.1.conv.pointwise_conv2.weight",
  "encoder.layers.1.feed_forward1.linear1.bias": "encoder.layers.1.feed_forward1.linear1.bias",
  "encoder.layers.1.feed_forward1.linear1.weight": "encoder.layers.1.feed_forward1.linear1.weight",
  "encoder.layers.1.feed_forward1.linear2.bias": "encoder.layers.1.feed_forward1.linear2.bias",
  "encoder.layers.1.feed_forward1.linear2.weight": "encoder.layers.1.feed_forward1.linear2.weight",
  "encoder.layers.1.feed_forward2.linear1.bias": "encoder.layers.1.feed_forward2.linear1.bias",
  "encoder.layers.1.feed_forward2.linear1.weight": "encoder.layers.1.feed_forward2.linear1.weight",
  "encoder.layers.1.feed_forward2.linear2.bias": "encoder.layers.1.feed_forward2.linear2.bias",
  "encoder.layers.1.feed_forward2.linear2.weight": "encoder.layers.1.feed_forward2.linear2.weight",
  "encoder.layers.1.norm_conv.bias": "encoder.layers.1.norm_conv.bias",
  "encoder.layers.1.norm_conv.weight": "encoder.layers.1.norm_conv.weight",
  "encoder.layers.1.norm_feed_forward1.bias": "encoder.layers.1.norm_feed_forward1.bias",
  "encoder.layers.1.norm_feed_forward1.weight": "encoder.layers.1.norm_feed_forward1.weight",
  "encoder.layers.1.norm_feed_forward2.bias": "encoder.layers.1.norm_feed_forward2.bias",
  "encoder.layers.1.norm_feed_forward2.weight": "encoder.layers.1.norm_feed_forward2.weight",
  "encoder.layers.1.norm_out.bias": "encoder.layers.1.norm_out.bias",
  "encoder.layers.1.norm_out.weight": "encoder.layers.1.norm_out.weight",
  "encoder.layers.1.norm_self_att.bias": "encoder.layers.1.norm_self_att.bias",
  "encoder.layers.1.norm_self_att.weight": "encoder.layers.1.norm_self_att.weight",
  "encoder.layers.1.self_attn.linear_out.bias": "encoder.layers.1.self_attn.out_proj.bias",
  "encoder.layers.1.self_attn.linear_out.weight": "encoder.layers.1.self_attn.out_proj.weight",
  "encoder.layers.1.self_attn.linear_pos.weight": "encoder.layers.1.self_attn.pos_proj.weight",
  "encoder.layers.1.self_attn.<qkv>.bias": "encoder.layers.1.self_attn.qkv_proj.bias",
  "encoder.layers.1.self_attn.<qkv>.weight": "encoder.layers.1.self_attn.qkv_proj.weight",
  "encoder.layers.1.self_attn.pos_bias_u": "encoder.layers.1.self_attn.pos_bias_u",
  "encoder.layers.1.self_attn.pos_bias_v": "encoder.layers.1.self_attn.pos_bias_v",
  "encoder.layers.10.conv.batch_norm.bias": "encoder.layers.10.conv.batch_norm.bias",
  "encoder.layers.10.conv.batch_norm.running_mean": "encoder.layers.10.conv.batch_norm.running_mean",
  "encoder.layers.10.conv.batch_norm.running_var": "encoder.layers.10.conv.batch_norm.running_var",
  "encoder.layers.10.conv.batch_norm.weight": "encoder.layers.10.conv.batch_norm.weight",
  "encoder.layers.10.conv.depthwise_conv.bias": "encoder.layers.10.conv.depthwise_conv.bias",
  "encoder.layers.10.conv.depthwise_conv.weight": "encoder.layers.10.conv.depthwise_conv.weight",
  "encoder.layers.10.conv.pointwise_conv1.bias": "encoder.layers.10.conv.pointwise_conv1.bias",
  "encoder.layers.10.conv.pointwise_conv1.weight": "encoder.layers.10.conv.pointwise_conv1.weight",
  "encoder.layers.10.conv.pointwise_conv2.bias": "encoder.layers.10.conv.pointwise_conv2.bias",
  "encoder.layers.10.conv.pointwise_conv2.weight": "encoder.layers.10.conv.pointwise_conv2.weight",
  "encoder.layers.10.feed_forward1.linear1.bias": "encoder.layers.10.feed_forward1.linear1.bias",
  "encoder.layers.10.feed_forward1.linear1.weight": "encoder.layers.10.feed_forward1.linear1.weight",
  "encoder.layers.10.feed_forward1.linear2.bias": "encoder.layers.10.feed_forward1.linear2.bias",
  "encoder.layers.10.feed_forward1.linear2.weight": "encoder.layers.10.feed_forward1.linear2.weight",
  "encoder.layers.10.feed_forward2.linear1.bias": "encoder.layers.10.feed_forward2.linear1.bias",
  "encoder.layers.10.feed_forward2.linear1.weight": "encoder.layers.10.feed_forward2.linear1.weight",
  "encoder.layers.10.feed_forward2.linear2.bias": "encoder.layers.10.feed_forward2.linear2.bias",
  "encoder.layers.10.feed_forward2.linear2.weight": "encoder.layers.10.feed_forward2.linear2.weight",
  "encoder.layers.10.norm_conv.bias": "encoder.layers.10.norm_conv.bias",
  "encoder.layers.10.norm_conv.weight": "encoder.layers.10.norm_conv.weight",
  "encoder.layers.10.norm_feed_forward1.bias": "encoder.layers.10.norm_feed_forward1.bias",
  "encoder.layers.10.norm_feed_forward1.weight": "encoder.layers.10.norm_feed_forward1.weight",
  "encoder.layers.10.norm_feed_forward2.bias": "encoder.layers.10.norm_feed_forward2.bias",
  "encoder.layers.10.norm_feed_forward2.weight": "encoder.layers.10.norm_feed_forward2.weight",
  "encoder.layers.10.norm_out.bias": "encoder.layers.10.norm_out.bias",
  "encoder.layers.10.norm_out.weight": "encoder.layers.10.norm_out.weight",
  "encoder.layers.10.norm_self_att.bias": "encoder.layers.10.norm_self_att.bias",
  "encoder.layers.10.norm_self_att.weight": "encoder.layers.10.norm_self_att.weight",
  "encoder.layers.10.self_attn.linear_out.bias": "encoder.layers.10.self_attn.out_proj.bias",
  "encoder.layers.10.self_attn.linear_out.weight": "encoder.layers.10.self_attn.out_proj.weight",
  "encoder.layers.10.self_attn.linear_pos.weight": "encoder.layers.10.self_attn.pos_proj.weight",
  "encoder.layers.10.self_attn.<qkv>.bias": "encoder.layers.10.self_attn.qkv_proj.bias",
  "encoder.layers.10.self_attn.<qkv>.weight": "encoder.layers.10.self_attn.qkv_proj.weight",
  "encoder.layers.10.self_attn.pos_bias_u": "encoder.layers.10.self_attn.pos_bias_u",
  "encoder.layers.10.self_attn.pos_bias_v": "encoder.layers.10.self_attn.pos_bias_v",
  "encoder.layers.11.conv.batch_norm.bias": "encoder.layers.11.conv.batch_norm.bias",
  "encoder.layers.11.conv.batch_norm.running_mean": "encoder.layers.11.conv.batch_norm.running_mean",
  "encoder.layers.11.conv.batch_norm.running_var": "encoder.layers.11.conv.batch_norm.running_var",
  "encoder.layers.11.conv.batch_norm.weight": "encoder.layers.11.conv.batch_norm.weight",
  "encoder.layers.11.conv.depthwise_conv.bias": "encoder.layers.11.conv.depthwise_conv.bias",
  "encoder.layers.11.conv.depthwise_conv.weight": "encoder.layers.11.conv.depthwise_conv.weight",
  "encoder.layers.11.conv.pointwise_conv1.bias": "encoder.layers.11.conv.pointwise_conv1.bias",
  "encoder.layers.11.conv.pointwise_conv1.weight": "encoder.layers.11.conv.pointwise_conv1.weight",
  "encoder.layers.11.conv.pointwise_conv2.bias": "encoder.layers.11.conv.pointwise_conv2.bias",
  "encoder.layers.11.conv.pointwise_conv2.weight": "encoder.layers.11.conv.pointwise_conv2.weight",
  "encoder.layers.11.feed_forward1.linear1.bias": "encoder.layers.11.feed_forward1.linear1.bias",
  "encoder.layers.11.feed_forward1.linear1.weight": "encoder.layers.11.feed_forward1.linear1.weight",
  "encoder.layers.11.feed_forward1.linear2.bias": "encoder.layers.11.feed_forward1.linear2.bias",
  "encoder.layers.11.feed_forward1.linear2.weight": "encoder.layers.11.feed_forward1.linear2.weight",
  "encoder.layers.11.feed_forward2.linear1.bias": "encoder.layers.11.feed_forward2.linear1.bias",
  "encoder.layers.11.feed_forward2.linear1.weight": "encoder.layers.11.feed_forward2.linear1.weight",
  "encoder.layers.11.feed_forward2.linear2.bias": "encoder.layers.11.feed_forward2.linear2.bias",
  "encoder.layers.11.feed_forward2.linear2.weight": "encoder.layers.11.feed_forward2.linear2.weight",
  "encoder.layers.11.norm_conv.bias": "encoder.layers.11.norm_conv.bias",
  "encoder.layers.11.norm_conv.weight": "encoder.layers.11.norm_conv.weight",
  "encoder.layers.11.norm_feed_forward1.bias": "encoder.layers.11.norm_feed_forward1.bias",
  "encoder.layers.11.norm_feed_forward1.weight": "encoder.layers.11.norm_feed_forward1.weight",
  "encoder.layers.11.norm_feed_forward2.bias": "encoder.layers.11.norm_feed_forward2.bias",
  "encoder.layers.11.norm_feed_forward2.weight": "encoder.layers.11.norm_feed_forward2.weight",
  "encoder.layers.11.norm_out.bias": "encoder.layers.11.norm_out.bias",
  "encoder.layers.11.norm_out.weight": "encoder.layers.11.norm_out.weight",
  "encoder.layers.11.norm_self_att.bias": "encoder.layers.11.norm_self_att.bias",
  "encoder.layers.11.norm_self_att.weight": "encoder.layers.11.norm_self_att.weight",
  "encoder.layers.11.self_attn.linear_out.bias": "encoder.layers.11.self_attn.out_proj.bias",
  "encoder.layers.11.self_attn.linear_out.weight": "encoder.layers.11.self_attn.out_proj.weight",
  "encoder.layers.11.self_attn.linear_pos.weight": "encoder.layers.11.self_attn.pos_proj.weight",
  "encoder.layers.11.self_attn.<qkv>.bias": "encoder.layers.11.self_attn.qkv_proj.bias",
  "encoder.layers.11.self_attn.<qkv>.weight": "encoder.layers.11.self_attn.qkv_proj.weight",
  "encoder.layers.11.self_attn.pos_bias_u": "encoder.layers.11.self_attn.pos_bias_u",
  "encoder.layers.11.self_attn.pos_bias_v": "encoder.layers.11.self_attn.pos_bias_v",
  "encoder.layers.12.conv.batch_norm.bias": "encoder.layers.12.conv.batch_norm.bias",
  "encoder.layers.12.conv.batch_norm.running_mean": "encoder.layers.12.conv.batch_norm.running_mean",
  "encoder.layers.12.conv.batch_norm.running_var": "encoder.layers.12.conv.batch_norm.running_var",
  "encoder.layers.12.conv.batch_norm.weight": "encoder.layers.12.conv.batch_norm.weight",
  "encoder.layers.12.conv.depthwise_conv.bias": "encoder.layers.12.conv.depthwise_conv.bias",
  "encoder.layers.12.conv.depthwise_conv.weight": "encoder.layers.12.conv.depthwise_conv.weight",
  "encoder.layers.12.conv.pointwise_conv1.bias": "encoder.layers.12.conv.pointwise_conv1.bias",
  "encoder.layers.12.conv.pointwise_conv1.weight": "encoder.layers.12.conv.pointwise_conv1.weight",
  "encoder.layers.12.conv.pointwise_conv2.bias": "encoder.layers.12.conv.pointwise_conv2.bias",
  "encoder.layers.12.conv.pointwise_conv2.weight": "encoder.layers.12.conv.pointwise_conv2.weight",
  "encoder.layers.12.feed_forward1.linear1.bias": "encoder.layers.12.feed_forward1.linear1.bias",
  "encoder.layers.12.feed_forward1.linear1.weight": "encoder.layers.12.feed_forward1.linear1.weight",
  "encoder.layers.12.feed_forward1.linear2.bias": "encoder.layers.12.feed_forward1.linear2.bias",
  "encoder.layers.12.feed_forward1.linear2.weight": "encoder.layers.12.feed_forward1.linear2.weight",
  "encoder.layers.12.feed_forward2.linear1.bias": "encoder.layers.12.feed_forward2.linear1.bias",
  "encoder.layers.12.feed_forward2.linear1.weight": "encoder.layers.12.feed_forward2.linear1.weight",
  "encoder.layers.12.feed_forward2.linear2.bias": "encoder.layers.12.feed_forward2.linear2.bias",
  "encoder.layers.12.feed_forward2.linear2.weight": "encoder.layers.12.feed_forward2.linear2.weight",
  "encoder.layers.12.norm_conv.bias": "encoder.layers.12.norm_conv.bias",
  "encoder.layers.12.norm_conv.weight": "encoder.layers.12.norm_conv.weight",
  "encoder.layers.12.norm_feed_forward1.bias": "encoder.layers.12.norm_feed_forward1.bias",
  "encoder.layers.12.norm_feed_forward1.weight": "encoder.layers.12.norm_feed_forward1.weight",
  "encoder.layers.12.norm_feed_forward2.bias": "encoder.layers.12.norm_feed_forward2.bias",
  "encoder.layers.12.norm_feed_forward2.weight": "encoder.layers.12.norm_feed_forward2.weight",
  "encoder.layers.12.norm_out.bias": "encoder.layers.12.norm_out.bias",
  "encoder.layers.12.norm_out.weight": "encoder.layers.12.norm_out.weight",
  "encoder.layers.12.norm_self_att.bias": "encoder.layers.12.norm_self_att.bias",
  "encoder.layers.12.norm_self_att.weight": "encoder.layers.12.norm_self_att.weight",
  "encoder.layers.12.self_attn.linear_out.bias": "encoder.layers.12.self_attn.out_proj.bias",
  "encoder.layers.12.self_attn.linear_out.weight": "encoder.layers.12.self_attn.out_proj.weight",
  "encoder.layers.12.self_attn.linear_pos.weight": "encoder.layers.12.self_attn.pos_proj.weight",
  "encoder.layers.12.self_attn.<qkv>.bias": "encoder.layers.12.self_attn.qkv_proj.bias",
  "encoder.layers.12.self_attn.<qkv>.weight": "encoder.layers.12.self_attn.qkv_proj.weight",
  "encoder.layers.12.self_attn.pos_bias_u": "encoder.layers.12.self_attn.pos_bias_u",
  "encoder.layers.12.self_attn.pos_bias_v": "encoder.layers.12.self_attn.pos_bias_v",
  "encoder.layers.13.conv.batch_norm.bias": "encoder.layers.13.conv.batch_norm.bias",
  "encoder.layers.13.conv.batch_norm.running_mean": "encoder.layers.13.conv.batch_norm.running_mean",
  "encoder.layers.13.conv.batch_norm.running_var": "encoder.layers.13.conv.batch_norm.running_var",
  "encoder.layers.13.conv.batch_norm.weight": "encoder.layers.13.conv.batch_norm.weight",
  "encoder.layers.13.conv.depthwise_conv.bias": "encoder.layers.13.conv.depthwise_conv.bias",
  "encoder.layers.13.conv.depthwise_conv.weight": "encoder.layers.13.conv.depthwise_conv.weight",
  "encoder.layers.13.conv.pointwise_conv1.bias": "encoder.layers.13.conv.pointwise_conv1.bias",
  "encoder.layers.13.conv.pointwise_conv1.weight": "encoder.layers.13.conv.pointwise_conv1.weight",
  "encoder.layers.13.conv.pointwise_conv2.bias": "encoder.layers.13.conv.pointwise_conv2.bias",
  "encoder.layers.13.conv.pointwise_conv2.weight": "encoder.layers.13.conv.pointwise_conv2.weight",
  "encoder.layers.13.feed_forward1.linear1.bias": "encoder.layers.13.feed_forward1.linear1.bias",
  "encoder.layers.13.feed_forward1.linear1.weight": "encoder.layers.13.feed_forward1.linear1.weight",
  "encoder.layers.13.feed_forward1.linear2.bias": "encoder.layers.13.feed_forward1.linear2.bias",
  "encoder.layers.13.feed_forward1.linear2.weight": "encoder.layers.13.feed_forward1.linear2.weight",
  "encoder.layers.13.feed_forward2.linear1.bias": "encoder.layers.13.feed_forward2.linear1.bias",
  "encoder.layers.13.feed_forward2.linear1.weight": "encoder.layers.13.feed_forward2.linear1.weight",
  "encoder.layers.13.feed_forward2.linear2.bias": "encoder.layers.13.feed_forward2.linear2.bias",
  "encoder.layers.13.feed_forward2.linear2.weight": "encoder.layers.13.feed_forward2.linear2.weight",
  "encoder.layers.13.norm_conv.bias": "encoder.layers.13.norm_conv.bias",
  "encoder.layers.13.norm_conv.weight": "encoder.layers.13.norm_conv.weight",
  "encoder.layers.13.norm_feed_forward1.bias": "encoder.layers.13.norm_feed_forward1.bias",
  "encoder.layers.13.norm_feed_forward1.weight": "encoder.layers.13.norm_feed_forward1.weight",
  "encoder.layers.13.norm_feed_forward2.bias": "encoder.layers.13.norm_feed_forward2.bias",
  "encoder.layers.13.norm_feed_forward2.weight": "encoder.layers.13.norm_feed_forward2.weight",
  "encoder.layers.13.norm_out.bias": "encoder.layers.13.norm_out.bias",
  "encoder.layers.13.norm_out.weight": "encoder.layers.13.norm_out.weight",
  "encoder.layers.13.norm_self_att.bias": "encoder.layers.13.norm_self_att.bias",
  "encoder.layers.13.norm_self_att.weight": "encoder.layers.13.norm_self_att.weight",
  "encoder.layers.13.self_attn.linear_out.bias": "encoder.layers.13.self_attn.out_proj.bias",
  "encoder.layers.13.self_attn.linear_out.weight": "encoder.layers.13.self_attn.out_proj.weight",
  "encoder.layers.13.self_attn.linear_pos.weight": "encoder.layers.13.self_attn.pos_proj.weight",
  "encoder.layers.13.self_attn.<qkv>.bias": "encoder.layers.13.self_attn.qkv_proj.bias",
  "encoder.layers.13.self_attn.<qkv>.weight": "encoder.layers.13.self_attn.qkv_proj.weight",
  "encoder.layers.13.self_attn.pos_bias_u": "encoder.layers.13.self_attn.pos_bias_u",
  "encoder.layers.13.self_attn.pos_bias_v": "encoder.layers.13.self_attn.pos_bias_v",
  "encoder.layers.14.conv.batch_norm.bias": "encoder.layers.14.conv.batch_norm.bias",
  "encoder.layers.14.conv.batch_norm.running_mean": "encoder.layers.14.conv.batch_norm.running_mean",
  "encoder.layers.14.conv.batch_norm.running_var": "encoder.layers.14.conv.batch_norm.running_var",
  "encoder.layers.14.conv.batch_norm.weight": "encoder.layers.14.conv.batch_norm.weight",
  "encoder.layers.14.conv.depthwise_conv.bias": "encoder.layers.14.conv.depthwise_conv.bias",
  "encoder.layers.14.conv.depthwise_conv.weight": "encoder.layers.14.conv.depthwise_conv.weight",
  "encoder.layers.14.conv.pointwise_conv1.bias": "encoder.layers.14.conv.pointwise_conv1.bias",
  "encoder.layers.14.conv.pointwise_conv1.weight": "encoder.layers.14.conv.pointwise_conv1.weight",
  "encoder.layers.14.conv.pointwise_conv2.bias": "encoder.layers.14.conv.pointwise_conv2.bias",
  "encoder.layers.14.conv.pointwise_conv2.weight": "encoder.layers.14.conv.pointwise_conv2.weight",
  "encoder.layers.14.feed_forward1.linear1.bias": "encoder.layers.14.feed_forward1.linear1.bias",
  "encoder.layers.14.feed_forward1.linear1.weight": "encoder.layers.14.feed_forward1.linear1.weight",
  "encoder.layers.14.feed_forward1.linear2.bias": "encoder.layers.14.feed_forward1.linear2.bias",
  "encoder.layers.14.feed_forward1.linear2.weight": "encoder.layers.14.feed_forward1.linear2.weight",
  "encoder.layers.14.feed_forward2.linear1.bias": "encoder.layers.14.feed_forward2.linear1.bias",
  "encoder.layers.14.feed_forward2.linear1.weight": "encoder.layers.14.feed_forward2.linear1.weight",
  "encoder.layers.14.feed_forward2.linear2.bias": "encoder.layers.14.feed_forward2.linear2.bias",
  "encoder.layers.14.feed_forward2.linear2.weight": "encoder.layers.14.feed_forward2.linear2.weight",
  "encoder.layers.14.norm_conv.bias": "encoder.layers.14.norm_conv.bias",
  "encoder.layers.14.norm_conv.weight": "encoder.layers.14.norm_conv.weight",
  "encoder.layers.14.norm_feed_forward1.bias": "encoder.layers.14.norm_feed_forward1.bias",
  "encoder.layers.14.norm_feed_forward1.weight": "encoder.layers.14.norm_feed_forward1.weight",
  "encoder.layers.14.norm_feed_forward2.bias": "encoder.layers.14.norm_feed_forward2.bias",
  "encoder.layers.14.norm_feed_forward2.weight": "encoder.layers.14.norm_feed_forward2.weight",
  "encoder.layers.14.norm_out.bias": "encoder.layers.14.norm_out.bias",
  "encoder.layers.14.norm_out.weight": "encoder.layers.14.norm_out.weight",
  "encoder.layers.14.norm_self_att.bias": "encoder.layers.14.norm_self_att.bias",
  "encoder.layers.14.norm_self_att.weight": "encoder.layers.14.norm_self_att.weight",
  "encoder.layers.14.self_attn.linear_out.bias": "encoder.layers.14.self_attn.out_proj.bias",
  "encoder.layers.14.self_attn.linear_out.weight": "encoder.layers.14.self_attn.out_proj.weight",
  "encoder.layers.14.self_attn.linear_pos.weight": "encoder.layers.14.self_attn.pos_proj.weight",
  "encoder.layers.14.self_attn.<qkv>.bias": "encoder.layers.14.self_attn.qkv_proj.bias",
  "encoder.layers.14.self_attn.<qkv>.weight": "encoder.layers.14.self_attn.qkv_proj.weight",
  "encoder.layers.14.self_attn.pos_bias_u": "encoder.layers.14.self_attn.pos_bias_u",
  "encoder.layers.14.self_attn.pos_bias_v": "encoder.layers.14.self_attn.pos_bias_v",
  "encoder.layers.15.conv.batch_norm.bias": "encoder.layers.15.conv.batch_norm.bias",
  "encoder.layers.15.conv.batch_norm.running_mean": "encoder.layers.15.conv.batch_norm.running_mean",
  "encoder.layers.15.conv.batch_norm.running_var": "encoder.layers.15.conv.batch_norm.running_var",
  "encoder.layers.15.conv.batch_norm.weight": "encoder.layers.15.conv.batch_norm.weight",
  "encoder.layers.15.conv.depthwise_conv.bias": "encoder.layers.15.conv.depthwise_conv.bias",
  "encoder.layers.15.conv.depthwise_conv.weight": "encoder.layers.15.conv.depthwise_conv.weight",
  "encoder.layers.15.conv.pointwise_conv1.bias": "encoder.layers.15.conv.pointwise_conv1.bias",
  "encoder.layers.15.conv.pointwise_conv1.weight": "encoder.layers.15.conv.pointwise_conv1.weight",
  "encoder.layers.15.conv.pointwise_conv2.bias": "encoder.layers.15.conv.pointwise_conv2.bias",
  "encoder.layers.15.conv.pointwise_conv2.weight": "encoder.layers.15.conv.pointwise_conv2.weight",
  "encoder.layers.15.feed_forward1.linear1.bias": "encoder.layers.15.feed_forward1.linear1.bias",
  "encoder.layers.15.feed_forward1.linear1.weight": "encoder.layers.15.feed_forward1.linear1.weight",
  "encoder.layers.15.feed_forward1.linear2.bias": "encoder.layers.15.feed_forward1.linear2.bias",
  "encoder.layers.15.feed_forward1.linear2.weight": "encoder.layers.15.feed_forward1.linear2.weight",
  "encoder.layers.15.feed_forward2.linear1.bias": "encoder.layers.15.feed_forward2.linear1.bias",
  "encoder.layers.15.feed_forward2.linear1.weight": "encoder.layers.15.feed_forward2.linear1.weight",
  "encoder.layers.15.feed_forward2.linear2.bias": "encoder.layers.15.feed_forward2.linear2.bias",
  "encoder.layers.15.feed_forward2.linear2.weight": "encoder.layers.15.feed_forward2.linear2.weight",
  "encoder.layers.15.norm_conv.bias": "encoder.layers.15.norm_conv.bias",
  "encoder.layers.15.norm_conv.weight": "encoder.layers.15.norm_conv.weight",
  "encoder.layers.15.norm_feed_forward1.bias": "encoder.layers.15.norm_feed_forward1.bias",
  "encoder.layers.15.norm_feed_forward1.weight": "encoder.layers.15.norm_feed_forward1.weight",
  "encoder.layers.15.norm_feed_forward2.bias": "encoder.layers.15.norm_feed_forward2.bias",
  "encoder.layers.15.norm_feed_forward2.weight": "encoder.layers.15.norm_feed_forward2.weight",
  "encoder.layers.15.norm_out.bias": "encoder.layers.15.norm_out.bias",
  "encoder.layers.15.norm_out.weight": "encoder.layers.15.norm_out.weight",
  "encoder.layers.15.norm_self_att.bias": "encoder.layers.15.norm_self_att.bias",
  "encoder.layers.15.norm_self_att.weight": "encoder.layers.15.norm_self_att.weight",
  "encoder.layers.15.self_attn.linear_out.bias": "encoder.layers.15.self_attn.out_proj.bias",
  "encoder.layers.15.self_attn.linear_out.weight": "encoder.layers.15.self_attn.out_proj.weight",
  "encoder.layers.15.self_attn.linear_pos.weight": "encoder.layers.15.self_attn.pos_proj.weight",
  "encoder.layers.15.self_attn.<qkv>.bias": "encoder.layers.15.self_attn.qkv_proj.bias",
  "encoder.layers.15.self_attn.<qkv>.weight": "encoder.layers.15.self_attn.qkv_proj.weight",
  "encoder.layers.15.self_attn.pos_bias_u": "encoder.layers.15.self_attn.pos_bias_u",
  "encoder.layers.15.self_attn.pos_bias_v": "encoder.layers.15.self_attn.pos_bias_v",
  "encoder.layers.16.conv.batch_norm.bias": "encoder.layers.16.conv.batch_norm.bias",
  "encoder.layers.16.conv.batch_norm.running_mean": "encoder.layers.16.conv.batch_norm.running_mean",
  "encoder.layers.16.conv.batch_norm.running_var": "encoder.layers.16.conv.batch_norm.running_var",
  "encoder.layers.16.conv.batch_norm.weight": "encoder.layers.16.conv.batch_norm.weight",
  "encoder.layers.16.conv.depthwise_conv.bias": "encoder.layers.16.conv.depthwise_conv.bias",
  "encoder.layers.16.conv.depthwise_conv.weight": "encoder.layers.16.conv.depthwise_conv.weight",
  "encoder.layers.16.conv.pointwise_conv1.bias": "encoder.layers.16.conv.pointwise_conv1.bias",
  "encoder.layers.16.conv.pointwise_conv1.weight": "encoder.layers.16.conv.pointwise_conv1.weight",
  "encoder.layers.16.conv.pointwise_conv2.bias": "encoder.layers.16.conv.pointwise_conv2.bias",
  "encoder.layers.16.conv.pointwise_conv2.weight": "encoder.layers.16.conv.pointwise_conv2.weight",
  "encoder.layers.16.feed_forward1.linear1.bias": "encoder.layers.16.feed_forward1.linear1.bias",
  "encoder.layers.16.feed_forward1.linear1.weight": "encoder.layers.16.feed_forward1.linear1.weight",
  "encoder.layers.16.feed_forward1.linear2.bias": "encoder.layers.16.feed_forward1.linear2.bias",
  "encoder.layers.16.feed_forward1.linear2.weight": "encoder.layers.16.feed_forward1.linear2.weight",
  "encoder.layers.16.feed_forward2.linear1.bias": "encoder.layers.16.feed_forward2.linear1.bias",
  "encoder.layers.16.feed_forward2.linear1.weight": "encoder.layers.16.feed_forward2.linear1.weight",
  "encoder.layers.16.feed_forward2.linear2.bias": "encoder.layers.16.feed_forward2.linear2.bias",
  "encoder.layers.16.feed_forward2.linear2.weight": "encoder.layers.16.feed_forward2.linear2.weight",
  "encoder.layers.16.norm_conv.bias": "encoder.layers.16.norm_conv.bias",
  "encoder.layers.16.norm_conv.weight": "encoder.layers.16.norm_conv.weight",
  "encoder.layers.16.norm_feed_forward1.bias": "encoder.layers.16.norm_feed_forward1.bias",
  "encoder.layers.16.norm_feed_forward1.weight": "encoder.layers.16.norm_feed_forward1.weight",
  "encoder.layers.16.norm_feed_forward2.bias": "encoder.layers.16.norm_feed_forward2.bias",
  "encoder.layers.16.norm_feed_forward2.weight": "encoder.layers.16.norm_feed_forward2.weight",
  "encoder.layers.16.norm_out.bias": "encoder.layers.16.norm_out.bias",
  "encoder.layers.16.norm_out.weight": "encoder.layers.16.norm_out.weight",
  "encoder.layers.16.norm_self_att.bias": "encoder.layers.16.norm_self_att.bias",
  "encoder.layers.16.norm_self_att.weight": "encoder.layers.16.norm_self_att.weight",
  "encoder.layers.16.self_attn.linear_out.bias": "encoder.layers.16.self_attn.out_proj.bias",
  "encoder.layers.16.self_attn.linear_out.weight": "encoder.layers.16.self_attn.out_proj.weight",
  "encoder.layers.16.self_attn.linear_pos.weight": "encoder.layers.16.self_attn.pos_proj.weight",
  "encoder.layers.16.self_attn.<qkv>.bias": "encoder.layers.16.self_attn.qkv_proj.bias",
  "encoder.layers.16.self_attn.<qkv>.weight": "encoder.layers.16.self_attn.qkv_proj.weight",
  "encoder.layers.16.self_attn.pos_bias_u": "encoder.layers.16.self_attn.pos_bias_u",
  "encoder.layers.16.self_attn.pos_bias_v": "encoder.layers.16.self_attn.pos_bias_v",
  "encoder.layers.17.conv.batch_norm.bias": "encoder.layers.17.conv.batch_norm.bias",
  "encoder.layers.17.conv.batch_norm.running_mean": "encoder.layers.17.conv.batch_norm.running_mean",
  "encoder.layers.17.conv.batch_norm.running_var": "encoder.layers.17.conv.batch_norm.running_var",
  "encoder.layers.17.conv.batch_norm.weight": "encoder.layers.17.conv.batch_norm.weight",
  "encoder.layers.17.conv.depthwise_conv.bias": "encoder.layers.17.conv.depthwise_conv.bias",
  "encoder.layers.17.conv.depthwise_conv.weight": "encoder.layers.17.conv.depthwise_conv.weight",
  "encoder.layers.17.conv.pointwise_conv1.bias": "encoder.layers.17.conv.pointwise_conv1.bias",
  "encoder.layers.17.conv.pointwise_conv1.weight": "encoder.layers.17.conv.pointwise_conv1.weight",
  "encoder.layers.17.conv.pointwise_conv2.bias": "encoder.layers.17.conv.pointwise_conv2.bias",
  "encoder.layers.17.conv.pointwise_conv2.weight": "encoder.layers.17.conv.pointwise_conv2.weight",
  "encoder.layers.17.feed_forward1.linear1.bias": "encoder.layers.17.feed_forward1.linear1.bias",
  "encoder.layers.17.feed_forward1.linear1.weight": "encoder.layers.17.feed_forward1.linear1.weight",
  "encoder.layers.17.feed_forward1.linear2.bias": "encoder.layers.17.feed_forward1.linear2.bias",
  "encoder.layers.17.feed_forward1.linear2.weight": "encoder.layers.17.feed_forward1.linear2.weight",
  "encoder.layers.17.feed_forward2.linear1.bias": "encoder.layers.17.feed_forward2.linear1.bias",
  "encoder.layers.17.feed_forward2.linear1.weight": "encoder.layers.17.feed_forward2.linear1.weight",
  "encoder.layers.17.feed_forward2.linear2.bias": "encoder.layers.17.feed_forward2.linear2.bias",
  "encoder.layers.17.feed_forward2.linear2.weight": "encoder.layers.17.feed_forward2.linear2.weight",
  "encoder.layers.17.norm_conv.bias": "encoder.layers.17.norm_conv.bias",
  "encoder.layers.17.norm_conv.weight": "encoder.layers.17.norm_conv.weight",
  "encoder.layers.17.norm_feed_forward1.bias": "encoder.layers.17.norm_feed_forward1.bias",
  "encoder.layers.17.norm_feed_forward1.weight": "encoder.layers.17.norm_feed_forward1.weight",
  "encoder.layers.17.norm_feed_forward2.bias": "encoder.layers.17.norm_feed_forward2.bias",
  "encoder.layers.17.norm_feed_forward2.weight": "encoder.layers.17.norm_feed_forward2.weight",
  "encoder.layers.17.norm_out.bias": "encoder.layers.17.norm_out.bias",
  "encoder.layers.17.norm_out.weight": "encoder.layers.17.norm_out.weight",
  "encoder.layers.17.norm_self_att.bias": "encoder.layers.17.norm_self_att.bias",
  "encoder.layers.17.norm_self_att.weight": "encoder.layers.17.norm_self_att.weight",
  "encoder.layers.17.self_attn.linear_out.bias": "encoder.layers.17.self_attn.out_proj.bias",
  "encoder.layers.17.self_attn.linear_out.weight": "encoder.layers.17.self_attn.out_proj.weight",
  "encoder.layers.17.self_attn.linear_pos.weight": "encoder.layers.17.self_attn.pos_proj.weight",
  "encoder.layers.17.self_attn.<qkv>.bias": "encoder.layers.17.self_attn.qkv_proj.bias",
  "encoder.layers.17.self_attn.<qkv>.weight": "encoder.layers.17.self_attn.qkv_proj.weight",
  "encoder.layers.17.self_attn.pos_bias_u": "encoder.layers.17.self_attn.pos_bias_u",
  "encoder.layers.17.self_attn.pos_bias_v": "encoder.layers.17.self_attn.pos_bias_v",
  "encoder.layers.18.conv.batch_norm.bias": "encoder.layers.18.conv.batch_norm.bias",
  "encoder.layers.18.conv.batch_norm.running_mean": "encoder.layers.18.conv.batch_norm.running_mean",
  "encoder.layers.18.conv.batch_norm.running_var": "encoder.layers.18.conv.batch_norm.running_var",
  "encoder.layers.18.conv.batch_norm.weight": "encoder.layers.18.conv.batch_norm.weight",
  "encoder.layers.18.conv.depthwise_conv.bias": "encoder.layers.18.conv.depthwise_conv.bias",
  "encoder.layers.18.conv.depthwise_conv.weight": "encoder.layers.18.conv.depthwise_conv.weight",
  "encoder.layers.18.conv.pointwise_conv1.bias": "encoder.layers.18.conv.pointwise_conv1.bias",
  "encoder.layers.18.conv.pointwise_conv1.weight": "encoder.layers.18.conv.pointwise_conv1.weight",
  "encoder.layers.18.conv.pointwise_conv2.bias": "encoder.layers.18.conv.pointwise_conv2.bias",
  "encoder.layers.18.conv.pointwise_conv2.weight": "encoder.layers.18.conv.pointwise_conv2.weight",
  "encoder.layers.18.feed_forward1.linear1.bias": "encoder.layers.18.feed_forward1.linear1.bias",
  "encoder.layers.18.feed_forward1.linear1.weight": "encoder.layers.18.feed_forward1.linear1.weight",
  "encoder.layers.18.feed_forward1.linear2.bias": "encoder.layers.18.feed_forward1.linear2.bias",
  "encoder.layers.18.feed_forward1.linear2.weight": "encoder.layers.18.feed_forward1.linear2.weight",
  "encoder.layers.18.feed_forward2.linear1.bias": "encoder.layers.18.feed_forward2.linear1.bias",
  "encoder.layers.18.feed_forward2.linear1.weight": "encoder.layers.18.feed_forward2.linear1.weight",
  "encoder.layers.18.feed_forward2.linear2.bias": "encoder.layers.18.feed_forward2.linear2.bias",
  "encoder.layers.18.feed_forward2.linear2.weight": "encoder.layers.18.feed_forward2.linear2.weight",
  "encoder.layers.18.norm_conv.bias": "encoder.layers.18.norm_conv.bias",
  "encoder.layers.18.norm_conv.weight": "encoder.layers.18.norm_conv.weight",
  "encoder.layers.18.norm_feed_forward1.bias": "encoder.layers.18.norm_feed_forward1.bias",
  "encoder.layers.18.norm_feed_forward1.weight": "encoder.layers.18.norm_feed_forward1.weight",
  "encoder.layers.18.norm_feed_forward2.bias": "encoder.layers.18.norm_feed_forward2.bias",
  "encoder.layers.18.norm_feed_forward2.weight": "encoder.layers.18.norm_feed_forward2.weight",
  "encoder.layers.18.norm_out.bias": "encoder.layers.18.norm_out.bias",
  "encoder.layers.18.norm_out.weight": "encoder.layers.18.norm_out.weight",
  "encoder.layers.18.norm_self_att.bias": "encoder.layers.18.norm_self_att.bias",
  "encoder.layers.18.norm_self_att.weight": "encoder.layers.18.norm_self_att.weight",
  "encoder.layers.18.self_attn.linear_out.bias": "encoder.layers.18.self_attn.out_proj.bias",
  "encoder.layers.18.self_attn.linear_out.weight": "encoder.layers.18.self_attn.out_proj.weight",
  "encoder.layers.18.self_attn.linear_pos.weight": "encoder.layers.18.self_attn.pos_proj.weight",
  "encoder.layers.18.self_attn.<qkv>.bias": "encoder.layers.18.self_attn.qkv_proj.bias",
  "encoder.layers.18.self_attn.<qkv>.weight": "encoder.layers.18.self_attn.qkv_proj.weight",
  "encoder.layers.18.self_attn.pos_bias_u": "encoder.layers.18.self_attn.pos_bias_u",
  "encoder.layers.18.self_attn.pos_bias_v": "encoder.layers.18.self_attn.pos_bias_v",
  "encoder.layers.19.conv.batch_norm.bias": "encoder.layers.19.conv.batch_norm.bias",
  "encoder.layers.19.conv.batch_norm.running_mean": "encoder.layers.19.conv.batch_norm.running_mean",
  "encoder.layers.19.conv.batch_norm.running_var": "encoder.layers.19.conv.batch_norm.running_var",
  "encoder.layers.19.conv.batch_norm.weight": "encoder.layers.19.conv.batch_norm.weight",
  "encoder.layers.19.conv.depthwise_conv.bias": "encoder.layers.19.conv.depthwise_conv.bias",
  "encoder.layers.19.conv.depthwise_conv.weight": "encoder.layers.19.conv.depthwise_conv.weight",
  "encoder.layers.19.conv.pointwise_conv1.bias": "encoder.layers.19.conv.pointwise_conv1.bias",
  "encoder.layers.19.conv.pointwise_conv1.weight": "encoder.layers.19.conv.pointwise_conv1.weight",
  "encoder.layers.19.conv.pointwise_conv2.bias": "encoder.layers.19.conv.pointwise_conv2.bias",
  "encoder.layers.19.conv.pointwise_conv2.weight": "encoder.layers.19.conv.pointwise_conv2.weight",
  "encoder.layers.19.feed_forward1.linear1.bias": "encoder.layers.19.feed_forward1.linear1.bias",
  "encoder.layers.19.feed_forward1.linear1.weight": "encoder.layers.19.feed_forward1.linear1.weight",
  "encoder.layers.19.feed_forward1.linear2.bias": "encoder.layers.19.feed_forward1.linear2.bias",
  "encoder.layers.19.feed_forward1.linear2.weight": "encoder.layers.19.feed_forward1.linear2.weight",
  "encoder.layers.19.feed_forward2.linear1.bias": "encoder.layers.19.feed_forward2.linear1.bias",
  "encoder.layers.19.feed_forward2.linear1.weight": "encoder.layers.19.feed_forward2.linear1.weight",
  "encoder.layers.19.feed_forward2.linear2.bias": "encoder.layers.19.feed_forward2.linear2.bias",
  "encoder.layers.19.feed_forward2.linear2.weight": "encoder.layers.19.feed_forward2.linear2.weight",
  "encoder.layers.19.norm_conv.bias": "encoder.layers.19.norm_conv.bias",
  "encoder.layers.19.norm_conv.weight": "encoder.layers.19.norm_conv.weight",
  "encoder.layers.19.norm_feed_forward1.bias": "encoder.layers.19.norm_feed_forward1.bias",
  "encoder.layers.19.norm_feed_forward1.weight": "encoder.layers.19.norm_feed_forward1.weight",
  "encoder.layers.19.norm_feed_forward2.bias": "encoder.layers.19.norm_feed_forward2.bias",
  "encoder.layers.19.norm_feed_forward2.weight": "encoder.layers.19.norm_feed_forward2.weight",
  "encoder.layers.19.norm_out.bias": "encoder.layers.19.norm_out.bias",
  "encoder.layers.19.norm_out.weight": "encoder.layers.19.norm_out.weight",
  "encoder.layers.19.norm_self_att.bias": "encoder.layers.19.norm_self_att.bias",
  "encoder.layers.19.norm_self_att.weight": "encoder.layers.19.norm_self_att.weight",
  "encoder.layers.19.self_attn.linear_out.bias": "encoder.layers.19.self_attn.out_proj.bias",
  "encoder.layers.19.self_attn.linear_out.weight": "encoder.layers.19.self_attn.out_proj.weight",
  "encoder.layers.19.self_attn.linear_pos.weight": "encoder.layers.19.self_attn.pos_proj.weight",
  "encoder.layers.19.self_attn.<qkv>.bias": "encoder.layers.19.self_attn.qkv_proj.bias",
  "encoder.layers.19.self_attn.<qkv>.weight": "encoder.layers.19.self_attn.qkv_proj.weight",
  "encoder.layers.19.self_attn.pos_bias_u": "encoder.layers.19.self_attn.pos_bias_u",
  "encoder.layers.19.self_attn.pos_bias_v": "encoder.layers.19.self_attn.pos_bias_v",
  "encoder.layers.2.conv.batch_norm.bias": "encoder.layers.2.conv.batch_norm.bias",
  "encoder.layers.2.conv.batch_norm.running_mean": "encoder.layers.2.conv.batch_norm.running_mean",
  "encoder.layers.2.conv.batch_norm.running_var": "encoder.layers.2.conv.batch_norm.running_var",
  "encoder.layers.2.conv.batch_norm.weight": "encoder.layers.2.conv.batch_norm.weight",
  "encoder.layers.2.conv.depthwise_conv.bias": "encoder.layers.2.conv.depthwise_conv.bias",
  "encoder.layers.2.conv.depthwise_conv.weight": "encoder.layers.2.conv.depthwise_conv.weight",
  "encoder.layers.2.conv.pointwise_conv1.bias": "encoder.layers.2.conv.pointwise_conv1.bias",
  "encoder.layers.2.conv.pointwise_conv1.weight": "encoder.layers.2.conv.pointwise_conv1.weight",
  "encoder.layers.2.conv.pointwise_conv2.bias": "encoder.layers.2.conv.pointwise_conv2.bias",
  "encoder.layers.2.conv.pointwise_conv2.weight": "encoder.layers.2.conv.pointwise_conv2.weight",
  "encoder.layers.2.feed_forward1.linear1.bias": "encoder.layers.2.feed_forward1.linear1.bias",
  "encoder.layers.2.feed_forward1.linear1.weight": "encoder.layers.2.feed_forward1.linear1.weight",
  "encoder.layers.2.feed_forward1.linear2.bias": "encoder.layers.2.feed_forward1.linear2.bias",
  "encoder.layers.2.feed_forward1.linear2.weight": "encoder.layers.2.feed_forward1.linear2.weight",
  "encoder.layers.2.feed_forward2.linear1.bias": "encoder.layers.2.feed_forward2.linear1.bias",
  "encoder.layers.2.feed_forward2.linear1.weight": "encoder.layers.2.feed_forward2.linear1.weight",
  "encoder.layers.2.feed_forward2.linear2.bias": "encoder.layers.2.feed_forward2.linear2.bias",
  "encoder.layers.2.feed_forward2.linear2.weight": "encoder.layers.2.feed_forward2.linear2.weight",
  "encoder.layers.2.norm_conv.bias": "encoder.layers.2.norm_conv.bias",
  "encoder.layers.2.norm_conv.weight": "encoder.layers.2.norm_conv.weight",
  "encoder.layers.2.norm_feed_forward1.bias": "encoder.layers.2.norm_feed_forward1.bias",
  "encoder.layers.2.norm_feed_forward1.weight": "encoder.layers.2.norm_feed_forward1.weight",
  "encoder.layers.2.norm_feed_forward2.bias": "encoder.layers.2.norm_feed_forward2.bias",
  "encoder.layers.2.norm_feed_forward2.weight": "encoder.layers.2.norm_feed_forward2.weight",
  "encoder.layers.2.norm_out.bias": "encoder.layers.2.norm_out.bias",
  "encoder.layers.2.norm_out.weight": "encoder.layers.2.norm_out.weight",
  "encoder.layers.2.norm_self_att.bias": "encoder.layers.2.norm_self_att.bias",
  "encoder.layers.2.norm_self_att.weight": "encoder.layers.2.norm_self_att.weight",
  "encoder.layers.2.self_attn.linear_out.bias": "encoder.layers.2.self_attn.out_proj.bias",
  "encoder.layers.2.self_attn.linear_out.weight": "encoder.layers.2.self_attn.out_proj.weight",
  "encoder.layers.2.self_attn.linear_pos.weight": "encoder.layers.2.self_attn.pos_proj.weight",
  "encoder.layers.2.self_attn.<qkv>.bias": "encoder.layers.2.self_attn.qkv_proj.bias",
  "encoder.layers.2.self_attn.<qkv>.weight": "encoder.layers.2.self_attn.qkv_proj.weight",
  "encoder.layers.2.self_attn.pos_bias_u": "encoder.layers.2.self_attn.pos_bias_u",
  "encoder.layers.2.self_attn.pos_bias_v": "encoder.layers.2.self_attn.pos_bias_v",
  "encoder.layers.20.conv.batch_norm.bias": "encoder.layers.20.conv.batch_norm.bias",
  "encoder.layers.20.conv.batch_norm.running_mean": "encoder.layers.20.conv.batch_norm.running_mean",
  "encoder.layers.20.conv.batch_norm.running_var": "encoder.layers.20.conv.batch_norm.running_var",
  "encoder.layers.20.conv.batch_norm.weight": "encoder.layers.20.conv.batch_norm.weight",
  "encoder.layers.20.conv.depthwise_conv.bias": "encoder.layers.20.conv.depthwise_conv.bias",
  "encoder.layers.20.conv.depthwise_conv.weight": "encoder.layers.20.conv.depthwise_conv.weight",
  "encoder.layers.20.conv.pointwise_conv1.bias": "encoder.layers.20.conv.pointwise_conv1.bias",
  "encoder.layers.20.conv.pointwise_conv1.weight": "encoder.layers.20.conv.pointwise_conv1.weight",
  "encoder.layers.20.conv.pointwise_conv2.bias": "encoder.layers.20.conv.pointwise_conv2.bias",
  "encoder.layers.20.conv.pointwise_conv2.weight": "encoder.layers.20.conv.pointwise_conv2.weight",
  "encoder.layers.20.feed_forward1.linear1.bias": "encoder.layers.20.feed_forward1.linear1.bias",
  "encoder.layers.20.feed_forward1.linear1.weight": "encoder.layers.20.feed_forward1.linear1.weight",
  "encoder.layers.20.feed_forward1.linear2.bias": "encoder.layers.20.feed_forward1.linear2.bias",
  "encoder.layers.20.feed_forward1.linear2.weight": "encoder.layers.20.feed_forward1.linear2.weight",
  "encoder.layers.20.feed_forward2.linear1.bias": "encoder.layers.20.feed_forward2.linear1.bias",
  "encoder.layers.20.feed_forward2.linear1.weight": "encoder.layers.20.feed_forward2.linear1.weight",
  "encoder.layers.20.feed_forward2.linear2.bias": "encoder.layers.20.feed_forward2.linear2.bias",
  "encoder.layers.20.feed_forward2.linear2.weight": "encoder.layers.20.feed_forward2.linear2.weight",
  "encoder.layers.20.norm_conv.bias": "encoder.layers.20.norm_conv.bias",
  "encoder.layers.20.norm_conv.weight": "encoder.layers.20.norm_conv.weight",
  "encoder.layers.20.norm_feed_forward1.bias": "encoder.layers.20.norm_feed_forward1.bias",
  "encoder.layers.20.norm_feed_forward1.weight": "encoder.layers.20.norm_feed_forward1.weight",
  "encoder.layers.20.norm_feed_forward2.bias": "encoder.layers.20.norm_feed_forward2.bias",
  "encoder.layers.20.norm_feed_forward2.weight": "encoder.layers.20.norm_feed_forward2.weight",
  "encoder.layers.20.norm_out.bias": "encoder.layers.20.norm_out.bias",
  "encoder.layers.20.norm_out.weight": "encoder.layers.20.norm_out.weight",
  "encoder.layers.20.norm_self_att.bias": "encoder.layers.20.norm_self_att.bias",
  "encoder.layers.20.norm_self_att.weight": "encoder.layers.20.norm_self_att.weight",
  "encoder.layers.20.self_attn.linear_out.bias": "encoder.layers.20.self_attn.out_proj.bias",
  "encoder.layers.20.self_attn.linear_out.weight": "encoder.layers.20.self_attn.out_proj.weight",
  "encoder.layers.20.self_attn.linear_pos.weight": "encoder.layers.20.self_attn.pos_proj.weight",
  "encoder.layers.20.self_attn.<qkv>.bias": "encoder.layers.20.self_attn.qkv_proj.bias",
  "encoder.layers.20.self_attn.<qkv>.weight": "encoder.layers.20.self_attn.qkv_proj.weight",
  "encoder.layers.20.self_attn.pos_bias_u": "encoder.layers.20.self_attn.pos_bias_u",
  "encoder.layers.20.self_attn.pos_bias_v": "encoder.layers.20.self_attn.pos_bias_v",
  "encoder.layers.21.conv.batch_norm.bias": "encoder.layers.21.conv.batch_norm.bias",
  "encoder.layers.21.conv.batch_norm.running_mean": "encoder.layers.21.conv.batch_norm.running_mean",
  "encoder.layers.21.conv.batch_norm.running_var": "encoder.layers.21.conv.batch_norm.running_var",
  "encoder.layers.21.conv.batch_norm.weight": "encoder.layers.21.conv.batch_norm.weight",
  "encoder.layers.21.conv.depthwise_conv.bias": "encoder.layers.21.conv.depthwise_conv.bias",
  "encoder.layers.21.conv.depthwise_conv.weight": "encoder.layers.21.conv.depthwise_conv.weight",
  "encoder.layers.21.conv.pointwise_conv1.bias": "encoder.layers.21.conv.pointwise_conv1.bias",
  "encoder.layers.21.conv.pointwise_conv1.weight": "encoder.layers.21.conv.pointwise_conv1.weight",
  "encoder.layers.21.conv.pointwise_conv2.bias": "encoder.layers.21.conv.pointwise_conv2.bias",
  "encoder.layers.21.conv.pointwise_conv2.weight": "encoder.layers.21.conv.pointwise_conv2.weight",
  "encoder.layers.21.feed_forward1.linear1.bias": "encoder.layers.21.feed_forward1.linear1.bias",
  "encoder.layers.21.feed_forward1.linear1.weight": "encoder.layers.21.feed_forward1.linear1.weight",
  "encoder.layers.21.feed_forward1.linear2.bias": "encoder.layers.21.feed_forward1.linear2.bias",
  "encoder.layers.21.feed_forward1.linear2.weight": "encoder.layers.21.feed_forward1.linear2.weight",
  "encoder.layers.21.feed_forward2.linear1.bias": "encoder.layers.21.feed_forward2.linear1.bias",
  "encoder.layers.21.feed_forward2.linear1.weight": "encoder.layers.21.feed_forward2.linear1.weight",
  "encoder.layers.21.feed_forward2.linear2.bias": "encoder.layers.21.feed_forward2.linear2.bias",
  "encoder.layers.21.feed_forward2.linear2.weight": "encoder.layers.21.feed_forward2.linear2.weight",
  "encoder.layers.21.norm_conv.bias": "encoder.layers.21.norm_conv.bias",
  "encoder.layers.21.norm_conv.weight": "encoder.layers.21.norm_conv.weight",
  "encoder.layers.21.norm_feed_forward1.bias": "encoder.layers.21.norm_feed_forward1.bias",
  "encoder.layers.21.norm_feed_forward1.weight": "encoder.layers.21.norm_feed_forward1.weight",
  "encoder.layers.21.norm_feed_forward2.bias": "encoder.layers.21.norm_feed_forward2.bias",
  "encoder.layers.21.norm_feed_forward2.weight": "encoder.layers.21.norm_feed_forward2.weight",
  "encoder.layers.21.norm_out.bias": "encoder.layers.21.norm_out.bias",
  "encoder.layers.21.norm_out.weight": "encoder.layers.21.norm_out.weight",
  "encoder.layers.21.norm_self_att.bias": "encoder.layers.21.norm_self_att.bias",
  "encoder.layers.21.norm_self_att.weight": "encoder.layers.21.norm_self_att.weight",
  "encoder.layers.21.self_attn.linear_out.bias": "encoder.layers.21.self_attn.out_proj.bias",
  "encoder.layers.21.self_attn.linear_out.weight": "encoder.layers.21.self_attn.out_proj.weight",
  "encoder.layers.21.self_attn.linear_pos.weight": "encoder.layers.21.self_attn.pos_proj.weight",
  "encoder.layers.21.self_attn.<qkv>.bias": "encoder.layers.21.self_attn.qkv_proj.bias",
  "encoder.layers.21.self_attn.<qkv>.weight": "encoder.layers.21.self_attn.qkv_proj.weight",
  "encoder.layers.21.self_attn.pos_bias_u": "encoder.layers.21.self_attn.pos_bias_u",
  "encoder.layers.21.self_attn.pos_bias_v": "encoder.layers.21.self_attn.pos_bias_v",
  "encoder.layers.22.conv.batch_norm.bias": "encoder.layers.22.conv.batch_norm.bias",
  "encoder.layers.22.conv.batch_norm.running_mean": "encoder.layers.22.conv.batch_norm.running_mean",
  "encoder.layers.22.conv.batch_norm.running_var": "encoder.layers.22.conv.batch_norm.running_var",
  "encoder.layers.22.conv.batch_norm.weight": "encoder.layers.22.conv.batch_norm.weight",
  "encoder.layers.22.conv.depthwise_conv.bias": "encoder.layers.22.conv.depthwise_conv.bias",
  "encoder.layers.22.conv.depthwise_conv.weight": "encoder.layers.22.conv.depthwise_conv.weight",
  "encoder.layers.22.conv.pointwise_conv1.bias": "encoder.layers.22.conv.pointwise_conv1.bias",
  "encoder.layers.22.conv.pointwise_conv1.weight": "encoder.layers.22.conv.pointwise_conv1.weight",
  "encoder.layers.22.conv.pointwise_conv2.bias": "encoder.layers.22.conv.pointwise_conv2.bias",
  "encoder.layers.22.conv.pointwise_conv2.weight": "encoder.layers.22.conv.pointwise_conv2.weight",
  "encoder.layers.22.feed_forward1.linear1.bias": "encoder.layers.22.feed_forward1.linear1.bias",
  "encoder.layers.22.feed_forward1.linear1.weight": "encoder.layers.22.feed_forward1.linear1.weight",
  "encoder.layers.22.feed_forward1.linear2.bias": "encoder.layers.22.feed_forward1.linear2.bias",
  "encoder.layers.22.feed_forward1.linear2.weight": "encoder.layers.22.feed_forward1.linear2.weight",
  "encoder.layers.22.feed_forward2.linear1.bias": "encoder.layers.22.feed_forward2.linear1.bias",
  "encoder.layers.22.feed_forward2.linear1.weight": "encoder.layers.22.feed_forward2.linear1.weight",
  "encoder.layers.22.feed_forward2.linear2.bias": "encoder.layers.22.feed_forward2.linear2.bias",
  "encoder.layers.22.feed_forward2.linear2.weight": "encoder.layers.22.feed_forward2.linear2.weight",
  "encoder.layers.22.norm_conv.bias": "encoder.layers.22.norm_conv.bias",
  "encoder.layers.22.norm_conv.weight": "encoder.layers.22.norm_conv.weight",
  "encoder.layers.22.norm_feed_forward1.bias": "encoder.layers.22.norm_feed_forward1.bias",
  "encoder.layers.22.norm_feed_forward1.weight": "encoder.layers.22.norm_feed_forward1.weight",
  "encoder.layers.22.norm_feed_forward2.bias": "encoder.layers.22.norm_feed_forward2.bias",
  "encoder.layers.22.norm_feed_forward2.weight": "encoder.layers.22.norm_feed_forward2.weight",
  "encoder.layers.22.norm_out.bias": "encoder.layers.22.norm_out.bias",
  "encoder.layers.22.norm_out.weight": "encoder.layers.22.norm_out.weight",
  "encoder.layers.22.norm_self_att.bias": "encoder.layers.22.norm_self_att.bias",
  "encoder.layers.22.norm_self_att.weight": "encoder.layers.22.norm_self_att.weight",
  "encoder.layers.22.self_attn.linear_out.bias": "encoder.layers.22.self_attn.out_proj.bias",
  "encoder.layers.22.self_attn.linear_out.weight": "encoder.layers.22.self_attn.out_proj.weight",
  "encoder.layers.22.self_attn.linear_pos.weight": "encoder.layers.22.self_attn.pos_proj.weight",
  "encoder.layers.22.self_attn.<qkv>.bias": "encoder.layers.22.self_attn.qkv_proj.bias",
  "encoder.layers.22.self_attn.<qkv>.weight": "encoder.layers.22.self_attn.qkv_proj.weight",
  "encoder.layers.22.self_attn.pos_bias_u": "encoder.layers.22.self_attn.pos_bias_u",
  "encoder.layers.22.self_attn.pos_bias_v": "encoder.layers.22.self_attn.pos_bias_v",
  "encoder.layers.23.conv.batch_norm.bias": "encoder.layers.23.conv.batch_norm.bias",
  "encoder.layers.23.conv.batch_norm.running_mean": "encoder.layers.23.conv.batch_norm.running_mean",
  "encoder.layers.23.conv.batch_norm.running_var": "encoder.layers.23.conv.batch_norm.running_var",
  "encoder.layers.23.conv.batch_norm.weight": "encoder.layers.23.conv.batch_norm.weight",
  "encoder.layers.23.conv.depthwise_conv.bias": "encoder.layers.23.conv.depthwise_conv.bias",
  "encoder.layers.23.conv.depthwise_conv.weight": "encoder.layers.23.conv.depthwise_conv.weight",
  "encoder.layers.23.conv.pointwise_conv1.bias": "encoder.layers.23.conv.pointwise_conv1.bias",
  "encoder.layers.23.conv.pointwise_conv1.weight": "encoder.layers.23.conv.pointwise_conv1.weight",
  "encoder.layers.23.conv.pointwise_conv2.bias": "encoder.layers.23.conv.pointwise_conv2.bias",
  "encoder.layers.23.conv.pointwise_conv2.weight": "encoder.layers.23.conv.pointwise_conv2.weight",
  "encoder.layers.23.feed_forward1.linear1.bias": "encoder.layers.23.feed_forward1.linear1.bias",
  "encoder.layers.23.feed_forward1.linear1.weight": "encoder.layers.23.feed_forward1.linear1.weight",
  "encoder.layers.23.feed_forward1.linear2.bias": "encoder.layers.23.feed_forward1.linear2.bias",
  "encoder.layers.23.feed_forward1.linear2.weight": "encoder.layers.23.feed_forward1.linear2.weight",
  "encoder.layers.23.feed_forward2.linear1.bias": "encoder.layers.23.feed_forward2.linear1.bias",
  "encoder.layers.23.feed_forward2.linear1.weight": "encoder.layers.23.feed_forward2.linear1.weight",
  "encoder.layers.23.feed_forward2.linear2.bias": "encoder.layers.23.feed_forward2.linear2.bias",
  "encoder.layers.23.feed_forward2.linear2.weight": "encoder.layers.23.feed_forward2.linear2.weight",
  "encoder.layers.23.norm_conv.bias": "encoder.layers.23.norm_conv.bias",
  "encoder.layers.23.norm_conv.weight": "encoder.layers.23.norm_conv.weight",
  "encoder.layers.23.norm_feed_forward1.bias": "encoder.layers.23.norm_feed_forward1.bias",
  "encoder.layers.23.norm_feed_forward1.weight": "encoder.layers.23.norm_feed_forward1.weight",
  "encoder.layers.23.norm_feed_forward2.bias": "encoder.layers.23.norm_feed_forward2.bias",
  "encoder.layers.23.norm_feed_forward2.weight": "encoder.layers.23.norm_feed_forward2.weight",
  "encoder.layers.23.norm_out.bias": "encoder.layers.23.norm_out.bias",
  "encoder.layers.23.norm_out.weight": "encoder.layers.23.norm_out.weight",
  "encoder.layers.23.norm_self_att.bias": "encoder.layers.23.norm_self_att.bias",
  "encoder.layers.23.norm_self_att.weight": "encoder.layers.23.norm_self_att.weight",
  "encoder.layers.23.self_attn.linear_out.bias": "encoder.layers.23.self_attn.out_proj.bias",
  "encoder.layers.23.self_attn.linear_out.weight": "encoder.layers.23.self_attn.out_proj.weight",
  "encoder.layers.23.self_attn.linear_pos.weight": "encoder.layers.23.self_attn.pos_proj.weight",
  "encoder.layers.23.self_attn.<qkv>.bias": "encoder.layers.23.self_attn.qkv_proj.bias",
  "encoder.layers.23.self_attn.<qkv>.weight": "encoder.layers.23.self_attn.qkv_proj.weight",
  "encoder.layers.23.self_attn.pos_bias_u": "encoder.layers.23.self_attn.pos_bias_u",
  "encoder.layers.23.self_attn.pos_bias_v": "encoder.layers.23.self_attn.pos_bias_v",
  "encoder.layers.24.conv.batch_norm.bias": "encoder.layers.24.conv.batch_norm.bias",
  "encoder.layers.24.conv.batch_norm.running_mean": "encoder.layers.24.conv.batch_norm.running_mean",
  "encoder.layers.24.conv.batch_norm.running_var": "encoder.layers.24.conv.batch_norm.running_var",
  "encoder.layers.24.conv.batch_norm.weight": "encoder.layers.24.conv.batch_norm.weight",
  "encoder.layers.24.conv.depthwise_conv.bias": "encoder.layers.24.conv.depthwise_conv.bias",
  "encoder.layers.24.conv.depthwise_conv.weight": "encoder.layers.24.conv.depthwise_conv.weight",
  "encoder.layers.24.conv.pointwise_conv1.bias": "encoder.layers.24.conv.pointwise_conv1.bias",
  "encoder.layers.24.conv.pointwise_conv1.weight": "encoder.layers.24.conv.pointwise_conv1.weight",
  "encoder.layers.24.conv.pointwise_conv2.bias": "encoder.layers.24.conv.pointwise_conv2.bias",
  "encoder.layers.24.conv.pointwise_conv2.weight": "encoder.layers.24.conv.pointwise_conv2.weight",
  "encoder.layers.24.feed_forward1.linear1.bias": "encoder.layers.24.feed_forward1.linear1.bias",
  "encoder.layers.24.feed_forward1.linear1.weight": "encoder.layers.24.feed_forward1.linear1.weight",
  "encoder.layers.24.feed_forward1.linear2.bias": "encoder.layers.24.feed_forward1.linear2.bias",
  "encoder.layers.24.feed_forward1.linear2.weight": "encoder.layers.24.feed_forward1.linear2.weight",
  "encoder.layers.24.feed_forward2.linear1.bias": "encoder.layers.24.feed_forward2.linear1.bias",
  "encoder.layers.24.feed_forward2.linear1.weight": "encoder.layers.24.feed_forward2.linear1.weight",
  "encoder.layers.24.feed_forward2.linear2.bias": "encoder.layers.24.feed_forward2.linear2.bias",
  "encoder.layers.24.feed_forward2.linear2.weight": "encoder.layers.24.feed_forward2.linear2.weight",
  "encoder.layers.24.norm_conv.bias": "encoder.layers.24.norm_conv.bias",
  "encoder.layers.24.norm_conv.weight": "encoder.layers.24.norm_conv.weight",
  "encoder.layers.24.norm_feed_forward1.bias": "encoder.layers.24.norm_feed_forward1.bias",
  "encoder.layers.24.norm_feed_forward1.weight": "encoder.layers.24.norm_feed_forward1.weight",
  "encoder.layers.24.norm_feed_forward2.bias": "encoder.layers.24.norm_feed_forward2.bias",
  "encoder.layers.24.norm_feed_forward2.weight": "encoder.layers.24.norm_feed_forward2.weight",
  "encoder.layers.24.norm_out.bias": "encoder.layers.24.norm_out.bias",
  "encoder.layers.24.norm_out.weight": "encoder.layers.24.norm_out.weight",
  "encoder.layers.24.norm_self_att.bias": "encoder.layers.24.norm_self_att.bias",
  "encoder.layers.24.norm_self_att.weight": "encoder.layers.24.norm_self_att.weight",
  "encoder.layers.24.self_attn.linear_out.bias": "encoder.layers.24.self_attn.out_proj.bias",
  "encoder.layers.24.self_attn.linear_out.weight": "encoder.layers.24.self_attn.out_proj.weight",
  "encoder.layers.24.self_attn.linear_pos.weight": "encoder.layers.24.self_attn.pos_proj.weight",
  "encoder.layers.24.self_attn.<qkv>.bias": "encoder.layers.24.self_attn.qkv_proj.bias",
  "encoder.layers.24.self_attn.<qkv>.weight": "encoder.layers.24.self_attn.qkv_proj.weight",
  "encoder.layers.24.self_attn.pos_bias_u": "encoder.layers.24.self_attn.pos_bias_u",
  "encoder.layers.24.self_attn.pos_bias_v": "encoder.layers.24.self_attn.pos_bias_v",
  "encoder.layers.25.conv.batch_norm.bias": "encoder.layers.25.conv.batch_norm.bias",
  "encoder.layers.25.conv.batch_norm.running_mean": "encoder.layers.25.conv.batch_norm.running_mean",
  "encoder.layers.25.conv.batch_norm.running_var": "encoder.layers.25.conv.batch_norm.running_var",
  "encoder.layers.25.conv.batch_norm.weight": "encoder.layers.25.conv.batch_norm.weight",
  "encoder.layers.25.conv.depthwise_conv.bias": "encoder.layers.25.conv.depthwise_conv.bias",
  "encoder.layers.25.conv.depthwise_conv.weight": "encoder.layers.25.conv.depthwise_conv.weight",
  "encoder.layers.25.conv.pointwise_conv1.bias": "encoder.layers.25.conv.pointwise_conv1.bias",
  "encoder.layers.25.conv.pointwise_conv1.weight": "encoder.layers.25.conv.pointwise_conv1.weight",
  "encoder.layers.25.conv.pointwise_conv2.bias": "encoder.layers.25.conv.pointwise_conv2.bias",
  "encoder.layers.25.conv.pointwise_conv2.weight": "encoder.layers.25.conv.pointwise_conv2.weight",
  "encoder.layers.25.feed_forward1.linear1.bias": "encoder.layers.25.feed_forward1.linear1.bias",
  "encoder.layers.25.feed_forward1.linear1.weight": "encoder.layers.25.feed_forward1.linear1.weight",
  "encoder.layers.25.feed_forward1.linear2.bias": "encoder.layers.25.feed_forward1.linear2.bias",
  "encoder.layers.25.feed_forward1.linear2.weight": "encoder.layers.25.feed_forward1.linear2.weight",
  "encoder.layers.25.feed_forward2.linear1.bias": "encoder.layers.25.feed_forward2.linear1.bias",
  "encoder.layers.25.feed_forward2.linear1.weight": "encoder.layers.25.feed_forward2.linear1.weight",
  "encoder.layers.25.feed_forward2.linear2.bias": "encoder.layers.25.feed_forward2.linear2.bias",
  "encoder.layers.25.feed_forward2.linear2.weight": "encoder.layers.25.feed_forward2.linear2.weight",
  "encoder.layers.25.norm_conv.bias": "encoder.layers.25.norm_conv.bias",
  "encoder.layers.25.norm_conv.weight": "encoder.layers.25.norm_conv.weight",
  "encoder.layers.25.norm_feed_forward1.bias": "encoder.layers.25.norm_feed_forward1.bias",
  "encoder.layers.25.norm_feed_forward1.weight": "encoder.layers.25.norm_feed_forward1.weight",
  "encoder.layers.25.norm_feed_forward2.bias": "encoder.layers.25.norm_feed_forward2.bias",
  "encoder.layers.25.norm_feed_forward2.weight": "encoder.layers.25.norm_feed_forward2.weight",
  "encoder.layers.25.norm_out.bias": "encoder.layers.25.norm_out.bias",
  "encoder.layers.25.norm_out.weight": "encoder.layers.25.norm_out.weight",
  "encoder.layers.25.norm_self_att.bias": "encoder.layers.25.norm_self_att.bias",
  "encoder.layers.25.norm_self_att.weight": "encoder.layers.25.norm_self_att.weight",
  "encoder.layers.25.self_attn.linear_out.bias": "encoder.layers.25.self_attn.out_proj.bias",
  "encoder.layers.25.self_attn.linear_out.weight": "encoder.layers.25.self_attn.out_proj.weight",
  "encoder.layers.25.self_attn.linear_pos.weight": "encoder.layers.25.self_attn.pos_proj.weight",
  "encoder.layers.25.self_attn.<qkv>.bias": "encoder.layers.25.self_attn.qkv_proj.bias",
  "encoder.layers.25.self_attn.<qkv>.weight": "encoder.layers.25.self_attn.qkv_proj.weight",
  "encoder.layers.25.self_attn.pos_bias_u": "encoder.layers.25.self_attn.pos_bias_u",
  "encoder.layers.25.self_attn.pos_bias_v": "encoder.layers.25.self_attn.pos_bias_v",
  "encoder.layers.26.conv.batch_norm.bias": "encoder.layers.26.conv.batch_norm.bias",
  "encoder.layers.26.conv.batch_norm.running_mean": "encoder.layers.26.conv.batch_norm.running_mean",
  "encoder.layers.26.conv.batch_norm.running_var": "encoder.layers.26.conv.batch_norm.running_var",
  "encoder.layers.26.conv.batch_norm.weight": "encoder.layers.26.conv.batch_norm.weight",
  "encoder.layers.26.conv.depthwise_conv.bias": "encoder.layers.26.conv.depthwise_conv.bias",
  "encoder.layers.26.conv.depthwise_conv.weight": "encoder.layers.26.conv.depthwise_conv.weight",
  "encoder.layers.26.conv.pointwise_conv1.bias": "encoder.layers.26.conv.pointwise_conv1.bias",
  "encoder.layers.26.conv.pointwise_conv1.weight": "encoder.layers.26.conv.pointwise_conv1.weight",
  "encoder.layers.26.conv.pointwise_conv2.bias": "encoder.layers.26.conv.pointwise_conv2.bias",
  "encoder.layers.26.conv.pointwise_conv2.weight": "encoder.layers.26.conv.pointwise_conv2.weight",
  "encoder.layers.26.feed_forward1.linear1.bias": "encoder.layers.26.feed_forward1.linear1.bias",
  "encoder.layers.26.feed_forward1.linear1.weight": "encoder.layers.26.feed_forward1.linear1.weight",
  "encoder.layers.26.feed_forward1.linear2.bias": "encoder.layers.26.feed_forward1.linear2.bias",
  "encoder.layers.26.feed_forward1.linear2.weight": "encoder.layers.26.feed_forward1.linear2.weight",
  "encoder.layers.26.feed_forward2.linear1.bias": "encoder.layers.26.feed_forward2.linear1.bias",
  "encoder.layers.26.feed_forward2.linear1.weight": "encoder.layers.26.feed_forward2.linear1.weight",
  "encoder.layers.26.feed_forward2.linear2.bias": "encoder.layers.26.feed_forward2.linear2.bias",
  "encoder.layers.26.feed_forward2.linear2.weight": "encoder.layers.26.feed_forward2.linear2.weight",
  "encoder.layers.26.norm_conv.bias": "encoder.layers.26.norm_conv.bias",
  "encoder.layers.26.norm_conv.weight": "encoder.layers.26.norm_conv.weight",
  "encoder.layers.26.norm_feed_forward1.bias": "encoder.layers.26.norm_feed_forward1.bias",
  "encoder.layers.26.norm_feed_forward1.weight": "encoder.layers.26.norm_feed_forward1.weight",
  "encoder.layers.26.norm_feed_forward2.bias": "encoder.layers.26.norm_feed_forward2.bias",
  "encoder.layers.26.norm_feed_forward2.weight": "encoder.layers.26.norm_feed_forward2.weight",
  "encoder.layers.26.norm_out.bias": "encoder.layers.26.norm_out.bias",
  "encoder.layers.26.norm_out.weight": "encoder.layers.26.norm_out.weight",
  "encoder.layers.26.norm_self_att.bias": "encoder.layers.26.norm_self_att.bias",
  "encoder.layers.26.norm_self_att.weight": "encoder.layers.26.norm_self_att.weight",
  "encoder.layers.26.self_attn.linear_out.bias": "encoder.layers.26.self_attn.out_proj.bias",
  "encoder.layers.26.self_attn.linear_out.weight": "encoder.layers.26.self_attn.out_proj.weight",
  "encoder.layers.26.self_attn.linear_pos.weight": "encoder.layers.26.self_attn.pos_proj.weight",
  "encoder.layers.26.self_attn.<qkv>.bias": "encoder.layers.26.self_attn.qkv_proj.bias",
  "encoder.layers.26.self_attn.<qkv>.weight": "encoder.layers.26.self_attn.qkv_proj.weight",
  "encoder.layers.26.self_attn.pos_bias_u": "encoder.layers.26.self_attn.pos_bias_u",
  "encoder.layers.26.self_attn.pos_bias_v": "encoder.layers.26.self_attn.pos_bias_v",
  "encoder.layers.27.conv.batch_norm.bias": "encoder.layers.27.conv.batch_norm.bias",
  "encoder.layers.27.conv.batch_norm.running_mean": "encoder.layers.27.conv.batch_norm.running_mean",
  "encoder.layers.27.conv.batch_norm.running_var": "encoder.layers.27.conv.batch_norm.running_var",
  "encoder.layers.27.conv.batch_norm.weight": "encoder.layers.27.conv.batch_norm.weight",
  "encoder.layers.27.conv.depthwise_conv.bias": "encoder.layers.27.conv.depthwise_conv.bias",
  "encoder.layers.27.conv.depthwise_conv.weight": "encoder.layers.27.conv.depthwise_conv.weight",
  "encoder.layers.27.conv.pointwise_conv1.bias": "encoder.layers.27.conv.pointwise_conv1.bias",
  "encoder.layers.27.conv.pointwise_conv1.weight": "encoder.layers.27.conv.pointwise_conv1.weight",
  "encoder.layers.27.conv.pointwise_conv2.bias": "encoder.layers.27.conv.pointwise_conv2.bias",
  "encoder.layers.27.conv.pointwise_conv2.weight": "encoder.layers.27.conv.pointwise_conv2.weight",
  "encoder.layers.27.feed_forward1.linear1.bias": "encoder.layers.27.feed_forward1.linear1.bias",
  "encoder.layers.27.feed_forward1.linear1.weight": "encoder.layers.27.feed_forward1.linear1.weight",
  "encoder.layers.27.feed_forward1.linear2.bias": "encoder.layers.27.feed_forward1.linear2.bias",
  "encoder.layers.27.feed_forward1.linear2.weight": "encoder.layers.27.feed_forward1.linear2.weight",
  "encoder.layers.27.feed_forward2.linear1.bias": "encoder.layers.27.feed_forward2.linear1.bias",
  "encoder.layers.27.feed_forward2.linear1.weight": "encoder.layers.27.feed_forward2.linear1.weight",
  "encoder.layers.27.feed_forward2.linear2.bias": "encoder.layers.27.feed_forward2.linear2.bias",
  "encoder.layers.27.feed_forward2.linear2.weight": "encoder.layers.27.feed_forward2.linear2.weight",
  "encoder.layers.27.norm_conv.bias": "encoder.layers.27.norm_conv.bias",
  "encoder.layers.27.norm_conv.weight": "encoder.layers.27.norm_conv.weight",
  "encoder.layers.27.norm_feed_forward1.bias": "encoder.layers.27.norm_feed_forward1.bias",
  "encoder.layers.27.norm_feed_forward1.weight": "encoder.layers.27.norm_feed_forward1.weight",
  "encoder.layers.27.norm_feed_forward2.bias": "encoder.layers.27.norm_feed_forward2.bias",
  "encoder.layers.27.norm_feed_forward2.weight": "encoder.layers.27.norm_feed_forward2.weight",
  "encoder.layers.27.norm_out.bias": "encoder.layers.27.norm_out.bias",
  "encoder.layers.27.norm_out.weight": "encoder.layers.27.norm_out.weight",
  "encoder.layers.27.norm_self_att.bias": "encoder.layers.27.norm_self_att.bias",
  "encoder.layers.27.norm_self_att.weight": "encoder.layers.27.norm_self_att.weight",
  "encoder.layers.27.self_attn.linear_out.bias": "encoder.layers.27.self_attn.out_proj.bias",
  "encoder.layers.27.self_attn.linear_out.weight": "encoder.layers.27.self_attn.out_proj.weight",
  "encoder.layers.27.self_attn.linear_pos.weight": "encoder.layers.27.self_attn.pos_proj.weight",
  "encoder.layers.27.self_attn.<qkv>.bias": "encoder.layers.27.self_attn.qkv_proj.bias",
  "encoder.layers.27.self_attn.<qkv>.weight": "encoder.layers.27.self_attn.qkv_proj.weight",
  "encoder.layers.27.self_attn.pos_bias_u": "encoder.layers.27.self_attn.pos_bias_u",
  "encoder.layers.27.self_attn.pos_bias_v": "encoder.layers.27.self_attn.pos_bias_v",
  "encoder.layers.28.conv.batch_norm.bias": "encoder.layers.28.conv.batch_norm.bias",
  "encoder.layers.28.conv.batch_norm.running_mean": "encoder.layers.28.conv.batch_norm.running_mean",
  "encoder.layers.28.conv.batch_norm.running_var": "encoder.layers.28.conv.batch_norm.running_var",
  "encoder.layers.28.conv.batch_norm.weight": "encoder.layers.28.conv.batch_norm.weight",
  "encoder.layers.28.conv.depthwise_conv.bias": "encoder.layers.28.conv.depthwise_conv.bias",
  "encoder.layers.28.conv.depthwise_conv.weight": "encoder.layers.28.conv.depthwise_conv.weight",
  "encoder.layers.28.conv.pointwise_conv1.bias": "encoder.layers.28.conv.pointwise_conv1.bias",
  "encoder.layers.28.conv.pointwise_conv1.weight": "encoder.layers.28.conv.pointwise_conv1.weight",
  "encoder.layers.28.conv.pointwise_conv2.bias": "encoder.layers.28.conv.pointwise_conv2.bias",
  "encoder.layers.28.conv.pointwise_conv2.weight": "encoder.layers.28.conv.pointwise_conv2.weight",
  "encoder.layers.28.feed_forward1.linear1.bias": "encoder.layers.28.feed_forward1.linear1.bias",
  "encoder.layers.28.feed_forward1.linear1.weight": "encoder.layers.28.feed_forward1.linear1.weight",
  "encoder.layers.28.feed_forward1.linear2.bias": "encoder.layers.28.feed_forward1.linear2.bias",
  "encoder.layers.28.feed_forward1.linear2.weight": "encoder.layers.28.feed_forward1.linear2.weight",
  "encoder.layers.28.feed_forward2.linear1.bias": "encoder.layers.28.feed_forward2.linear1.bias",
  "encoder.layers.28.feed_forward2.linear1.weight": "encoder.layers.28.feed_forward2.linear1.weight",
  "encoder.layers.28.feed_forward2.linear2.bias": "encoder.layers.28.feed_forward2.linear2.bias",
  "encoder.layers.28.feed_forward2.linear2.weight": "encoder.layers.28.feed_forward2.linear2.weight",
  "encoder.layers.28.norm_conv.bias": "encoder.layers.28.norm_conv.bias",
  "encoder.layers.28.norm_conv.weight": "encoder.layers.28.norm_conv.weight",
  "encoder.layers.28.norm_feed_forward1.bias": "encoder.layers.28.norm_feed_forward1.bias",
  "encoder.layers.28.norm_feed_forward1.weight": "encoder.layers.28.norm_feed_forward1.weight",
  "encoder.layers.28.norm_feed_forward2.bias": "encoder.layers.28.norm_feed_forward2.bias",
  "encoder.layers.28.norm_feed_forward2.weight": "encoder.layers.28.norm_feed_forward2.weight",
  "encoder.layers.28.norm_out.bias": "encoder.layers.28.norm_out.bias",
  "encoder.layers.28.norm_out.weight": "encoder.layers.28.norm_out.weight",
  "encoder.layers.28.norm_self_att.bias": "encoder.layers.28.norm_self_att.bias",
  "encoder.layers.28.norm_self_att.weight": "encoder.layers.28.norm_self_att.weight",
  "encoder.layers.28.self_attn.linear_out.bias": "encoder.layers.28.self_attn.out_proj.bias",
  "encoder.layers.28.self_attn.linear_out.weight": "encoder.layers.28.self_attn.out_proj.weight",
  "encoder.layers.28.self_attn.linear_pos.weight": "encoder.layers.28.self_attn.pos_proj.weight",
  "encoder.layers.28.self_attn.<qkv>.bias": "encoder.layers.28.self_attn.qkv_proj.bias",
  "encoder.layers.28.self_attn.<qkv>.weight": "encoder.layers.28.self_attn.qkv_proj.weight",
  "encoder.layers.28.self_attn.pos_bias_u": "encoder.layers.28.self_attn.pos_bias_u",
  "encoder.layers.28.self_attn.pos_bias_v": "encoder.layers.28.self_attn.pos_bias_v",
  "encoder.layers.29.conv.batch_norm.bias": "encoder.layers.29.conv.batch_norm.bias",
  "encoder.layers.29.conv.batch_norm.running_mean": "encoder.layers.29.conv.batch_norm.running_mean",
  "encoder.layers.29.conv.batch_norm.running_var": "encoder.layers.29.conv.batch_norm.running_var",
  "encoder.layers.29.conv.batch_norm.weight": "encoder.layers.29.conv.batch_norm.weight",
  "encoder.layers.29.conv.depthwise_conv.bias": "encoder.layers.29.conv.depthwise_conv.bias",
  "encoder.layers.29.conv.depthwise_conv.weight": "encoder.layers.29.conv.depthwise_conv.weight",
  "encoder.layers.29.conv.pointwise_conv1.bias": "encoder.layers.29.conv.pointwise_conv1.bias",
  "encoder.layers.29.conv.pointwise_conv1.weight": "encoder.layers.29.conv.pointwise_conv1.weight",
  "encoder.layers.29.conv.pointwise_conv2.bias": "encoder.layers.29.conv.pointwise_conv2.bias",
  "encoder.layers.29.conv.pointwise_conv2.weight": "encoder.layers.29.conv.pointwise_conv2.weight",
  "encoder.layers.29.feed_forward1.linear1.bias": "encoder.layers.29.feed_forward1.linear1.bias",
  "encoder.layers.29.feed_forward1.linear1.weight": "encoder.layers.29.feed_forward1.linear1.weight",
  "encoder.layers.29.feed_forward1.linear2.bias": "encoder.layers.29.feed_forward1.linear2.bias",
  "encoder.layers.29.feed_forward1.linear2.weight": "encoder.layers.29.feed_forward1.linear2.weight",
  "encoder.layers.29.feed_forward2.linear1.bias": "encoder.layers.29.feed_forward2.linear1.bias",
  "encoder.layers.29.feed_forward2.linear1.weight": "encoder.layers.29.feed_forward2.linear1.weight",
  "encoder.layers.29.feed_forward2.linear2.bias": "encoder.layers.29.feed_forward2.linear2.bias",
  "encoder.layers.29.feed_forward2.linear2.weight": "encoder.layers.29.feed_forward2.linear2.weight",
  "encoder.layers.29.norm_conv.bias": "encoder.layers.29.norm_conv.bias",
  "encoder.layers.29.norm_conv.weight": "encoder.layers.29.norm_conv.weight",
  "encoder.layers.29.norm_feed_forward1.bias": "encoder.layers.29.norm_feed_forward1.bias",
  "encoder.layers.29.norm_feed_forward1.weight": "encoder.layers.29.norm_feed_forward1.weight",
  "encoder.layers.29.norm_feed_forward2.bias": "encoder.layers.29.norm_feed_forward2.bias",
  "encoder.layers.29.norm_feed_forward2.weight": "encoder.layers.29.norm_feed_forward2.weight",
  "encoder.layers.29.norm_out.bias": "encoder.layers.29.norm_out.bias",
  "encoder.layers.29.norm_out.weight": "encoder.layers.29.norm_out.weight",
  "encoder.layers.29.norm_self_att.bias": "encoder.layers.29.norm_self_att.bias",
  "encoder.layers.29.norm_self_att.weight": "encoder.layers.29.norm_self_att.weight",
  "encoder.layers.29.self_attn.linear_out.bias": "encoder.layers.29.self_attn.out_proj.bias",
  "encoder.layers.29.self_attn.linear_out.weight": "encoder.layers.29.self_attn.out_proj.weight",
  "encoder.layers.29.self_attn.linear_pos.weight": "encoder.layers.29.self_attn.pos_proj.weight",
  "encoder.layers.29.self_attn.<qkv>.bias": "encoder.layers.29.self_attn.qkv_proj.bias",
  "encoder.layers.29.self_attn.<qkv>.weight": "encoder.layers.29.self_attn.qkv_proj.weight",
  "encoder.layers.29.self_attn.pos_bias_u": "encoder.layers.29.self_attn.pos_bias_u",
  "encoder.layers.29.self_attn.pos_bias_v": "encoder.layers.29.self_attn.pos_bias_v",
  "encoder.layers.3.conv.batch_norm.bias": "encoder.layers.3.conv.batch_norm.bias",
  "encoder.layers.3.conv.batch_norm.running_mean": "encoder.layers.3.conv.batch_norm.running_mean",
  "encoder.layers.3.conv.batch_norm.running_var": "encoder.layers.3.conv.batch_norm.running_var",
  "encoder.layers.3.conv.batch_norm.weight": "encoder.layers.3.conv.batch_norm.weight",
  "encoder.layers.3.conv.depthwise_conv.bias": "encoder.layers.3.conv.depthwise_conv.bias",
  "encoder.layers.3.conv.depthwise_conv.weight": "encoder.layers.3.conv.depthwise_conv.weight",
  "encoder.layers.3.conv.pointwise_conv1.bias": "encoder.layers.3.conv.pointwise_conv1.bias",
  "encoder.layers.3.conv.pointwise_conv1.weight": "encoder.layers.3.conv.pointwise_conv1.weight",
  "encoder.layers.3.conv.pointwise_conv2.bias": "encoder.layers.3.conv.pointwise_conv2.bias",
  "encoder.layers.3.conv.pointwise_conv2.weight": "encoder.layers.3.conv.pointwise_conv2.weight",
  "encoder.layers.3.feed_forward1.linear1.bias": "encoder.layers.3.feed_forward1.linear1.bias",
  "encoder.layers.3.feed_forward1.linear1.weight": "encoder.layers.3.feed_forward1.linear1.weight",
  "encoder.layers.3.feed_forward1.linear2.bias": "encoder.layers.3.feed_forward1.linear2.bias",
  "encoder.layers.3.feed_forward1.linear2.weight": "encoder.layers.3.feed_forward1.linear2.weight",
  "encoder.layers.3.feed_forward2.linear1.bias": "encoder.layers.3.feed_forward2.linear1.bias",
  "encoder.layers.3.feed_forward2.linear1.weight": "encoder.layers.3.feed_forward2.linear1.weight",
  "encoder.layers.3.feed_forward2.linear2.bias": "encoder.layers.3.feed_forward2.linear2.bias",
  "encoder.layers.3.feed_forward2.linear2.weight": "encoder.layers.3.feed_forward2.linear2.weight",
  "encoder.layers.3.norm_conv.bias": "encoder.layers.3.norm_conv.bias",
  "encoder.layers.3.norm_conv.weight": "encoder.layers.3.norm_conv.weight",
  "encoder.layers.3.norm_feed_forward1.bias": "encoder.layers.3.norm_feed_forward1.bias",
  "encoder.layers.3.norm_feed_forward1.weight": "encoder.layers.3.norm_feed_forward1.weight",
  "encoder.layers.3.norm_feed_forward2.bias": "encoder.layers.3.norm_feed_forward2.bias",
  "encoder.layers.3.norm_feed_forward2.weight": "encoder.layers.3.norm_feed_forward2.weight",
  "encoder.layers.3.norm_out.bias": "encoder.layers.3.norm_out.bias",
  "encoder.layers.3.norm_out.weight": "encoder.layers.3.norm_out.weight",
  "encoder.layers.3.norm_self_att.bias": "encoder.layers.3.norm_self_att.bias",
  "encoder.layers.3.norm_self_att.weight": "encoder.layers.3.norm_self_att.weight",
  "encoder.layers.3.self_attn.linear_out.bias": "encoder.layers.3.self_attn.out_proj.bias",
  "encoder.layers.3.self_attn.linear_out.weight": "encoder.layers.3.self_attn.out_proj.weight",
  "encoder.layers.3.self_attn.linear_pos.weight": "encoder.layers.3.self_attn.pos_proj.weight",
  "encoder.layers.3.self_attn.<qkv>.bias": "encoder.layers.3.self_attn.qkv_proj.bias",
  "encoder.layers.3.self_attn.<qkv>.weight": "encoder.layers.3.self_attn.qkv_proj.weight",
  "encoder.layers.3.self_attn.pos_bias_u": "encoder.layers.3.self_attn.pos_bias_u",
  "encoder.layers.3.self_attn.pos_bias_v": "encoder.layers.3.self_attn.pos_bias_v",
  "encoder.layers.30.conv.batch_norm.bias": "encoder.layers.30.conv.batch_norm.bias",
  "encoder.layers.30.conv.batch_norm.running_mean": "encoder.layers.30.conv.batch_norm.running_mean",
  "encoder.layers.30.conv.batch_norm.running_var": "encoder.layers.30.conv.batch_norm.running_var",
  "encoder.layers.30.conv.batch_norm.weight": "encoder.layers.30.conv.batch_norm.weight",
  "encoder.layers.30.conv.depthwise_conv.bias": "encoder.layers.30.conv.depthwise_conv.bias",
  "encoder.layers.30.conv.depthwise_conv.weight": "encoder.layers.30.conv.depthwise_conv.weight",
  "encoder.layers.30.conv.pointwise_conv1.bias": "encoder.layers.30.conv.pointwise_conv1.bias",
  "encoder.layers.30.conv.pointwise_conv1.weight": "encoder.layers.30.conv.pointwise_conv1.weight",
  "encoder.layers.30.conv.pointwise_conv2.bias": "encoder.layers.30.conv.pointwise_conv2.bias",
  "encoder.layers.30.conv.pointwise_conv2.weight": "encoder.layers.30.conv.pointwise_conv2.weight",
  "encoder.layers.30.feed_forward1.linear1.bias": "encoder.layers.30.feed_forward1.linear1.bias",
  "encoder.layers.30.feed_forward1.linear1.weight": "encoder.layers.30.feed_forward1.linear1.weight",
  "encoder.layers.30.feed_forward1.linear2.bias": "encoder.layers.30.feed_forward1.linear2.bias",
  "encoder.layers.30.feed_forward1.linear2.weight": "encoder.layers.30.feed_forward1.linear2.weight",
  "encoder.layers.30.feed_forward2.linear1.bias": "encoder.layers.30.feed_forward2.linear1.bias",
  "encoder.layers.30.feed_forward2.linear1.weight": "encoder.layers.30.feed_forward2.linear1.weight",
  "encoder.layers.30.feed_forward2.linear2.bias": "encoder.layers.30.feed_forward2.linear2.bias",
  "encoder.layers.30.feed_forward2.linear2.weight": "encoder.layers.30.feed_forward2.linear2.weight",
  "encoder.layers.30.norm_conv.bias": "encoder.layers.30.norm_conv.bias",
  "encoder.layers.30.norm_conv.weight": "encoder.layers.30.norm_conv.weight",
  "encoder.layers.30.norm_feed_forward1.bias": "encoder.layers.30.norm_feed_forward1.bias",
  "encoder.layers.30.norm_feed_forward1.weight": "encoder.layers.30.norm_feed_forward1.weight",
  "encoder.layers.30.norm_feed_forward2.bias": "encoder.layers.30.norm_feed_forward2.bias",
  "encoder.layers.30.norm_feed_forward2.weight": "encoder.layers.30.norm_feed_forward2.weight",
  "encoder.layers.30.norm_out.bias": "encoder.layers.30.norm_out.bias",
  "encoder.layers.30.norm_out.weight": "encoder.layers.30.norm_out.weight",
  "encoder.layers.30.norm_self_att.bias": "encoder.layers.30.norm_self_att.bias",
  "encoder.layers.30.norm_self_att.weight": "encoder.layers.30.norm_self_att.weight",
  "encoder.layers.30.self_attn.linear_out.bias": "encoder.layers.30.self_attn.out_proj.bias",
  "encoder.layers.30.self_attn.linear_out.weight": "encoder.layers.30.self_attn.out_proj.weight",
  "encoder.layers.30.self_attn.linear_pos.weight": "encoder.layers.30.self_attn.pos_proj.weight",
  "encoder.layers.30.self_attn.<qkv>.bias": "encoder.layers.30.self_attn.qkv_proj.bias",
  "encoder.layers.30.self_attn.<qkv>.weight": "encoder.layers.30.self_attn.qkv_proj.weight",
  "encoder.layers.30.self_attn.pos_bias_u": "encoder.layers.30.self_attn.pos_bias_u",
  "encoder.layers.30.self_attn.pos_bias_v": "encoder.layers.30.self_attn.pos_bias_v",
  "encoder.layers.31.conv.batch_norm.bias": "encoder.layers.31.conv.batch_norm.bias",
  "encoder.layers.31.conv.batch_norm.running_mean": "encoder.layers.31.conv.batch_norm.running_mean",
  "encoder.layers.31.conv.batch_norm.running_var": "encoder.layers.31.conv.batch_norm.running_var",
  "encoder.layers.31.conv.batch_norm.weight": "encoder.layers.31.conv.batch_norm.weight",
  "encoder.layers.31.conv.depthwise_conv.bias": "encoder.layers.31.conv.depthwise_conv.bias",
  "encoder.layers.31.conv.depthwise_conv.weight": "encoder.layers.31.conv.depthwise_conv.weight",
  "encoder.layers.31.conv.pointwise_conv1.bias": "encoder.layers.31.conv.pointwise_conv1.bias",
  "encoder.layers.31.conv.pointwise_conv1.weight": "encoder.layers.31.conv.pointwise_conv1.weight",
  "encoder.layers.31.conv.pointwise_conv2.bias": "encoder.layers.31.conv.pointwise_conv2.bias",
  "encoder.layers.31.conv.pointwise_conv2.weight": "encoder.layers.31.conv.pointwise_conv2.weight",
  "encoder.layers.31.feed_forward1.linear1.bias": "encoder.layers.31.feed_forward1.linear1.bias",
  "encoder.layers.31.feed_forward1.linear1.weight": "encoder.layers.31.feed_forward1.linear1.weight",
  "encoder.layers.31.feed_forward1.linear2.bias": "encoder.layers.31.feed_forward1.linear2.bias",
  "encoder.layers.31.feed_forward1.linear2.weight": "encoder.layers.31.feed_forward1.linear2.weight",
  "encoder.layers.31.feed_forward2.linear1.bias": "encoder.layers.31.feed_forward2.linear1.bias",
  "encoder.layers.31.feed_forward2.linear1.weight": "encoder.layers.31.feed_forward2.linear1.weight",
  "encoder.layers.31.feed_forward2.linear2.bias": "encoder.layers.31.feed_forward2.linear2.bias",
  "encoder.layers.31.feed_forward2.linear2.weight": "encoder.layers.31.feed_forward2.linear2.weight",
  "encoder.layers.31.norm_conv.bias": "encoder.layers.31.norm_conv.bias",
  "encoder.layers.31.norm_conv.weight": "encoder.layers.31.norm_conv.weight",
  "encoder.layers.31.norm_feed_forward1.bias": "encoder.layers.31.norm_feed_forward1.bias",
  "encoder.layers.31.norm_feed_forward1.weight": "encoder.layers.31.norm_feed_forward1.weight",
  "encoder.layers.31.norm_feed_forward2.bias": "encoder.layers.31.norm_feed_forward2.bias",
  "encoder.layers.31.norm_feed_forward2.weight": "encoder.layers.31.norm_feed_forward2.weight",
  "encoder.layers.31.norm_out.bias": "encoder.layers.31.norm_out.bias",
  "encoder.layers.31.norm_out.weight": "encoder.layers.31.norm_out.weight",
  "encoder.layers.31.norm_self_att.bias": "encoder.layers.31.norm_self_att.bias",
  "encoder.layers.31.norm_self_att.weight": "encoder.layers.31.norm_self_att.weight",
  "encoder.layers.31.self_attn.linear_out.bias": "encoder.layers.31.self_attn.out_proj.bias",
  "encoder.layers.31.self_attn.linear_out.weight": "encoder.layers.31.self_attn.out_proj.weight",
  "encoder.layers.31.self_attn.linear_pos.weight": "encoder.layers.31.self_attn.pos_proj.weight",
  "encoder.layers.31.self_attn.<qkv>.bias": "encoder.layers.31.self_attn.qkv_proj.bias",
  "encoder.layers.31.self_attn.<qkv>.weight": "encoder.layers.31.self_attn.qkv_proj.weight",
  "encoder.layers.31.self_attn.pos_bias_u": "encoder.layers.31.self_attn.pos_bias_u",
  "encoder.layers.31.self_attn.pos_bias_v": "encoder.layers.31.self_attn.pos_bias_v",
  "encoder.layers.32.conv.batch_norm.bias": "encoder.layers.32.conv.batch_norm.bias",
  "encoder.layers.32.conv.batch_norm.running_mean": "encoder.layers.32.conv.batch_norm.running_mean",
  "encoder.layers.32.conv.batch_norm.running_var": "encoder.layers.32.conv.batch_norm.running_var",
  "encoder.layers.32.conv.batch_norm.weight": "encoder.layers.32.conv.batch_norm.weight",
  "encoder.layers.32.conv.depthwise_conv.bias": "encoder.layers.32.conv.depthwise_conv.bias",
  "encoder.layers.32.conv.depthwise_conv.weight": "encoder.layers.32.conv.depthwise_conv.weight",
  "encoder.layers.32.conv.pointwise_conv1.bias": "encoder.layers.32.conv.pointwise_conv1.bias",
  "encoder.layers.32.conv.pointwise_conv1.weight": "encoder.layers.32.conv.pointwise_conv1.weight",
  "encoder.layers.32.conv.pointwise_conv2.bias": "encoder.layers.32.conv.pointwise_conv2.bias",
  "encoder.layers.32.conv.pointwise_conv2.weight": "encoder.layers.32.conv.pointwise_conv2.weight",
  "encoder.layers.32.feed_forward1.linear1.bias": "encoder.layers.32.feed_forward1.linear1.bias",
  "encoder.layers.32.feed_forward1.linear1.weight": "encoder.layers.32.feed_forward1.linear1.weight",
  "encoder.layers.32.feed_forward1.linear2.bias": "encoder.layers.32.feed_forward1.linear2.bias",
  "encoder.layers.32.feed_forward1.linear2.weight": "encoder.layers.32.feed_forward1.linear2.weight",
  "encoder.layers.32.feed_forward2.linear1.bias": "encoder.layers.32.feed_forward2.linear1.bias",
  "encoder.layers.32.feed_forward2.linear1.weight": "encoder.layers.32.feed_forward2.linear1.weight",
  "encoder.layers.32.feed_forward2.linear2.bias": "encoder.layers.32.feed_forward2.linear2.bias",
  "encoder.layers.32.feed_forward2.linear2.weight": "encoder.layers.32.feed_forward2.linear2.weight",
  "encoder.layers.32.norm_conv.bias": "encoder.layers.32.norm_conv.bias",
  "encoder.layers.32.norm_conv.weight": "encoder.layers.32.norm_conv.weight",
  "encoder.layers.32.norm_feed_forward1.bias": "encoder.layers.32.norm_feed_forward1.bias",
  "encoder.layers.32.norm_feed_forward1.weight": "encoder.layers.32.norm_feed_forward1.weight",
  "encoder.layers.32.norm_feed_forward2.bias": "encoder.layers.32.norm_feed_forward2.bias",
  "encoder.layers.32.norm_feed_forward2.weight": "encoder.layers.32.norm_feed_forward2.weight",
  "encoder.layers.32.norm_out.bias": "encoder.layers.32.norm_out.bias",
  "encoder.layers.32.norm_out.weight": "encoder.layers.32.norm_out.weight",
  "encoder.layers.32.norm_self_att.bias": "encoder.layers.32.norm_self_att.bias",
  "encoder.layers.32.norm_self_att.weight": "encoder.layers.32.norm_self_att.weight",
  "encoder.layers.32.self_attn.linear_out.bias": "encoder.layers.32.self_attn.out_proj.bias",
  "encoder.layers.32.self_attn.linear_out.weight": "encoder.layers.32.self_attn.out_proj.weight",
  "encoder.layers.32.self_attn.linear_pos.weight": "encoder.layers.32.self_attn.pos_proj.weight",
  "encoder.layers.32.self_attn.<qkv>.bias": "encoder.layers.32.self_attn.qkv_proj.bias",
  "encoder.layers.32.self_attn.<qkv>.weight": "encoder.layers.32.self_attn.qkv_proj.weight",
  "encoder.layers.32.self_attn.pos_bias_u": "encoder.layers.32.self_attn.pos_bias_u",
  "encoder.layers.32.self_attn.pos_bias_v": "encoder.layers.32.self_attn.pos_bias_v",
  "encoder.layers.33.conv.batch_norm.bias": "encoder.layers.33.conv.batch_norm.bias",
  "encoder.layers.33.conv.batch_norm.running_mean": "encoder.layers.33.conv.batch_norm.running_mean",
  "encoder.layers.33.conv.batch_norm.running_var": "encoder.layers.33.conv.batch_norm.running_var",
  "encoder.layers.33.conv.batch_norm.weight": "encoder.layers.33.conv.batch_norm.weight",
  "encoder.layers.33.conv.depthwise_conv.bias": "encoder.layers.33.conv.depthwise_conv.bias",
  "encoder.layers.33.conv.depthwise_conv.weight": "encoder.layers.33.conv.depthwise_conv.weight",
  "encoder.layers.33.conv.pointwise_conv1.bias": "encoder.layers.33.conv.pointwise_conv1.bias",
  "encoder.layers.33.conv.pointwise_conv1.weight": "encoder.layers.33.conv.pointwise_conv1.weight",
  "encoder.layers.33.conv.pointwise_conv2.bias": "encoder.layers.33.conv.pointwise_conv2.bias",
  "encoder.layers.33.conv.pointwise_conv2.weight": "encoder.layers.33.conv.pointwise_conv2.weight",
  "encoder.layers.33.feed_forward1.linear1.bias": "encoder.layers.33.feed_forward1.linear1.bias",
  "encoder.layers.33.feed_forward1.linear1.weight": "encoder.layers.33.feed_forward1.linear1.weight",
  "encoder.layers.33.feed_forward1.linear2.bias": "encoder.layers.33.feed_forward1.linear2.bias",
  "encoder.layers.33.feed_forward1.linear2.weight": "encoder.layers.33.feed_forward1.linear2.weight",
  "encoder.layers.33.feed_forward2.linear1.bias": "encoder.layers.33.feed_forward2.linear1.bias",
  "encoder.layers.33.feed_forward2.linear1.weight": "encoder.layers.33.feed_forward2.linear1.weight",
  "encoder.layers.33.feed_forward2.linear2.bias": "encoder.layers.33.feed_forward2.linear2.bias",
  "encoder.layers.33.feed_forward2.linear2.weight": "encoder.layers.33.feed_forward2.linear2.weight",
  "encoder.layers.33.norm_conv.bias": "encoder.layers.33.norm_conv.bias",
  "encoder.layers.33.norm_conv.weight": "encoder.layers.33.norm_conv.weight",
  "encoder.layers.33.norm_feed_forward1.bias": "encoder.layers.33.norm_feed_forward1.bias",
  "encoder.layers.33.norm_feed_forward1.weight": "encoder.layers.33.norm_feed_forward1.weight",
  "encoder.layers.33.norm_feed_forward2.bias": "encoder.layers.33.norm_feed_forward2.bias",
  "encoder.layers.33.norm_feed_forward2.weight": "encoder.layers.33.norm_feed_forward2.weight",
  "encoder.layers.33.norm_out.bias": "encoder.layers.33.norm_out.bias",
  "encoder.layers.33.norm_out.weight": "encoder.layers.33.norm_out.weight",
  "encoder.layers.33.norm_self_att.bias": "encoder.layers.33.norm_self_att.bias",
  "encoder.layers.33.norm_self_att.weight": "encoder.layers.33.norm_self_att.weight",
  "encoder.layers.33.self_attn.linear_out.bias": "encoder.layers.33.self_attn.out_proj.bias",
  "encoder.layers.33.self_attn.linear_out.weight": "encoder.layers.33.self_attn.out_proj.weight",
  "encoder.layers.33.self_attn.linear_pos.weight": "encoder.layers.33.self_attn.pos_proj.weight",
  "encoder.layers.33.self_attn.<qkv>.bias": "encoder.layers.33.self_attn.qkv_proj.bias",
  "encoder.layers.33.self_attn.<qkv>.weight": "encoder.layers.33.self_attn.qkv_proj.weight",
  "encoder.layers.33.self_attn.pos_bias_u": "encoder.layers.33.self_attn.pos_bias_u",
  "encoder.layers.33.self_attn.pos_bias_v": "encoder.layers.33.self_attn.pos_bias_v",
  "encoder.layers.34.conv.batch_norm.bias": "encoder.layers.34.conv.batch_norm.bias",
  "encoder.layers.34.conv.batch_norm.running_mean": "encoder.layers.34.conv.batch_norm.running_mean",
  "encoder.layers.34.conv.batch_norm.running_var": "encoder.layers.34.conv.batch_norm.running_var",
  "encoder.layers.34.conv.batch_norm.weight": "encoder.layers.34.conv.batch_norm.weight",
  "encoder.layers.34.conv.depthwise_conv.bias": "encoder.layers.34.conv.depthwise_conv.bias",
  "encoder.layers.34.conv.depthwise_conv.weight": "encoder.layers.34.conv.depthwise_conv.weight",
  "encoder.layers.34.conv.pointwise_conv1.bias": "encoder.layers.34.conv.pointwise_conv1.bias",
  "encoder.layers.34.conv.pointwise_conv1.weight": "encoder.layers.34.conv.pointwise_conv1.weight",
  "encoder.layers.34.conv.pointwise_conv2.bias": "encoder.layers.34.conv.pointwise_conv2.bias",
  "encoder.layers.34.conv.pointwise_conv2.weight": "encoder.layers.34.conv.pointwise_conv2.weight",
  "encoder.layers.34.feed_forward1.linear1.bias": "encoder.layers.34.feed_forward1.linear1.bias",
  "encoder.layers.34.feed_forward1.linear1.weight": "encoder.layers.34.feed_forward1.linear1.weight",
  "encoder.layers.34.feed_forward1.linear2.bias": "encoder.layers.34.feed_forward1.linear2.bias",
  "encoder.layers.34.feed_forward1.linear2.weight": "encoder.layers.34.feed_forward1.linear2.weight",
  "encoder.layers.34.feed_forward2.linear1.bias": "encoder.layers.34.feed_forward2.linear1.bias",
  "encoder.layers.34.feed_forward2.linear1.weight": "encoder.layers.34.feed_forward2.linear1.weight",
  "encoder.layers.34.feed_forward2.linear2.bias": "encoder.layers.34.feed_forward2.linear2.bias",
  "encoder.layers.34.feed_forward2.linear2.weight": "encoder.layers.34.feed_forward2.linear2.weight",
  "encoder.layers.34.norm_conv.bias": "encoder.layers.34.norm_conv.bias",
  "encoder.layers.34.norm_conv.weight": "encoder.layers.34.norm_conv.weight",
  "encoder.layers.34.norm_feed_forward1.bias": "encoder.layers.34.norm_feed_forward1.bias",
  "encoder.layers.34.norm_feed_forward1.weight": "encoder.layers.34.norm_feed_forward1.weight",
  "encoder.layers.34.norm_feed_forward2.bias": "encoder.layers.34.norm_feed_forward2.bias",
  "encoder.layers.34.norm_feed_forward2.weight": "encoder.layers.34.norm_feed_forward2.weight",
  "encoder.layers.34.norm_out.bias": "encoder.layers.34.norm_out.bias",
  "encoder.layers.34.norm_out.weight": "encoder.layers.34.norm_out.weight",
  "encoder.layers.34.norm_self_att.bias": "encoder.layers.34.norm_self_att.bias",
  "encoder.layers.34.norm_self_att.weight": "encoder.layers.34.norm_self_att.weight",
  "encoder.layers.34.self_attn.linear_out.bias": "encoder.layers.34.self_attn.out_proj.bias",
  "encoder.layers.34.self_attn.linear_out.weight": "encoder.layers.34.self_attn.out_proj.weight",
  "encoder.layers.34.self_attn.linear_pos.weight": "encoder.layers.34.self_attn.pos_proj.weight",
  "encoder.layers.34.self_attn.<qkv>.bias": "encoder.layers.34.self_attn.qkv_proj.bias",
  "encoder.layers.34.self_attn.<qkv>.weight": "encoder.layers.34.self_attn.qkv_proj.weight",
  "encoder.layers.34.self_attn.pos_bias_u": "encoder.layers.34.self_attn.pos_bias_u",
  "encoder.layers.34.self_attn.pos_bias_v": "encoder.layers.34.self_attn.pos_bias_v",
  "encoder.layers.35.conv.batch_norm.bias": "encoder.layers.35.conv.batch_norm.bias",
  "encoder.layers.35.conv.batch_norm.running_mean": "encoder.layers.35.conv.batch_norm.running_mean",
  "encoder.layers.35.conv.batch_norm.running_var": "encoder.layers.35.conv.batch_norm.running_var",
  "encoder.layers.35.conv.batch_norm.weight": "encoder.layers.35.conv.batch_norm.weight",
  "encoder.layers.35.conv.depthwise_conv.bias": "encoder.layers.35.conv.depthwise_conv.bias",
  "encoder.layers.35.conv.depthwise_conv.weight": "encoder.layers.35.conv.depthwise_conv.weight",
  "encoder.layers.35.conv.pointwise_conv1.bias": "encoder.layers.35.conv.pointwise_conv1.bias",
  "encoder.layers.35.conv.pointwise_conv1.weight": "encoder.layers.35.conv.pointwise_conv1.weight",
  "encoder.layers.35.conv.pointwise_conv2.bias": "encoder.layers.35.conv.pointwise_conv2.bias",
  "encoder.layers.35.conv.pointwise_conv2.weight": "encoder.layers.35.conv.pointwise_conv2.weight",
  "encoder.layers.35.feed_forward1.linear1.bias": "encoder.layers.35.feed_forward1.linear1.bias",
  "encoder.layers.35.feed_forward1.linear1.weight": "encoder.layers.35.feed_forward1.linear1.weight",
  "encoder.layers.35.feed_forward1.linear2.bias": "encoder.layers.35.feed_forward1.linear2.bias",
  "encoder.layers.35.feed_forward1.linear2.weight": "encoder.layers.35.feed_forward1.linear2.weight",
  "encoder.layers.35.feed_forward2.linear1.bias": "encoder.layers.35.feed_forward2.linear1.bias",
  "encoder.layers.35.feed_forward2.linear1.weight": "encoder.layers.35.feed_forward2.linear1.weight",
  "encoder.layers.35.feed_forward2.linear2.bias": "encoder.layers.35.feed_forward2.linear2.bias",
  "encoder.layers.35.feed_forward2.linear2.weight": "encoder.layers.35.feed_forward2.linear2.weight",
  "encoder.layers.35.norm_conv.bias": "encoder.layers.35.norm_conv.bias",
  "encoder.layers.35.norm_conv.weight": "encoder.layers.35.norm_conv.weight",
  "encoder.layers.35.norm_feed_forward1.bias": "encoder.layers.35.norm_feed_forward1.bias",
  "encoder.layers.35.norm_feed_forward1.weight": "encoder.layers.35.norm_feed_forward1.weight",
  "encoder.layers.35.norm_feed_forward2.bias": "encoder.layers.35.norm_feed_forward2.bias",
  "encoder.layers.35.norm_feed_forward2.weight": "encoder.layers.35.norm_feed_forward2.weight",
  "encoder.layers.35.norm_out.bias": "encoder.layers.35.norm_out.bias",
  "encoder.layers.35.norm_out.weight": "encoder.layers.35.norm_out.weight",
  "encoder.layers.35.norm_self_att.bias": "encoder.layers.35.norm_self_att.bias",
  "encoder.layers.35.norm_self_att.weight": "encoder.layers.35.norm_self_att.weight",
  "encoder.layers.35.self_attn.linear_out.bias": "encoder.layers.35.self_attn.out_proj.bias",
  "encoder.layers.35.self_attn.linear_out.weight": "encoder.layers.35.self_attn.out_proj.weight",
  "encoder.layers.35.self_attn.linear_pos.weight": "encoder.layers.35.self_attn.pos_proj.weight",
  "encoder.layers.35.self_attn.<qkv>.bias": "encoder.layers.35.self_attn.qkv_proj.bias",
  "encoder.layers.35.self_attn.<qkv>.weight": "encoder.layers.35.self_attn.qkv_proj.weight",
  "encoder.layers.35.self_attn.pos_bias_u": "encoder.layers.35.self_attn.pos_bias_u",
  "encoder.layers.35.self_attn.pos_bias_v": "encoder.layers.35.self_attn.pos_bias_v",
  "encoder.layers.36.conv.batch_norm.bias": "encoder.layers.36.conv.batch_norm.bias",
  "encoder.layers.36.conv.batch_norm.running_mean": "encoder.layers.36.conv.batch_norm.running_mean",
  "encoder.layers.36.conv.batch_norm.running_var": "encoder.layers.36.conv.batch_norm.running_var",
  "encoder.layers.36.conv.batch_norm.weight": "encoder.layers.36.conv.batch_norm.weight",
  "encoder.layers.36.conv.depthwise_conv.bias": "encoder.layers.36.conv.depthwise_conv.bias",
  "encoder.layers.36.conv.depthwise_conv.weight": "encoder.layers.36.conv.depthwise_conv.weight",
  "encoder.layers.36.conv.pointwise_conv1.bias": "encoder.layers.36.conv.pointwise_conv1.bias",
  "encoder.layers.36.conv.pointwise_conv1.weight": "encoder.layers.36.conv.pointwise_conv1.weight",
  "encoder.layers.36.conv.pointwise_conv2.bias": "encoder.layers.36.conv.pointwise_conv2.bias",
  "encoder.layers.36.conv.pointwise_conv2.weight": "encoder.layers.36.conv.pointwise_conv2.weight",
  "encoder.layers.36.feed_forward1.linear1.bias": "encoder.layers.36.feed_forward1.linear1.bias",
  "encoder.layers.36.feed_forward1.linear1.weight": "encoder.layers.36.feed_forward1.linear1.weight",
  "encoder.layers.36.feed_forward1.linear2.bias": "encoder.layers.36.feed_forward1.linear2.bias",
  "encoder.layers.36.feed_forward1.linear2.weight": "encoder.layers.36.feed_forward1.linear2.weight",
  "encoder.layers.36.feed_forward2.linear1.bias": "encoder.layers.36.feed_forward2.linear1.bias",
  "encoder.layers.36.feed_forward2.linear1.weight": "encoder.layers.36.feed_forward2.linear1.weight",
  "encoder.layers.36.feed_forward2.linear2.bias": "encoder.layers.36.feed_forward2.linear2.bias",
  "encoder.layers.36.feed_forward2.linear2.weight": "encoder.layers.36.feed_forward2.linear2.weight",
  "encoder.layers.36.norm_conv.bias": "encoder.layers.36.norm_conv.bias",
  "encoder.layers.36.norm_conv.weight": "encoder.layers.36.norm_conv.weight",
  "encoder.layers.36.norm_feed_forward1.bias": "encoder.layers.36.norm_feed_forward1.bias",
  "encoder.layers.36.norm_feed_forward1.weight": "encoder.layers.36.norm_feed_forward1.weight",
  "encoder.layers.36.norm_feed_forward2.bias": "encoder.layers.36.norm_feed_forward2.bias",
  "encoder.layers.36.norm_feed_forward2.weight": "encoder.layers.36.norm_feed_forward2.weight",
  "encoder.layers.36.norm_out.bias": "encoder.layers.36.norm_out.bias",
  "encoder.layers.36.norm_out.weight": "encoder.layers.36.norm_out.weight",
  "encoder.layers.36.norm_self_att.bias": "encoder.layers.36.norm_self_att.bias",
  "encoder.layers.36.norm_self_att.weight": "encoder.layers.36.norm_self_att.weight",
  "encoder.layers.36.self_attn.linear_out.bias": "encoder.layers.36.self_attn.out_proj.bias",
  "encoder.layers.36.self_attn.linear_out.weight": "encoder.layers.36.self_attn.out_proj.weight",
  "encoder.layers.36.self_attn.linear_pos.weight": "encoder.layers.36.self_attn.pos_proj.weight",
  "encoder.layers.36.self_attn.<qkv>.bias": "encoder.layers.36.self_attn.qkv_proj.bias",
  "encoder.layers.36.self_attn.<qkv>.weight": "encoder.layers.36.self_attn.qkv_proj.weight",
  "encoder.layers.36.self_attn.pos_bias_u": "encoder.layers.36.self_attn.pos_bias_u",
  "encoder.layers.36.self_attn.pos_bias_v": "encoder.layers.36.self_attn.pos_bias_v",
  "encoder.layers.37.conv.batch_norm.bias": "encoder.layers.37.conv.batch_norm.bias",
  "encoder.layers.37.conv.batch_norm.running_mean": "encoder.layers.37.conv.batch_norm.running_mean",
  "encoder.layers.37.conv.batch_norm.running_var": "encoder.layers.37.conv.batch_norm.running_var",
  "encoder.layers.37.conv.batch_norm.weight": "encoder.layers.37.conv.batch_norm.weight",
  "encoder.layers.37.conv.depthwise_conv.bias": "encoder.layers.37.conv.depthwise_conv.bias",
  "encoder.layers.37.conv.depthwise_conv.weight": "encoder.layers.37.conv.depthwise_conv.weight",
  "encoder.layers.37.conv.pointwise_conv1.bias": "encoder.layers.37.conv.pointwise_conv1.bias",
  "encoder.layers.37.conv.pointwise_conv1.weight": "encoder.layers.37.conv.pointwise_conv1.weight",
  "encoder.layers.37.conv.pointwise_conv2.bias": "encoder.layers.37.conv.pointwise_conv2.bias",
  "encoder.layers.37.conv.pointwise_conv2.weight": "encoder.layers.37.conv.pointwise_conv2.weight",
  "encoder.layers.37.feed_forward1.linear1.bias": "encoder.layers.37.feed_forward1.linear1.bias",
  "encoder.layers.37.feed_forward1.linear1.weight": "encoder.layers.37.feed_forward1.linear1.weight",
  "encoder.layers.37.feed_forward1.linear2.bias": "encoder.layers.37.feed_forward1.linear2.bias",
  "encoder.layers.37.feed_forward1.linear2.weight": "encoder.layers.37.feed_forward1.linear2.weight",
  "encoder.layers.37.feed_forward2.linear1.bias": "encoder.layers.37.feed_forward2.linear1.bias",
  "encoder.layers.37.feed_forward2.linear1.weight": "encoder.layers.37.feed_forward2.linear1.weight",
  "encoder.layers.37.feed_forward2.linear2.bias": "encoder.layers.37.feed_forward2.linear2.bias",
  "encoder.layers.37.feed_forward2.linear2.weight": "encoder.layers.37.feed_forward2.linear2.weight",
  "encoder.layers.37.norm_conv.bias": "encoder.layers.37.norm_conv.bias",
  "encoder.layers.37.norm_conv.weight": "encoder.layers.37.norm_conv.weight",
  "encoder.layers.37.norm_feed_forward1.bias": "encoder.layers.37.norm_feed_forward1.bias",
  "encoder.layers.37.norm_feed_forward1.weight": "encoder.layers.37.norm_feed_forward1.weight",
  "encoder.layers.37.norm_feed_forward2.bias": "encoder.layers.37.norm_feed_forward2.bias",
  "encoder.layers.37.norm_feed_forward2.weight": "encoder.layers.37.norm_feed_forward2.weight",
  "encoder.layers.37.norm_out.bias": "encoder.layers.37.norm_out.bias",
  "encoder.layers.37.norm_out.weight": "encoder.layers.37.norm_out.weight",
  "encoder.layers.37.norm_self_att.bias": "encoder.layers.37.norm_self_att.bias",
  "encoder.layers.37.norm_self_att.weight": "encoder.layers.37.norm_self_att.weight",
  "encoder.layers.37.self_attn.linear_out.bias": "encoder.layers.37.self_attn.out_proj.bias",
  "encoder.layers.37.self_attn.linear_out.weight": "encoder.layers.37.self_attn.out_proj.weight",
  "encoder.layers.37.self_attn.linear_pos.weight": "encoder.layers.37.self_attn.pos_proj.weight",
  "encoder.layers.37.self_attn.<qkv>.bias": "encoder.layers.37.self_attn.qkv_proj.bias",
  "encoder.layers.37.self_attn.<qkv>.weight": "encoder.layers.37.self_attn.qkv_proj.weight",
  "encoder.layers.37.self_attn.pos_bias_u": "encoder.layers.37.self_attn.pos_bias_u",
  "encoder.layers.37.self_attn.pos_bias_v": "encoder.layers.37.self_attn.pos_bias_v",
  "encoder.layers.38.conv.batch_norm.bias": "encoder.layers.38.conv.batch_norm.bias",
  "encoder.layers.38.conv.batch_norm.running_mean": "encoder.layers.38.conv.batch_norm.running_mean",
  "encoder.layers.38.conv.batch_norm.running_var": "encoder.layers.38.conv.batch_norm.running_var",
  "encoder.layers.38.conv.batch_norm.weight": "encoder.layers.38.conv.batch_norm.weight",
  "encoder.layers.38.conv.depthwise_conv.bias": "encoder.layers.38.conv.depthwise_conv.bias",
  "encoder.layers.38.conv.depthwise_conv.weight": "encoder.layers.38.conv.depthwise_conv.weight",
  "encoder.layers.38.conv.pointwise_conv1.bias": "encoder.layers.38.conv.pointwise_conv1.bias",
  "encoder.layers.38.conv.pointwise_conv1.weight": "encoder.layers.38.conv.pointwise_conv1.weight",
  "encoder.layers.38.conv.pointwise_conv2.bias": "encoder.layers.38.conv.pointwise_conv2.bias",
  "encoder.layers.38.conv.pointwise_conv2.weight": "encoder.layers.38.conv.pointwise_conv2.weight",
  "encoder.layers.38.feed_forward1.linear1.bias": "encoder.layers.38.feed_forward1.linear1.bias",
  "encoder.layers.38.feed_forward1.linear1.weight": "encoder.layers.38.feed_forward1.linear1.weight",
  "encoder.layers.38.feed_forward1.linear2.bias": "encoder.layers.38.feed_forward1.linear2.bias",
  "encoder.layers.38.feed_forward1.linear2.weight": "encoder.layers.38.feed_forward1.linear2.weight",
  "encoder.layers.38.feed_forward2.linear1.bias": "encoder.layers.38.feed_forward2.linear1.bias",
  "encoder.layers.38.feed_forward2.linear1.weight": "encoder.layers.38.feed_forward2.linear1.weight",
  "encoder.layers.38.feed_forward2.linear2.bias": "encoder.layers.38.feed_forward2.linear2.bias",
  "encoder.layers.38.feed_forward2.linear2.weight": "encoder.layers.38.feed_forward2.linear2.weight",
  "encoder.layers.38.norm_conv.bias": "encoder.layers.38.norm_conv.bias",
  "encoder.layers.38.norm_conv.weight": "encoder.layers.38.norm_conv.weight",
  "encoder.layers.38.norm_feed_forward1.bias": "encoder.layers.38.norm_feed_forward1.bias",
  "encoder.layers.38.norm_feed_forward1.weight": "encoder.layers.38.norm_feed_forward1.weight",
  "encoder.layers.38.norm_feed_forward2.bias": "encoder.layers.38.norm_feed_forward2.bias",
  "encoder.layers.38.norm_feed_forward2.weight": "encoder.layers.38.norm_feed_forward2.weight",
  "encoder.layers.38.norm_out.bias": "encoder.layers.38.norm_out.bias",
  "encoder.layers.38.norm_out.weight": "encoder.layers.38.norm_out.weight",
  "encoder.layers.38.norm_self_att.bias": "encoder.layers.38.norm_self_att.bias",
  "encoder.layers.38.norm_self_att.weight": "encoder.layers.38.norm_self_att.weight",
  "encoder.layers.38.self_attn.linear_out.bias": "encoder.layers.38.self_attn.out_proj.bias",
  "encoder.layers.38.self_attn.linear_out.weight": "encoder.layers.38.self_attn.out_proj.weight",
  "encoder.layers.38.self_attn.linear_pos.weight": "encoder.layers.38.self_attn.pos_proj.weight",
  "encoder.layers.38.self_attn.<qkv>.bias": "encoder.layers.38.self_attn.qkv_proj.bias",
  "encoder.layers.38.self_attn.<qkv>.weight": "encoder.layers.38.self_attn.qkv_proj.weight",
  "encoder.layers.38.self_attn.pos_bias_u": "encoder.layers.38.self_attn.pos_bias_u",
  "encoder.layers.38.self_attn.pos_bias_v": "encoder.layers.38.self_attn.pos_bias_v",
  "encoder.layers.39.conv.batch_norm.bias": "encoder.layers.39.conv.batch_norm.bias",
  "encoder.layers.39.conv.batch_norm.running_mean": "encoder.layers.39.conv.batch_norm.running_mean",
  "encoder.layers.39.conv.batch_norm.running_var": "encoder.layers.39.conv.batch_norm.running_var",
  "encoder.layers.39.conv.batch_norm.weight": "encoder.layers.39.conv.batch_norm.weight",
  "encoder.layers.39.conv.depthwise_conv.bias": "encoder.layers.39.conv.depthwise_conv.bias",
  "encoder.layers.39.conv.depthwise_conv.weight": "encoder.layers.39.conv.depthwise_conv.weight",
  "encoder.layers.39.conv.pointwise_conv1.bias": "encoder.layers.39.conv.pointwise_conv1.bias",
  "encoder.layers.39.conv.pointwise_conv1.weight": "encoder.layers.39.conv.pointwise_conv1.weight",
  "encoder.layers.39.conv.pointwise_conv2.bias": "encoder.layers.39.conv.pointwise_conv2.bias",
  "encoder.layers.39.conv.pointwise_conv2.weight": "encoder.layers.39.conv.pointwise_conv2.weight",
  "encoder.layers.39.feed_forward1.linear1.bias": "encoder.layers.39.feed_forward1.linear1.bias",
  "encoder.layers.39.feed_forward1.linear1.weight": "encoder.layers.39.feed_forward1.linear1.weight",
  "encoder.layers.39.feed_forward1.linear2.bias": "encoder.layers.39.feed_forward1.linear2.bias",
  "encoder.layers.39.feed_forward1.linear2.weight": "encoder.layers.39.feed_forward1.linear2.weight",
  "encoder.layers.39.feed_forward2.linear1.bias": "encoder.layers.39.feed_forward2.linear1.bias",
  "encoder.layers.39.feed_forward2.linear1.weight": "encoder.layers.39.feed_forward2.linear1.weight",
  "encoder.layers.39.feed_forward2.linear2.bias": "encoder.layers.39.feed_forward2.linear2.bias",
  "encoder.layers.39.feed_forward2.linear2.weight": "encoder.layers.39.feed_forward2.linear2.weight",
  "encoder.layers.39.norm_conv.bias": "encoder.layers.39.norm_conv.bias",
  "encoder.layers.39.norm_conv.weight": "encoder.layers.39.norm_conv.weight",
  "encoder.layers.39.norm_feed_forward1.bias": "encoder.layers.39.norm_feed_forward1.bias",
  "encoder.layers.39.norm_feed_forward1.weight": "encoder.layers.39.norm_feed_forward1.weight",
  "encoder.layers.39.norm_feed_forward2.bias": "encoder.layers.39.norm_feed_forward2.bias",
  "encoder.layers.39.norm_feed_forward2.weight": "encoder.layers.39.norm_feed_forward2.weight",
  "encoder.layers.39.norm_out.bias": "encoder.layers.39.norm_out.bias",
  "encoder.layers.39.norm_out.weight": "encoder.layers.39.norm_out.weight",
  "encoder.layers.39.norm_self_att.bias": "encoder.layers.39.norm_self_att.bias",
  "encoder.layers.39.norm_self_att.weight": "encoder.layers.39.norm_self_att.weight",
  "encoder.layers.39.self_attn.linear_out.bias": "encoder.layers.39.self_attn.out_proj.bias",
  "encoder.layers.39.self_attn.linear_out.weight": "encoder.layers.39.self_attn.out_proj.weight",
  "encoder.layers.39.self_attn.linear_pos.weight": "encoder.layers.39.self_attn.pos_proj.weight",
  "encoder.layers.39.self_attn.<qkv>.bias": "encoder.layers.39.self_attn.qkv_proj.bias",
  "encoder.layers.39.self_attn.<qkv>.weight": "encoder.layers.39.self_attn.qkv_proj.weight",
  "encoder.layers.39.self_attn.pos_bias_u": "encoder.layers.39.self_attn.pos_bias_u",
  "encoder.layers.39.self_attn.pos_bias_v": "encoder.layers.39.self_attn.pos_bias_v",
  "encoder.layers.4.conv.batch_norm.bias": "encoder.layers.4.conv.batch_norm.bias",
  "encoder.layers.4.conv.batch_norm.running_mean": "encoder.layers.4.conv.batch_norm.running_mean",
  "encoder.layers.4.conv.batch_norm.running_var": "encoder.layers.4.conv.batch_norm.running_var",
  "encoder.layers.4.conv.batch_norm.weight": "encoder.layers.4.conv.batch_norm.weight",
  "encoder.layers.4.conv.depthwise_conv.bias": "encoder.layers.4.conv.depthwise_conv.bias",
  "encoder.layers.4.conv.depthwise_conv.weight": "encoder.layers.4.conv.depthwise_conv.weight",
  "encoder.layers.4.conv.pointwise_conv1.bias": "encoder.layers.4.conv.pointwise_conv1.bias",
  "encoder.layers.4.conv.pointwise_conv1.weight": "encoder.layers.4.conv.pointwise_conv1.weight",
  "encoder.layers.4.conv.pointwise_conv2.bias": "encoder.layers.4.conv.pointwise_conv2.bias",
  "encoder.layers.4.conv.pointwise_conv2.weight": "encoder.layers.4.conv.pointwise_conv2.weight",
  "encoder.layers.4.feed_forward1.linear1.bias": "encoder.layers.4.feed_forward1.linear1.bias",
  "encoder.layers.4.feed_forward1.linear1.weight": "encoder.layers.4.feed_forward1.linear1.weight",
  "encoder.layers.4.feed_forward1.linear2.bias": "encoder.layers.4.feed_forward1.linear2.bias",
  "encoder.layers.4.feed_forward1.linear2.weight": "encoder.layers.4.feed_forward1.linear2.weight",
  "encoder.layers.4.feed_forward2.linear1.bias": "encoder.layers.4.feed_forward2.linear1.bias",
  "encoder.layers.4.feed_forward2.linear1.weight": "encoder.layers.4.feed_forward2.linear1.weight",
  "encoder.layers.4.feed_forward2.linear2.bias": "encoder.layers.4.feed_forward2.linear2.bias",
  "encoder.layers.4.feed_forward2.linear2.weight": "encoder.layers.4.feed_forward2.linear2.weight",
  "encoder.layers.4.norm_conv.bias": "encoder.layers.4.norm_conv.bias",
  "encoder.layers.4.norm_conv.weight": "encoder.layers.4.norm_conv.weight",
  "encoder.layers.4.norm_feed_forward1.bias": "encoder.layers.4.norm_feed_forward1.bias",
  "encoder.layers.4.norm_feed_forward1.weight": "encoder.layers.4.norm_feed_forward1.weight",
  "encoder.layers.4.norm_feed_forward2.bias": "encoder.layers.4.norm_feed_forward2.bias",
  "encoder.layers.4.norm_feed_forward2.weight": "encoder.layers.4.norm_feed_forward2.weight",
  "encoder.layers.4.norm_out.bias": "encoder.layers.4.norm_out.bias",
  "encoder.layers.4.norm_out.weight": "encoder.layers.4.norm_out.weight",
  "encoder.layers.4.norm_self_att.bias": "encoder.layers.4.norm_self_att.bias",
  "encoder.layers.4.norm_self_att.weight": "encoder.layers.4.norm_self_att.weight",
  "encoder.layers.4.self_attn.linear_out.bias": "encoder.layers.4.self_attn.out_proj.bias",
  "encoder.layers.4.self_attn.linear_out.weight": "encoder.layers.4.self_attn.out_proj.weight",
  "encoder.layers.4.self_attn.linear_pos.weight": "encoder.layers.4.self_attn.pos_proj.weight",
  "encoder.layers.4.self_attn.<qkv>.bias": "encoder.layers.4.self_attn.qkv_proj.bias",
  "encoder.layers.4.self_attn.<qkv>.weight": "encoder.layers.4.self_attn.qkv_proj.weight",
  "encoder.layers.4.self_attn.pos_bias_u": "encoder.layers.4.self_attn.pos_bias_u",
  "encoder.layers.4.self_attn.pos_bias_v": "encoder.layers.4.self_attn.pos_bias_v",
  "encoder.layers.40.conv.batch_norm.bias": "encoder.layers.40.conv.batch_norm.bias",
  "encoder.layers.40.conv.batch_norm.running_mean": "encoder.layers.40.conv.batch_norm.running_mean",
  "encoder.layers.40.conv.batch_norm.running_var": "encoder.layers.40.conv.batch_norm.running_var",
  "encoder.layers.40.conv.batch_norm.weight": "encoder.layers.40.conv.batch_norm.weight",
  "encoder.layers.40.conv.depthwise_conv.bias": "encoder.layers.40.conv.depthwise_conv.bias",
  "encoder.layers.40.conv.depthwise_conv.weight": "encoder.layers.40.conv.depthwise_conv.weight",
  "encoder.layers.40.conv.pointwise_conv1.bias": "encoder.layers.40.conv.pointwise_conv1.bias",
  "encoder.layers.40.conv.pointwise_conv1.weight": "encoder.layers.40.conv.pointwise_conv1.weight",
  "encoder.layers.40.conv.pointwise_conv2.bias": "encoder.layers.40.conv.pointwise_conv2.bias",
  "encoder.layers.40.conv.pointwise_conv2.weight": "encoder.layers.40.conv.pointwise_conv2.weight",
  "encoder.layers.40.feed_forward1.linear1.bias": "encoder.layers.40.feed_forward1.linear1.bias",
  "encoder.layers.40.feed_forward1.linear1.weight": "encoder.layers.40.feed_forward1.linear1.weight",
  "encoder.layers.40.feed_forward1.linear2.bias": "encoder.layers.40.feed_forward1.linear2.bias",
  "encoder.layers.40.feed_forward1.linear2.weight": "encoder.layers.40.feed_forward1.linear2.weight",
  "encoder.layers.40.feed_forward2.linear1.bias": "encoder.layers.40.feed_forward2.linear1.bias",
  "encoder.layers.40.feed_forward2.linear1.weight": "encoder.layers.40.feed_forward2.linear1.weight",
  "encoder.layers.40.feed_forward2.linear2.bias": "encoder.layers.40.feed_forward2.linear2.bias",
  "encoder.layers.40.feed_forward2.linear2.weight": "encoder.layers.40.feed_forward2.linear2.weight",
  "encoder.layers.40.norm_conv.bias": "encoder.layers.40.norm_conv.bias",
  "encoder.layers.40.norm_conv.weight": "encoder.layers.40.norm_conv.weight",
  "encoder.layers.40.norm_feed_forward1.bias": "encoder.layers.40.norm_feed_forward1.bias",
  "encoder.layers.40.norm_feed_forward1.weight": "encoder.layers.40.norm_feed_forward1.weight",
  "encoder.layers.40.norm_feed_forward2.bias": "encoder.layers.40.norm_feed_forward2.bias",
  "encoder.layers.40.norm_feed_forward2.weight": "encoder.layers.40.norm_feed_forward2.weight",
  "encoder.layers.40.norm_out.bias": "encoder.layers.40.norm_out.bias",
  "encoder.layers.40.norm_out.weight": "encoder.layers.40.norm_out.weight",
  "encoder.layers.40.norm_self_att.bias": "encoder.layers.40.norm_self_att.bias",
  "encoder.layers.40.norm_self_att.weight": "encoder.layers.40.norm_self_att.weight",
  "encoder.layers.40.self_attn.linear_out.bias": "encoder.layers.40.self_attn.out_proj.bias",
  "encoder.layers.40.self_attn.linear_out.weight": "encoder.layers.40.self_attn.out_proj.weight",
  "encoder.layers.40.self_attn.linear_pos.weight": "encoder.layers.40.self_attn.pos_proj.weight",
  "encoder.layers.40.self_attn.<qkv>.bias": "encoder.layers.40.self_attn.qkv_proj.bias",
  "encoder.layers.40.self_attn.<qkv>.weight": "encoder.layers.40.self_attn.qkv_proj.weight",
  "encoder.layers.40.self_attn.pos_bias_u": "encoder.layers.40.self_attn.pos_bias_u",
  "encoder.layers.40.self_attn.pos_bias_v": "encoder.layers.40.self_attn.pos_bias_v",
  "encoder.layers.41.conv.batch_norm.bias": "encoder.layers.41.conv.batch_norm.bias",
  "encoder.layers.41.conv.batch_norm.running_mean": "encoder.layers.41.conv.batch_norm.running_mean",
  "encoder.layers.41.conv.batch_norm.running_var": "encoder.layers.41.conv.batch_norm.running_var",
  "encoder.layers.41.conv.batch_norm.weight": "encoder.layers.41.conv.batch_norm.weight",
  "encoder.layers.41.conv.depthwise_conv.bias": "encoder.layers.41.conv.depthwise_conv.bias",
  "encoder.layers.41.conv.depthwise_conv.weight": "encoder.layers.41.conv.depthwise_conv.weight",
  "encoder.layers.41.conv.pointwise_conv1.bias": "encoder.layers.41.conv.pointwise_conv1.bias",
  "encoder.layers.41.conv.pointwise_conv1.weight": "encoder.layers.41.conv.pointwise_conv1.weight",
  "encoder.layers.41.conv.pointwise_conv2.bias": "encoder.layers.41.conv.pointwise_conv2.bias",
  "encoder.layers.41.conv.pointwise_conv2.weight": "encoder.layers.41.conv.pointwise_conv2.weight",
  "encoder.layers.41.feed_forward1.linear1.bias": "encoder.layers.41.feed_forward1.linear1.bias",
  "encoder.layers.41.feed_forward1.linear1.weight": "encoder.layers.41.feed_forward1.linear1.weight",
  "encoder.layers.41.feed_forward1.linear2.bias": "encoder.layers.41.feed_forward1.linear2.bias",
  "encoder.layers.41.feed_forward1.linear2.weight": "encoder.layers.41.feed_forward1.linear2.weight",
  "encoder.layers.41.feed_forward2.linear1.bias": "encoder.layers.41.feed_forward2.linear1.bias",
  "encoder.layers.41.feed_forward2.linear1.weight": "encoder.layers.41.feed_forward2.linear1.weight",
  "encoder.layers.41.feed_forward2.linear2.bias": "encoder.layers.41.feed_forward2.linear2.bias",
  "encoder.layers.41.feed_forward2.linear2.weight": "encoder.layers.41.feed_forward2.linear2.weight",
  "encoder.layers.41.norm_conv.bias": "encoder.layers.41.norm_conv.bias",
  "encoder.layers.41.norm_conv.weight": "encoder.layers.41.norm_conv.weight",
  "encoder.layers.41.norm_feed_forward1.bias": "encoder.layers.41.norm_feed_forward1.bias",
  "encoder.layers.41.norm_feed_forward1.weight": "encoder.layers.41.norm_feed_forward1.weight",
  "encoder.layers.41.norm_feed_forward2.bias": "encoder.layers.41.norm_feed_forward2.bias",
  "encoder.layers.41.norm_feed_forward2.weight": "encoder.layers.41.norm_feed_forward2.weight",
  "encoder.layers.41.norm_out.bias": "encoder.layers.41.norm_out.bias",
  "encoder.layers.41.norm_out.weight": "encoder.layers.41.norm_out.weight",
  "encoder.layers.41.norm_self_att.bias": "encoder.layers.41.norm_self_att.bias",
  "encoder.layers.41.norm_self_att.weight": "encoder.layers.41.norm_self_att.weight",
  "encoder.layers.41.self_attn.linear_out.bias": "encoder.layers.41.self_attn.out_proj.bias",
  "encoder.layers.41.self_attn.linear_out.weight": "encoder.layers.41.self_attn.out_proj.weight",
  "encoder.layers.41.self_attn.linear_pos.weight": "encoder.layers.41.self_attn.pos_proj.weight",
  "encoder.layers.41.self_attn.<qkv>.bias": "encoder.layers.41.self_attn.qkv_proj.bias",
  "encoder.layers.41.self_attn.<qkv>.weight": "encoder.layers.41.self_attn.qkv_proj.weight",
  "encoder.layers.41.self_attn.pos_bias_u": "encoder.layers.41.self_attn.pos_bias_u",
  "encoder.layers.41.self_attn.pos_bias_v": "encoder.layers.41.self_attn.pos_bias_v",
  "encoder.layers.42.conv.batch_norm.bias": "encoder.layers.42.conv.batch_norm.bias",
  "encoder.layers.42.conv.batch_norm.running_mean": "encoder.layers.42.conv.batch_norm.running_mean",
  "encoder.layers.42.conv.batch_norm.running_var": "encoder.layers.42.conv.batch_norm.running_var",
  "encoder.layers.42.conv.batch_norm.weight": "encoder.layers.42.conv.batch_norm.weight",
  "encoder.layers.42.conv.depthwise_conv.bias": "encoder.layers.42.conv.depthwise_conv.bias",
  "encoder.layers.42.conv.depthwise_conv.weight": "encoder.layers.42.conv.depthwise_conv.weight",
  "encoder.layers.42.conv.pointwise_conv1.bias": "encoder.layers.42.conv.pointwise_conv1.bias",
  "encoder.layers.42.conv.pointwise_conv1.weight": "encoder.layers.42.conv.pointwise_conv1.weight",
  "encoder.layers.42.conv.pointwise_conv2.bias": "encoder.layers.42.conv.pointwise_conv2.bias",
  "encoder.layers.42.conv.pointwise_conv2.weight": "encoder.layers.42.conv.pointwise_conv2.weight",
  "encoder.layers.42.feed_forward1.linear1.bias": "encoder.layers.42.feed_forward1.linear1.bias",
  "encoder.layers.42.feed_forward1.linear1.weight": "encoder.layers.42.feed_forward1.linear1.weight",
  "encoder.layers.42.feed_forward1.linear2.bias": "encoder.layers.42.feed_forward1.linear2.bias",
  "encoder.layers.42.feed_forward1.linear2.weight": "encoder.layers.42.feed_forward1.linear2.weight",
  "encoder.layers.42.feed_forward2.linear1.bias": "encoder.layers.42.feed_forward2.linear1.bias",
  "encoder.layers.42.feed_forward2.linear1.weight": "encoder.layers.42.feed_forward2.linear1.weight",
  "encoder.layers.42.feed_forward2.linear2.bias": "encoder.layers.42.feed_forward2.linear2.bias",
  "encoder.layers.42.feed_forward2.linear2.weight": "encoder.layers.42.feed_forward2.linear2.weight",
  "encoder.layers.42.norm_conv.bias": "encoder.layers.42.norm_conv.bias",
  "encoder.layers.42.norm_conv.weight": "encoder.layers.42.norm_conv.weight",
  "encoder.layers.42.norm_feed_forward1.bias": "encoder.layers.42.norm_feed_forward1.bias",
  "encoder.layers.42.norm_feed_forward1.weight": "encoder.layers.42.norm_feed_forward1.weight",
  "encoder.layers.42.norm_feed_forward2.bias": "encoder.layers.42.norm_feed_forward2.bias",
  "encoder.layers.42.norm_feed_forward2.weight": "encoder.layers.42.norm_feed_forward2.weight",
  "encoder.layers.42.norm_out.bias": "encoder.layers.42.norm_out.bias",
  "encoder.layers.42.norm_out.weight": "encoder.layers.42.norm_out.weight",
  "encoder.layers.42.norm_self_att.bias": "encoder.layers.42.norm_self_att.bias",
  "encoder.layers.42.norm_self_att.weight": "encoder.layers.42.norm_self_att.weight",
  "encoder.layers.42.self_attn.linear_out.bias": "encoder.layers.42.self_attn.out_proj.bias",
  "encoder.layers.42.self_attn.linear_out.weight": "encoder.layers.42.self_attn.out_proj.weight",
  "encoder.layers.42.self_attn.linear_pos.weight": "encoder.layers.42.self_attn.pos_proj.weight",
  "encoder.layers.42.self_attn.<qkv>.bias": "encoder.layers.42.self_attn.qkv_proj.bias",
  "encoder.layers.42.self_attn.<qkv>.weight": "encoder.layers.42.self_attn.qkv_proj.weight",
  "encoder.layers.42.self_attn.pos_bias_u": "encoder.layers.42.self_attn.pos_bias_u",
  "encoder.layers.42.self_attn.pos_bias_v": "encoder.layers.42.self_attn.pos_bias_v",
  "encoder.layers.43.conv.batch_norm.bias": "encoder.layers.43.conv.batch_norm.bias",
  "encoder.layers.43.conv.batch_norm.running_mean": "encoder.layers.43.conv.batch_norm.running_mean",
  "encoder.layers.43.conv.batch_norm.running_var": "encoder.layers.43.conv.batch_norm.running_var",
  "encoder.layers.43.conv.batch_norm.weight": "encoder.layers.43.conv.batch_norm.weight",
  "encoder.layers.43.conv.depthwise_conv.bias": "encoder.layers.43.conv.depthwise_conv.bias",
  "encoder.layers.43.conv.depthwise_conv.weight": "encoder.layers.43.conv.depthwise_conv.weight",
  "encoder.layers.43.conv.pointwise_conv1.bias": "encoder.layers.43.conv.pointwise_conv1.bias",
  "encoder.layers.43.conv.pointwise_conv1.weight": "encoder.layers.43.conv.pointwise_conv1.weight",
  "encoder.layers.43.conv.pointwise_conv2.bias": "encoder.layers.43.conv.pointwise_conv2.bias",
  "encoder.layers.43.conv.pointwise_conv2.weight": "encoder.layers.43.conv.pointwise_conv2.weight",
  "encoder.layers.43.feed_forward1.linear1.bias": "encoder.layers.43.feed_forward1.linear1.bias",
  "encoder.layers.43.feed_forward1.linear1.weight": "encoder.layers.43.feed_forward1.linear1.weight",
  "encoder.layers.43.feed_forward1.linear2.bias": "encoder.layers.43.feed_forward1.linear2.bias",
  "encoder.layers.43.feed_forward1.linear2.weight": "encoder.layers.43.feed_forward1.linear2.weight",
  "encoder.layers.43.feed_forward2.linear1.bias": "encoder.layers.43.feed_forward2.linear1.bias",
  "encoder.layers.43.feed_forward2.linear1.weight": "encoder.layers.43.feed_forward2.linear1.weight",
  "encoder.layers.43.feed_forward2.linear2.bias": "encoder.layers.43.feed_forward2.linear2.bias",
  "encoder.layers.43.feed_forward2.linear2.weight": "encoder.layers.43.feed_forward2.linear2.weight",
  "encoder.layers.43.norm_conv.bias": "encoder.layers.43.norm_conv.bias",
  "encoder.layers.43.norm_conv.weight": "encoder.layers.43.norm_conv.weight",
  "encoder.layers.43.norm_feed_forward1.bias": "encoder.layers.43.norm_feed_forward1.bias",
  "encoder.layers.43.norm_feed_forward1.weight": "encoder.layers.43.norm_feed_forward1.weight",
  "encoder.layers.43.norm_feed_forward2.bias": "encoder.layers.43.norm_feed_forward2.bias",
  "encoder.layers.43.norm_feed_forward2.weight": "encoder.layers.43.norm_feed_forward2.weight",
  "encoder.layers.43.norm_out.bias": "encoder.layers.43.norm_out.bias",
  "encoder.layers.43.norm_out.weight": "encoder.layers.43.norm_out.weight",
  "encoder.layers.43.norm_self_att.bias": "encoder.layers.43.norm_self_att.bias",
  "encoder.layers.43.norm_self_att.weight": "encoder.layers.43.norm_self_att.weight",
  "encoder.layers.43.self_attn.linear_out.bias": "encoder.layers.43.self_attn.out_proj.bias",
  "encoder.layers.43.self_attn.linear_out.weight": "encoder.layers.43.self_attn.out_proj.weight",
  "encoder.layers.43.self_attn.linear_pos.weight": "encoder.layers.43.self_attn.pos_proj.weight",
  "encoder.layers.43.self_attn.<qkv>.bias": "encoder.layers.43.self_attn.qkv_proj.bias",
  "encoder.layers.43.self_attn.<qkv>.weight": "encoder.layers.43.self_attn.qkv_proj.weight",
  "encoder.layers.43.self_attn.pos_bias_u": "encoder.layers.43.self_attn.pos_bias_u",
  "encoder.layers.43.self_attn.pos_bias_v": "encoder.layers.43.self_attn.pos_bias_v",
  "encoder.layers.44.conv.batch_norm.bias": "encoder.layers.44.conv.batch_norm.bias",
  "encoder.layers.44.conv.batch_norm.running_mean": "encoder.layers.44.conv.batch_norm.running_mean",
  "encoder.layers.44.conv.batch_norm.running_var": "encoder.layers.44.conv.batch_norm.running_var",
  "encoder.layers.44.conv.batch_norm.weight": "encoder.layers.44.conv.batch_norm.weight",
  "encoder.layers.44.conv.depthwise_conv.bias": "encoder.layers.44.conv.depthwise_conv.bias",
  "encoder.layers.44.conv.depthwise_conv.weight": "encoder.layers.44.conv.depthwise_conv.weight",
  "encoder.layers.44.conv.pointwise_conv1.bias": "encoder.layers.44.conv.pointwise_conv1.bias",
  "encoder.layers.44.conv.pointwise_conv1.weight": "encoder.layers.44.conv.pointwise_conv1.weight",
  "encoder.layers.44.conv.pointwise_conv2.bias": "encoder.layers.44.conv.pointwise_conv2.bias",
  "encoder.layers.44.conv.pointwise_conv2.weight": "encoder.layers.44.conv.pointwise_conv2.weight",
  "encoder.layers.44.feed_forward1.linear1.bias": "encoder.layers.44.feed_forward1.linear1.bias",
  "encoder.layers.44.feed_forward1.linear1.weight": "encoder.layers.44.feed_forward1.linear1.weight",
  "encoder.layers.44.feed_forward1.linear2.bias": "encoder.layers.44.feed_forward1.linear2.bias",
  "encoder.layers.44.feed_forward1.linear2.weight": "encoder.layers.44.feed_forward1.linear2.weight",
  "encoder.layers.44.feed_forward2.linear1.bias": "encoder.layers.44.feed_forward2.linear1.bias",
  "encoder.layers.44.feed_forward2.linear1.weight": "encoder.layers.44.feed_forward2.linear1.weight",
  "encoder.layers.44.feed_forward2.linear2.bias": "encoder.layers.44.feed_forward2.linear2.bias",
  "encoder.layers.44.feed_forward2.linear2.weight": "encoder.layers.44.feed_forward2.linear2.weight",
  "encoder.layers.44.norm_conv.bias": "encoder.layers.44.norm_conv.bias",
  "encoder.layers.44.norm_conv.weight": "encoder.layers.44.norm_conv.weight",
  "encoder.layers.44.norm_feed_forward1.bias": "encoder.layers.44.norm_feed_forward1.bias",
  "encoder.layers.44.norm_feed_forward1.weight": "encoder.layers.44.norm_feed_forward1.weight",
  "encoder.layers.44.norm_feed_forward2.bias": "encoder.layers.44.norm_feed_forward2.bias",
  "encoder.layers.44.norm_feed_forward2.weight": "encoder.layers.44.norm_feed_forward2.weight",
  "encoder.layers.44.norm_out.bias": "encoder.layers.44.norm_out.bias",
  "encoder.layers.44.norm_out.weight": "encoder.layers.44.norm_out.weight",
  "encoder.layers.44.norm_self_att.bias": "encoder.layers.44.norm_self_att.bias",
  "encoder.layers.44.norm_self_att.weight": "encoder.layers.44.norm_self_att.weight",
  "encoder.layers.44.self_attn.linear_out.bias": "encoder.layers.44.self_attn.out_proj.bias",
  "encoder.layers.44.self_attn.linear_out.weight": "encoder.layers.44.self_attn.out_proj.weight",
  "encoder.layers.44.self_attn.linear_pos.weight": "encoder.layers.44.self_attn.pos_proj.weight",
  "encoder.layers.44.self_attn.<qkv>.bias": "encoder.layers.44.self_attn.qkv_proj.bias",
  "encoder.layers.44.self_attn.<qkv>.weight": "encoder.layers.44.self_attn.qkv_proj.weight",
  "encoder.layers.44.self_attn.pos_bias_u": "encoder.layers.44.self_attn.pos_bias_u",
  "encoder.layers.44.self_attn.pos_bias_v": "encoder.layers.44.self_attn.pos_bias_v",
  "encoder.layers.45.conv.batch_norm.bias": "encoder.layers.45.conv.batch_norm.bias",
  "encoder.layers.45.conv.batch_norm.running_mean": "encoder.layers.45.conv.batch_norm.running_mean",
  "encoder.layers.45.conv.batch_norm.running_var": "encoder.layers.45.conv.batch_norm.running_var",
  "encoder.layers.45.conv.batch_norm.weight": "encoder.layers.45.conv.batch_norm.weight",
  "encoder.layers.45.conv.depthwise_conv.bias": "encoder.layers.45.conv.depthwise_conv.bias",
  "encoder.layers.45.conv.depthwise_conv.weight": "encoder.layers.45.conv.depthwise_conv.weight",
  "encoder.layers.45.conv.pointwise_conv1.bias": "encoder.layers.45.conv.pointwise_conv1.bias",
  "encoder.layers.45.conv.pointwise_conv1.weight": "encoder.layers.45.conv.pointwise_conv1.weight",
  "encoder.layers.45.conv.pointwise_conv2.bias": "encoder.layers.45.conv.pointwise_conv2.bias",
  "encoder.layers.45.conv.pointwise_conv2.weight": "encoder.layers.45.conv.pointwise_conv2.weight",
  "encoder.layers.45.feed_forward1.linear1.bias": "encoder.layers.45.feed_forward1.linear1.bias",
  "encoder.layers.45.feed_forward1.linear1.weight": "encoder.layers.45.feed_forward1.linear1.weight",
  "encoder.layers.45.feed_forward1.linear2.bias": "encoder.layers.45.feed_forward1.linear2.bias",
  "encoder.layers.45.feed_forward1.linear2.weight": "encoder.layers.45.feed_forward1.linear2.weight",
  "encoder.layers.45.feed_forward2.linear1.bias": "encoder.layers.45.feed_forward2.linear1.bias",
  "encoder.layers.45.feed_forward2.linear1.weight": "encoder.layers.45.feed_forward2.linear1.weight",
  "encoder.layers.45.feed_forward2.linear2.bias": "encoder.layers.45.feed_forward2.linear2.bias",
  "encoder.layers.45.feed_forward2.linear2.weight": "encoder.layers.45.feed_forward2.linear2.weight",
  "encoder.layers.45.norm_conv.bias": "encoder.layers.45.norm_conv.bias",
  "encoder.layers.45.norm_conv.weight": "encoder.layers.45.norm_conv.weight",
  "encoder.layers.45.norm_feed_forward1.bias": "encoder.layers.45.norm_feed_forward1.bias",
  "encoder.layers.45.norm_feed_forward1.weight": "encoder.layers.45.norm_feed_forward1.weight",
  "encoder.layers.45.norm_feed_forward2.bias": "encoder.layers.45.norm_feed_forward2.bias",
  "encoder.layers.45.norm_feed_forward2.weight": "encoder.layers.45.norm_feed_forward2.weight",
  "encoder.layers.45.norm_out.bias": "encoder.layers.45.norm_out.bias",
  "encoder.layers.45.norm_out.weight": "encoder.layers.45.norm_out.weight",
  "encoder.layers.45.norm_self_att.bias": "encoder.layers.45.norm_self_att.bias",
  "encoder.layers.45.norm_self_att.weight": "encoder.layers.45.norm_self_att.weight",
  "encoder.layers.45.self_attn.linear_out.bias": "encoder.layers.45.self_attn.out_proj.bias",
  "encoder.layers.45.self_attn.linear_out.weight": "encoder.layers.45.self_attn.out_proj.weight",
  "encoder.layers.45.self_attn.linear_pos.weight": "encoder.layers.45.self_attn.pos_proj.weight",
  "encoder.layers.45.self_attn.<qkv>.bias": "encoder.layers.45.self_attn.qkv_proj.bias",
  "encoder.layers.45.self_attn.<qkv>.weight": "encoder.layers.45.self_attn.qkv_proj.weight",
  "encoder.layers.45.self_attn.pos_bias_u": "encoder.layers.45.self_attn.pos_bias_u",
  "encoder.layers.45.self_attn.pos_bias_v": "encoder.layers.45.self_attn.pos_bias_v",
  "encoder.layers.46.conv.batch_norm.bias": "encoder.layers.46.conv.batch_norm.bias",
  "encoder.layers.46.conv.batch_norm.running_mean": "encoder.layers.46.conv.batch_norm.running_mean",
  "encoder.layers.46.conv.batch_norm.running_var": "encoder.layers.46.conv.batch_norm.running_var",
  "encoder.layers.46.conv.batch_norm.weight": "encoder.layers.46.conv.batch_norm.weight",
  "encoder.layers.46.conv.depthwise_conv.bias": "encoder.layers.46.conv.depthwise_conv.bias",
  "encoder.layers.46.conv.depthwise_conv.weight": "encoder.layers.46.conv.depthwise_conv.weight",
  "encoder.layers.46.conv.pointwise_conv1.bias": "encoder.layers.46.conv.pointwise_conv1.bias",
  "encoder.layers.46.conv.pointwise_conv1.weight": "encoder.layers.46.conv.pointwise_conv1.weight",
  "encoder.layers.46.conv.pointwise_conv2.bias": "encoder.layers.46.conv.pointwise_conv2.bias",
  "encoder.layers.46.conv.pointwise_conv2.weight": "encoder.layers.46.conv.pointwise_conv2.weight",
  "encoder.layers.46.feed_forward1.linear1.bias": "encoder.layers.46.feed_forward1.linear1.bias",
  "encoder.layers.46.feed_forward1.linear1.weight": "encoder.layers.46.feed_forward1.linear1.weight",
  "encoder.layers.46.feed_forward1.linear2.bias": "encoder.layers.46.feed_forward1.linear2.bias",
  "encoder.layers.46.feed_forward1.linear2.weight": "encoder.layers.46.feed_forward1.linear2.weight",
  "encoder.layers.46.feed_forward2.linear1.bias": "encoder.layers.46.feed_forward2.linear1.bias",
  "encoder.layers.46.feed_forward2.linear1.weight": "encoder.layers.46.feed_forward2.linear1.weight",
  "encoder.layers.46.feed_forward2.linear2.bias": "encoder.layers.46.feed_forward2.linear2.bias",
  "encoder.layers.46.feed_forward2.linear2.weight": "encoder.layers.46.feed_forward2.linear2.weight",
  "encoder.layers.46.norm_conv.bias": "encoder.layers.46.norm_conv.bias",
  "encoder.layers.46.norm_conv.weight": "encoder.layers.46.norm_conv.weight",
  "encoder.layers.46.norm_feed_forward1.bias": "encoder.layers.46.norm_feed_forward1.bias",
  "encoder.layers.46.norm_feed_forward1.weight": "encoder.layers.46.norm_feed_forward1.weight",
  "encoder.layers.46.norm_feed_forward2.bias": "encoder.layers.46.norm_feed_forward2.bias",
  "encoder.layers.46.norm_feed_forward2.weight": "encoder.layers.46.norm_feed_forward2.weight",
  "encoder.layers.46.norm_out.bias": "encoder.layers.46.norm_out.bias",
  "encoder.layers.46.norm_out.weight": "encoder.layers.46.norm_out.weight",
  "encoder.layers.46.norm_self_att.bias": "encoder.layers.46.norm_self_att.bias",
  "encoder.layers.46.norm_self_att.weight": "encoder.layers.46.norm_self_att.weight",
  "encoder.layers.46.self_attn.linear_out.bias": "encoder.layers.46.self_attn.out_proj.bias",
  "encoder.layers.46.self_attn.linear_out.weight": "encoder.layers.46.self_attn.out_proj.weight",
  "encoder.layers.46.self_attn.linear_pos.weight": "encoder.layers.46.self_attn.pos_proj.weight",
  "encoder.layers.46.self_attn.<qkv>.bias": "encoder.layers.46.self_attn.qkv_proj.bias",
  "encoder.layers.46.self_attn.<qkv>.weight": "encoder.layers.46.self_attn.qkv_proj.weight",
  "encoder.layers.46.self_attn.pos_bias_u": "encoder.layers.46.self_attn.pos_bias_u",
  "encoder.layers.46.self_attn.pos_bias_v": "encoder.layers.46.self_attn.pos_bias_v",
  "encoder.layers.47.conv.batch_norm.bias": "encoder.layers.47.conv.batch_norm.bias",
  "encoder.layers.47.conv.batch_norm.running_mean": "encoder.layers.47.conv.batch_norm.running_mean",
  "encoder.layers.47.conv.batch_norm.running_var": "encoder.layers.47.conv.batch_norm.running_var",
  "encoder.layers.47.conv.batch_norm.weight": "encoder.layers.47.conv.batch_norm.weight",
  "encoder.layers.47.conv.depthwise_conv.bias": "encoder.layers.47.conv.depthwise_conv.bias",
  "encoder.layers.47.conv.depthwise_conv.weight": "encoder.layers.47.conv.depthwise_conv.weight",
  "encoder.layers.47.conv.pointwise_conv1.bias": "encoder.layers.47.conv.pointwise_conv1.bias",
  "encoder.layers.47.conv.pointwise_conv1.weight": "encoder.layers.47.conv.pointwise_conv1.weight",
  "encoder.layers.47.conv.pointwise_conv2.bias": "encoder.layers.47.conv.pointwise_conv2.bias",
  "encoder.layers.47.conv.pointwise_conv2.weight": "encoder.layers.47.conv.pointwise_conv2.weight",
  "encoder.layers.47.feed_forward1.linear1.bias": "encoder.layers.47.feed_forward1.linear1.bias",
  "encoder.layers.47.feed_forward1.linear1.weight": "encoder.layers.47.feed_forward1.linear1.weight",
  "encoder.layers.47.feed_forward1.linear2.bias": "encoder.layers.47.feed_forward1.linear2.bias",
  "encoder.layers.47.feed_forward1.linear2.weight": "encoder.layers.47.feed_forward1.linear2.weight",
  "encoder.layers.47.feed_forward2.linear1.bias": "encoder.layers.47.feed_forward2.linear1.bias",
  "encoder.layers.47.feed_forward2.linear1.weight": "encoder.layers.47.feed_forward2.linear1.weight",
  "encoder.layers.47.feed_forward2.linear2.bias": "encoder.layers.47.feed_forward2.linear2.bias",
  "encoder.layers.47.feed_forward2.linear2.weight": "encoder.layers.47.feed_forward2.linear2.weight",
  "encoder.layers.47.norm_conv.bias": "encoder.layers.47.norm_conv.bias",
  "encoder.layers.47.norm_conv.weight": "encoder.layers.47.norm_conv.weight",
  "encoder.layers.47.norm_feed_forward1.bias": "encoder.layers.47.norm_feed_forward1.bias",
  "encoder.layers.47.norm_feed_forward1.weight": "encoder.layers.47.norm_feed_forward1.weight",
  "encoder.layers.47.norm_feed_forward2.bias": "encoder.layers.47.norm_feed_forward2.bias",
  "encoder.layers.47.norm_feed_forward2.weight": "encoder.layers.47.norm_feed_forward2.weight",
  "encoder.layers.47.norm_out.bias": "encoder.layers.47.norm_out.bias",
  "encoder.layers.47.norm_out.weight": "encoder.layers.47.norm_out.weight",
  "encoder.layers.47.norm_self_att.bias": "encoder.layers.47.norm_self_att.bias",
  "encoder.layers.47.norm_self_att.weight": "encoder.layers.47.norm_self_att.weight",
  "encoder.layers.47.self_attn.linear_out.bias": "encoder.layers.47.self_attn.out_proj.bias",
  "encoder.layers.47.self_attn.linear_out.weight": "encoder.layers.47.self_attn.out_proj.weight",
  "encoder.layers.47.self_attn.linear_pos.weight": "encoder.layers.47.self_attn.pos_proj.weight",
  "encoder.layers.47.self_attn.<qkv>.bias": "encoder.layers.47.self_attn.qkv_proj.bias",
  "encoder.layers.47.self_attn.<qkv>.weight": "encoder.layers.47.self_attn.qkv_proj.weight",
  "encoder.layers.47.self_attn.pos_bias_u": "encoder.layers.47.self_attn.pos_bias_u",
  "encoder.layers.47.self_attn.pos_bias_v": "encoder.layers.47.self_attn.pos_bias_v",
  "encoder.layers.5.conv.batch_norm.bias": "encoder.layers.5.conv.batch_norm.bias",
  "encoder.layers.5.conv.batch_norm.running_mean": "encoder.layers.5.conv.batch_norm.running_mean",
  "encoder.layers.5.conv.batch_norm.running_var": "encoder.layers.5.conv.batch_norm.running_var",
  "encoder.layers.5.conv.batch_norm.weight": "encoder.layers.5.conv.batch_norm.weight",
  "encoder.layers.5.conv.depthwise_conv.bias": "encoder.layers.5.conv.depthwise_conv.bias",
  "encoder.layers.5.conv.depthwise_conv.weight": "encoder.layers.5.conv.depthwise_conv.weight",
  "encoder.layers.5.conv.pointwise_conv1.bias": "encoder.layers.5.conv.pointwise_conv1.bias",
  "encoder.layers.5.conv.pointwise_conv1.weight": "encoder.layers.5.conv.pointwise_conv1.weight",
  "encoder.layers.5.conv.pointwise_conv2.bias": "encoder.layers.5.conv.pointwise_conv2.bias",
  "encoder.layers.5.conv.pointwise_conv2.weight": "encoder.layers.5.conv.pointwise_conv2.weight",
  "encoder.layers.5.feed_forward1.linear1.bias": "encoder.layers.5.feed_forward1.linear1.bias",
  "encoder.layers.5.feed_forward1.linear1.weight": "encoder.layers.5.feed_forward1.linear1.weight",
  "encoder.layers.5.feed_forward1.linear2.bias": "encoder.layers.5.feed_forward1.linear2.bias",
  "encoder.layers.5.feed_forward1.linear2.weight": "encoder.layers.5.feed_forward1.linear2.weight",
  "encoder.layers.5.feed_forward2.linear1.bias": "encoder.layers.5.feed_forward2.linear1.bias",
  "encoder.layers.5.feed_forward2.linear1.weight": "encoder.layers.5.feed_forward2.linear1.weight",
  "encoder.layers.5.feed_forward2.linear2.bias": "encoder.layers.5.feed_forward2.linear2.bias",
  "encoder.layers.5.feed_forward2.linear2.weight": "encoder.layers.5.feed_forward2.linear2.weight",
  "encoder.layers.5.norm_conv.bias": "encoder.layers.5.norm_conv.bias",
  "encoder.layers.5.norm_conv.weight": "encoder.layers.5.norm_conv.weight",
  "encoder.layers.5.norm_feed_forward1.bias": "encoder.layers.5.norm_feed_forward1.bias",
  "encoder.layers.5.norm_feed_forward1.weight": "encoder.layers.5.norm_feed_forward1.weight",
  "encoder.layers.5.norm_feed_forward2.bias": "encoder.layers.5.norm_feed_forward2.bias",
  "encoder.layers.5.norm_feed_forward2.weight": "encoder.layers.5.norm_feed_forward2.weight",
  "encoder.layers.5.norm_out.bias": "encoder.layers.5.norm_out.bias",
  "encoder.layers.5.norm_out.weight": "encoder.layers.5.norm_out.weight",
  "encoder.layers.5.norm_self_att.bias": "encoder.layers.5.norm_self_att.bias",
  "encoder.layers.5.norm_self_att.weight": "encoder.layers.5.norm_self_att.weight",
  "encoder.layers.5.self_attn.linear_out.bias": "encoder.layers.5.self_attn.out_proj.bias",
  "encoder.layers.5.self_attn.linear_out.weight": "encoder.layers.5.self_attn.out_proj.weight",
  "encoder.layers.5.self_attn.linear_pos.weight": "encoder.layers.5.self_attn.pos_proj.weight",
  "encoder.layers.5.self_attn.<qkv>.bias": "encoder.layers.5.self_attn.qkv_proj.bias",
  "encoder.layers.5.self_attn.<qkv>.weight": "encoder.layers.5.self_attn.qkv_proj.weight",
  "encoder.layers.5.self_attn.pos_bias_u": "encoder.layers.5.self_attn.pos_bias_u",
  "encoder.layers.5.self_attn.pos_bias_v": "encoder.layers.5.self_attn.pos_bias_v",
  "encoder.layers.6.conv.batch_norm.bias": "encoder.layers.6.conv.batch_norm.bias",
  "encoder.layers.6.conv.batch_norm.running_mean": "encoder.layers.6.conv.batch_norm.running_mean",
  "encoder.layers.6.conv.batch_norm.running_var": "encoder.layers.6.conv.batch_norm.running_var",
  "encoder.layers.6.conv.batch_norm.weight": "encoder.layers.6.conv.batch_norm.weight",
  "encoder.layers.6.conv.depthwise_conv.bias": "encoder.layers.6.conv.depthwise_conv.bias",
  "encoder.layers.6.conv.depthwise_conv.weight": "encoder.layers.6.conv.depthwise_conv.weight",
  "encoder.layers.6.conv.pointwise_conv1.bias": "encoder.layers.6.conv.pointwise_conv1.bias",
  "encoder.layers.6.conv.pointwise_conv1.weight": "encoder.layers.6.conv.pointwise_conv1.weight",
  "encoder.layers.6.conv.pointwise_conv2.bias": "encoder.layers.6.conv.pointwise_conv2.bias",
  "encoder.layers.6.conv.pointwise_conv2.weight": "encoder.layers.6.conv.pointwise_conv2.weight",
  "encoder.layers.6.feed_forward1.linear1.bias": "encoder.layers.6.feed_forward1.linear1.bias",
  "encoder.layers.6.feed_forward1.linear1.weight": "encoder.layers.6.feed_forward1.linear1.weight",
  "encoder.layers.6.feed_forward1.linear2.bias": "encoder.layers.6.feed_forward1.linear2.bias",
  "encoder.layers.6.feed_forward1.linear2.weight": "encoder.layers.6.feed_forward1.linear2.weight",
  "encoder.layers.6.feed_forward2.linear1.bias": "encoder.layers.6.feed_forward2.linear1.bias",
  "encoder.layers.6.feed_forward2.linear1.weight": "encoder.layers.6.feed_forward2.linear1.weight",
  "encoder.layers.6.feed_forward2.linear2.bias": "encoder.layers.6.feed_forward2.linear2.bias",
  "encoder.layers.6.feed_forward2.linear2.weight": "encoder.layers.6.feed_forward2.linear2.weight",
  "encoder.layers.6.norm_conv.bias": "encoder.layers.6.norm_conv.bias",
  "encoder.layers.6.norm_conv.weight": "encoder.layers.6.norm_conv.weight",
  "encoder.layers.6.norm_feed_forward1.bias": "encoder.layers.6.norm_feed_forward1.bias",
  "encoder.layers.6.norm_feed_forward1.weight": "encoder.layers.6.norm_feed_forward1.weight",
  "encoder.layers.6.norm_feed_forward2.bias": "encoder.layers.6.norm_feed_forward2.bias",
  "encoder.layers.6.norm_feed_forward2.weight": "encoder.layers.6.norm_feed_forward2.weight",
  "encoder.layers.6.norm_out.bias": "encoder.layers.6.norm_out.bias",
  "encoder.layers.6.norm_out.weight": "encoder.layers.6.norm_out.weight",
  "encoder.layers.6.norm_self_att.bias": "encoder.layers.6.norm_self_att.bias",
  "encoder.layers.6.norm_self_att.weight": "encoder.layers.6.norm_self_att.weight",
  "encoder.layers.6.self_attn.linear_out.bias": "encoder.layers.6.self_attn.out_proj.bias",
  "encoder.layers.6.self_attn.linear_out.weight": "encoder.layers.6.self_attn.out_proj.weight",
  "encoder.layers.6.self_attn.linear_pos.weight": "encoder.layers.6.self_attn.pos_proj.weight",
  "encoder.layers.6.self_attn.<qkv>.bias": "encoder.layers.6.self_attn.qkv_proj.bias",
  "encoder.layers.6.self_attn.<qkv>.weight": "encoder.layers.6.self_attn.qkv_proj.weight",
  "encoder.layers.6.self_attn.pos_bias_u": "encoder.layers.6.self_attn.pos_bias_u",
  "encoder.layers.6.self_attn.pos_bias_v": "encoder.layers.6.self_attn.pos_bias_v",
  "encoder.layers.7.conv.batch_norm.bias": "encoder.layers.7.conv.batch_norm.bias",
  "encoder.layers.7.conv.batch_norm.running_mean": "encoder.layers.7.conv.batch_norm.running_mean",
  "encoder.layers.7.conv.batch_norm.running_var": "encoder.layers.7.conv.batch_norm.running_var",
  "encoder.layers.7.conv.batch_norm.weight": "encoder.layers.7.conv.batch_norm.weight",
  "encoder.layers.7.conv.depthwise_conv.bias": "encoder.layers.7.conv.depthwise_conv.bias",
  "encoder.layers.7.conv.depthwise_conv.weight": "encoder.layers.7.conv.depthwise_conv.weight",
  "encoder.layers.7.conv.pointwise_conv1.bias": "encoder.layers.7.conv.pointwise_conv1.bias",
  "encoder.layers.7.conv.pointwise_conv1.weight": "encoder.layers.7.conv.pointwise_conv1.weight",
  "encoder.layers.7.conv.pointwise_conv2.bias": "encoder.layers.7.conv.pointwise_conv2.bias",
  "encoder.layers.7.conv.pointwise_conv2.weight": "encoder.layers.7.conv.pointwise_conv2.weight",
  "encoder.layers.7.feed_forward1.linear1.bias": "encoder.layers.7.feed_forward1.linear1.bias",
  "encoder.layers.7.feed_forward1.linear1.weight": "encoder.layers.7.feed_forward1.linear1.weight",
  "encoder.layers.7.feed_forward1.linear2.bias": "encoder.layers.7.feed_forward1.linear2.bias",
  "encoder.layers.7.feed_forward1.linear2.weight": "encoder.layers.7.feed_forward1.linear2.weight",
  "encoder.layers.7.feed_forward2.linear1.bias": "encoder.layers.7.feed_forward2.linear1.bias",
  "encoder.layers.7.feed_forward2.linear1.weight": "encoder.layers.7.feed_forward2.linear1.weight",
  "encoder.layers.7.feed_forward2.linear2.bias": "encoder.layers.7.feed_forward2.linear2.bias",
  "encoder.layers.7.feed_forward2.linear2.weight": "encoder.layers.7.feed_forward2.linear2.weight",
  "encoder.layers.7.norm_conv.bias": "encoder.layers.7.norm_conv.bias",
  "encoder.layers.7.norm_conv.weight": "encoder.layers.7.norm_conv.weight",
  "encoder.layers.7.norm_feed_forward1.bias": "encoder.layers.7.norm_feed_forward1.bias",
  "encoder.layers.7.norm_feed_forward1.weight": "encoder.layers.7.norm_feed_forward1.weight",
  "encoder.layers.7.norm_feed_forward2.bias": "encoder.layers.7.norm_feed_forward2.bias",
  "encoder.layers.7.norm_feed_forward2.weight": "encoder.layers.7.norm_feed_forward2.weight",
  "encoder.layers.7.norm_out.bias": "encoder.layers.7.norm_out.bias",
  "encoder.layers.7.norm_out.weight": "encoder.layers.7.norm_out.weight",
  "encoder.layers.7.norm_self_att.bias": "encoder.layers.7.norm_self_att.bias",
  "encoder.layers.7.norm_self_att.weight": "encoder.layers.7.norm_self_att.weight",
  "encoder.layers.7.self_attn.linear_out.bias": "encoder.layers.7.self_attn.out_proj.bias",
  "encoder.layers.7.self_attn.linear_out.weight": "encoder.layers.7.self_attn.out_proj.weight",
  "encoder.layers.7.self_attn.linear_pos.weight": "encoder.layers.7.self_attn.pos_proj.weight",
  "encoder.layers.7.self_attn.<qkv>.bias": "encoder.layers.7.self_attn.qkv_proj.bias",
  "encoder.layers.7.self_attn.<qkv>.weight": "encoder.layers.7.self_attn.qkv_proj.weight",
  "encoder.layers.7.self_attn.pos_bias_u": "encoder.layers.7.self_attn.pos_bias_u",
  "encoder.layers.7.self_attn.pos_bias_v": "encoder.layers.7.self_attn.pos_bias_v",
  "encoder.layers.8.conv.batch_norm.bias": "encoder.layers.8.conv.batch_norm.bias",
  "encoder.layers.8.conv.batch_norm.running_mean": "encoder.layers.8.conv.batch_norm.running_mean",
  "encoder.layers.8.conv.batch_norm.running_var": "encoder.layers.8.conv.batch_norm.running_var",
  "encoder.layers.8.conv.batch_norm.weight": "encoder.layers.8.conv.batch_norm.weight",
  "encoder.layers.8.conv.depthwise_conv.bias": "encoder.layers.8.conv.depthwise_conv.bias",
  "encoder.layers.8.conv.depthwise_conv.weight": "encoder.layers.8.conv.depthwise_conv.weight",
  "encoder.layers.8.conv.pointwise_conv1.bias": "encoder.layers.8.conv.pointwise_conv1.bias",
  "encoder.layers.8.conv.pointwise_conv1.weight": "encoder.layers.8.conv.pointwise_conv1.weight",
  "encoder.layers.8.conv.pointwise_conv2.bias": "encoder.layers.8.conv.pointwise_conv2.bias",
  "encoder.layers.8.conv.pointwise_conv2.weight": "encoder.layers.8.conv.pointwise_conv2.weight",
  "encoder.layers.8.feed_forward1.linear1.bias": "encoder.layers.8.feed_forward1.linear1.bias",
  "encoder.layers.8.feed_forward1.linear1.weight": "encoder.layers.8.feed_forward1.linear1.weight",
  "encoder.layers.8.feed_forward1.linear2.bias": "encoder.layers.8.feed_forward1.linear2.bias",
  "encoder.layers.8.feed_forward1.linear2.weight": "encoder.layers.8.feed_forward1.linear2.weight",
  "encoder.layers.8.feed_forward2.linear1.bias": "encoder.layers.8.feed_forward2.linear1.bias",
  "encoder.layers.8.feed_forward2.linear1.weight": "encoder.layers.8.feed_forward2.linear1.weight",
  "encoder.layers.8.feed_forward2.linear2.bias": "encoder.layers.8.feed_forward2.linear2.bias",
  "encoder.layers.8.feed_forward2.linear2.weight": "encoder.layers.8.feed_forward2.linear2.weight",
  "encoder.layers.8.norm_conv.bias": "encoder.layers.8.norm_conv.bias",
  "encoder.layers.8.norm_conv.weight": "encoder.layers.8.norm_conv.weight",
  "encoder.layers.8.norm_feed_forward1.bias": "encoder.layers.8.norm_feed_forward1.bias",
  "encoder.layers.8.norm_feed_forward1.weight": "encoder.layers.8.norm_feed_forward1.weight",
  "encoder.layers.8.norm_feed_forward2.bias": "encoder.layers.8.norm_feed_forward2.bias",
  "encoder.layers.8.norm_feed_forward2.weight": "encoder.layers.8.norm_feed_forward2.weight",
  "encoder.layers.8.norm_out.bias": "encoder.layers.8.norm_out.bias",
  "encoder.layers.8.norm_out.weight": "encoder.layers.8.norm_out.weight",
  "encoder.layers.8.norm_self_att.bias": "encoder.layers.8.norm_self_att.bias",
  "encoder.layers.8.norm_self_att.weight": "encoder.layers.8.norm_self_att.weight",
  "encoder.layers.8.self_attn.linear_out.bias": "encoder.layers.8.self_attn.out_proj.bias",
  "encoder.layers.8.self_attn.linear_out.weight": "encoder.layers.8.self_attn.out_proj.weight",
  "encoder.layers.8.self_attn.linear_pos.weight": "encoder.layers.8.self_attn.pos_proj.weight",
  "encoder.layers.8.self_attn.<qkv>.bias": "encoder.layers.8.self_attn.qkv_proj.bias",
  "encoder.layers.8.self_attn.<qkv>.weight": "encoder.layers.8.self_attn.qkv_proj.weight",
  "encoder.layers.8.self_attn.pos_bias_u": "encoder.layers.8.self_attn.pos_bias_u",
  "encoder.layers.8.self_attn.pos_bias_v": "encoder.layers.8.self_attn.pos_bias_v",
  "encoder.layers.9.conv.batch_norm.bias": "encoder.layers.9.conv.batch_norm.bias",
  "encoder.layers.9.conv.batch_norm.running_mean": "encoder.layers.9.conv.batch_norm.running_mean",
  "encoder.layers.9.conv.batch_norm.running_var": "encoder.layers.9.conv.batch_norm.running_var",
  "encoder.layers.9.conv.batch_norm.weight": "encoder.layers.9.conv.batch_norm.weight",
  "encoder.layers.9.conv.depthwise_conv.bias": "encoder.layers.9.conv.depthwise_conv.bias",
  "encoder.layers.9.conv.depthwise_conv.weight": "encoder.layers.9.conv.depthwise_conv.weight",
  "encoder.layers.9.conv.pointwise_conv1.bias": "encoder.layers.9.conv.pointwise_conv1.bias",
  "encoder.layers.9.conv.pointwise_conv1.weight": "encoder.layers.9.conv.pointwise_conv1.weight",
  "encoder.layers.9.conv.pointwise_conv2.bias": "encoder.layers.9.conv.pointwise_conv2.bias",
  "encoder.layers.9.conv.pointwise_conv2.weight": "encoder.layers.9.conv.pointwise_conv2.weight",
  "encoder.layers.9.feed_forward1.linear1.bias": "encoder.layers.9.feed_forward1.linear1.bias",
  "encoder.layers.9.feed_forward1.linear1.weight": "encoder.layers.9.feed_forward1.linear1.weight",
  "encoder.layers.9.feed_forward1.linear2.bias": "encoder.layers.9.feed_forward1.linear2.bias",
  "encoder.layers.9.feed_forward1.linear2.weight": "encoder.layers.9.feed_forward1.linear2.weight",
  "encoder.layers.9.feed_forward2.linear1.bias": "encoder.layers.9.feed_forward2.linear1.bias",
  "encoder.layers.9.feed_forward2.linear1.weight": "encoder.layers.9.feed_forward2.linear1.weight",
  "encoder.layers.9.feed_forward2.linear2.bias": "encoder.layers.9.feed_forward2.linear2.bias",
  "encoder.layers.9.feed_forward2.linear2.weight": "encoder.layers.9.feed_forward2.linear2.weight",
  "encoder.layers.9.norm_conv.bias": "encoder.layers.9.norm_conv.bias",
  "encoder.layers.9.norm_conv.weight": "encoder.layers.9.norm_conv.weight",
  "encoder.layers.9.norm_feed_forward1.bias": "encoder.layers.9.norm_feed_forward1.bias",
  "encoder.layers.9.norm_feed_forward1.weight": "encoder.layers.9.norm_feed_forward1.weight",
  "encoder.layers.9.norm_feed_forward2.bias": "encoder.layers.9.norm_feed_forward2.bias",
  "encoder.layers.9.norm_feed_forward2.weight": "encoder.layers.9.norm_feed_forward2.weight",
  "encoder.layers.9.norm_out.bias": "encoder.layers.9.norm_out.bias",
  "encoder.layers.9.norm_out.weight": "encoder.layers.9.norm_out.weight",
  "encoder.layers.9.norm_self_att.bias": "encoder.layers.9.norm_self_att.bias",
  "encoder.layers.9.norm_self_att.weight": "encoder.layers.9.norm_self_att.weight",
  "encoder.layers.9.self_attn.linear_out.bias": "encoder.layers.9.self_attn.out_proj.bias",
  "encoder.layers.9.self_attn.linear_out.weight": "encoder.layers.9.self_attn.out_proj.weight",
  "encoder.layers.9.self_attn.linear_pos.weight": "encoder.layers.9.self_attn.pos_proj.weight",
  "encoder.layers.9.self_attn.<qkv>.bias": "encoder.layers.9.self_attn.qkv_proj.bias",
  "encoder.layers.9.self_attn.<qkv>.weight": "encoder.layers.9.self_attn.qkv_proj.weight",
  "encoder.layers.9.self_attn.pos_bias_u": "encoder.layers.9.self_attn.pos_bias_u",
  "encoder.layers.9.self_attn.pos_bias_v": "encoder.layers.9.self_attn.pos_bias_v",
  "encoder.pre_encode.conv.0.bias": "encoder.subsampling.conv.0.bias",
  "encoder.pre_encode.conv.0.weight": "encoder.subsampling.conv.0.weight",
  "encoder.pre_encode.conv.2.bias": "encoder.subsampling.conv.2.bias",
  "encoder.pre_encode.conv.2.weight": "encoder.subsampling.conv.2.weight",
  "encoder.pre_encode.conv.3.bias": "encoder.subsampling.conv.3.bias",
  "encoder.pre_encode.conv.3.weight": "encoder.subsampling.conv.3.weight",
  "encoder.pre_encode.conv.5.bias": "encoder.subsampling.conv.5.bias",
  "encoder.pre_encode.conv.5.weight": "encoder.subsampling.conv.5.weight",
  "encoder.pre_encode.conv.6.bias": "encoder.subsampling.conv.6.bias",
  "encoder.pre_encode.conv.6.weight": "encoder.subsampling.conv.6.weight",
  "encoder.pre_encode.out.bias": "encoder.subsampling.out.bias",
  "encoder.pre_encode.out.weight": "encoder.subsampling.out.weight",
  "encoder_decoder_proj.bias": "bridge_proj.bias",
  "encoder_decoder_proj.weight": "bridge_proj.weight",
  "log_softmax.mlp.layer0.bias": "lm_head.bias",
  "log_softmax.mlp.layer0.weight": "lm_head.weight",
  "preprocessor.featurizer.fb": "preprocessor.featurizer.fb",
  "preprocessor.featurizer.window": "preprocessor.featurizer.window",
  "transf_decoder._decoder.final_layer_norm.bias": "decoder.core.final_layer_norm.bias",
  "transf_decoder._decoder.final_layer_norm.weight": "decoder.core.final_layer_norm.weight",
  "transf_decoder._decoder.layers.0.first_sub_layer.out_projection.bias": "decoder.core.layers.0.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.0.first_sub_layer.out_projection.weight": "decoder.core.layers.0.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.0.first_sub_layer.<qkv>.bias": "decoder.core.layers.0.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.0.first_sub_layer.<qkv>.weight": "decoder.core.layers.0.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.0.layer_norm_1.bias": "decoder.core.layers.0.layer_norm_1.bias",
  "transf_decoder._decoder.layers.0.layer_norm_1.weight": "decoder.core.layers.0.layer_norm_1.weight",
  "transf_decoder._decoder.layers.0.layer_norm_2.bias": "decoder.core.layers.0.layer_norm_2.bias",
  "transf_decoder._decoder.layers.0.layer_norm_2.weight": "decoder.core.layers.0.layer_norm_2.weight",
  "transf_decoder._decoder.layers.0.layer_norm_3.bias": "decoder.core.layers.0.layer_norm_3.bias",
  "transf_decoder._decoder.layers.0.layer_norm_3.weight": "decoder.core.layers.0.layer_norm_3.weight",
  "transf_decoder._decoder.layers.0.second_sub_layer.out_projection.bias": "decoder.core.layers.0.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.0.second_sub_layer.out_projection.weight": "decoder.core.layers.0.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.0.second_sub_layer.<qkv>.bias": "decoder.core.layers.0.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.0.second_sub_layer.<qkv>.weight": "decoder.core.layers.0.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.0.third_sub_layer.dense_in.bias": "decoder.core.layers.0.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.0.third_sub_layer.dense_in.weight": "decoder.core.layers.0.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.0.third_sub_layer.dense_out.bias": "decoder.core.layers.0.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.0.third_sub_layer.dense_out.weight": "decoder.core.layers.0.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.1.first_sub_layer.out_projection.bias": "decoder.core.layers.1.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.1.first_sub_layer.out_projection.weight": "decoder.core.layers.1.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.1.first_sub_layer.<qkv>.bias": "decoder.core.layers.1.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.1.first_sub_layer.<qkv>.weight": "decoder.core.layers.1.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.1.layer_norm_1.bias": "decoder.core.layers.1.layer_norm_1.bias",
  "transf_decoder._decoder.layers.1.layer_norm_1.weight": "decoder.core.layers.1.layer_norm_1.weight",
  "transf_decoder._decoder.layers.1.layer_norm_2.bias": "decoder.core.layers.1.layer_norm_2.bias",
  "transf_decoder._decoder.layers.1.layer_norm_2.weight": "decoder.core.layers.1.layer_norm_2.weight",
  "transf_decoder._decoder.layers.1.layer_norm_3.bias": "decoder.core.layers.1.layer_norm_3.bias",
  "transf_decoder._decoder.layers.1.layer_norm_3.weight": "decoder.core.layers.1.layer_norm_3.weight",
  "transf_decoder._decoder.layers.1.second_sub_layer.out_projection.bias": "decoder.core.layers.1.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.1.second_sub_layer.out_projection.weight": "decoder.core.layers.1.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.1.second_sub_layer.<qkv>.bias": "decoder.core.layers.1.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.1.second_sub_layer.<qkv>.weight": "decoder.core.layers.1.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.1.third_sub_layer.dense_in.bias": "decoder.core.layers.1.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.1.third_sub_layer.dense_in.weight": "decoder.core.layers.1.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.1.third_sub_layer.dense_out.bias": "decoder.core.layers.1.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.1.third_sub_layer.dense_out.weight": "decoder.core.layers.1.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.2.first_sub_layer.out_projection.bias": "decoder.core.layers.2.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.2.first_sub_layer.out_projection.weight": "decoder.core.layers.2.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.2.first_sub_layer.<qkv>.bias": "decoder.core.layers.2.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.2.first_sub_layer.<qkv>.weight": "decoder.core.layers.2.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.2.layer_norm_1.bias": "decoder.core.layers.2.layer_norm_1.bias",
  "transf_decoder._decoder.layers.2.layer_norm_1.weight": "decoder.core.layers.2.layer_norm_1.weight",
  "transf_decoder._decoder.layers.2.layer_norm_2.bias": "decoder.core.layers.2.layer_norm_2.bias",
  "transf_decoder._decoder.layers.2.layer_norm_2.weight": "decoder.core.layers.2.layer_norm_2.weight",
  "transf_decoder._decoder.layers.2.layer_norm_3.bias": "decoder.core.layers.2.layer_norm_3.bias",
  "transf_decoder._decoder.layers.2.layer_norm_3.weight": "decoder.core.layers.2.layer_norm_3.weight",
  "transf_decoder._decoder.layers.2.second_sub_layer.out_projection.bias": "decoder.core.layers.2.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.2.second_sub_layer.out_projection.weight": "decoder.core.layers.2.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.2.second_sub_layer.<qkv>.bias": "decoder.core.layers.2.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.2.second_sub_layer.<qkv>.weight": "decoder.core.layers.2.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.2.third_sub_layer.dense_in.bias": "decoder.core.layers.2.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.2.third_sub_layer.dense_in.weight": "decoder.core.layers.2.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.2.third_sub_layer.dense_out.bias": "decoder.core.layers.2.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.2.third_sub_layer.dense_out.weight": "decoder.core.layers.2.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.3.first_sub_layer.out_projection.bias": "decoder.core.layers.3.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.3.first_sub_layer.out_projection.weight": "decoder.core.layers.3.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.3.first_sub_layer.<qkv>.bias": "decoder.core.layers.3.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.3.first_sub_layer.<qkv>.weight": "decoder.core.layers.3.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.3.layer_norm_1.bias": "decoder.core.layers.3.layer_norm_1.bias",
  "transf_decoder._decoder.layers.3.layer_norm_1.weight": "decoder.core.layers.3.layer_norm_1.weight",
  "transf_decoder._decoder.layers.3.layer_norm_2.bias": "decoder.core.layers.3.layer_norm_2.bias",
  "transf_decoder._decoder.layers.3.layer_norm_2.weight": "decoder.core.layers.3.layer_norm_2.weight",
  "transf_decoder._decoder.layers.3.layer_norm_3.bias": "decoder.core.layers.3.layer_norm_3.bias",
  "transf_decoder._decoder.layers.3.layer_norm_3.weight": "decoder.core.layers.3.layer_norm_3.weight",
  "transf_decoder._decoder.layers.3.second_sub_layer.out_projection.bias": "decoder.core.layers.3.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.3.second_sub_layer.out_projection.weight": "decoder.core.layers.3.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.3.second_sub_layer.<qkv>.bias": "decoder.core.layers.3.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.3.second_sub_layer.<qkv>.weight": "decoder.core.layers.3.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.3.third_sub_layer.dense_in.bias": "decoder.core.layers.3.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.3.third_sub_layer.dense_in.weight": "decoder.core.layers.3.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.3.third_sub_layer.dense_out.bias": "decoder.core.layers.3.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.3.third_sub_layer.dense_out.weight": "decoder.core.layers.3.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.4.first_sub_layer.out_projection.bias": "decoder.core.layers.4.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.4.first_sub_layer.out_projection.weight": "decoder.core.layers.4.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.4.first_sub_layer.<qkv>.bias": "decoder.core.layers.4.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.4.first_sub_layer.<qkv>.weight": "decoder.core.layers.4.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.4.layer_norm_1.bias": "decoder.core.layers.4.layer_norm_1.bias",
  "transf_decoder._decoder.layers.4.layer_norm_1.weight": "decoder.core.layers.4.layer_norm_1.weight",
  "transf_decoder._decoder.layers.4.layer_norm_2.bias": "decoder.core.layers.4.layer_norm_2.bias",
  "transf_decoder._decoder.layers.4.layer_norm_2.weight": "decoder.core.layers.4.layer_norm_2.weight",
  "transf_decoder._decoder.layers.4.layer_norm_3.bias": "decoder.core.layers.4.layer_norm_3.bias",
  "transf_decoder._decoder.layers.4.layer_norm_3.weight": "decoder.core.layers.4.layer_norm_3.weight",
  "transf_decoder._decoder.layers.4.second_sub_layer.out_projection.bias": "decoder.core.layers.4.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.4.second_sub_layer.out_projection.weight": "decoder.core.layers.4.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.4.second_sub_layer.<qkv>.bias": "decoder.core.layers.4.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.4.second_sub_layer.<qkv>.weight": "decoder.core.layers.4.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.4.third_sub_layer.dense_in.bias": "decoder.core.layers.4.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.4.third_sub_layer.dense_in.weight": "decoder.core.layers.4.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.4.third_sub_layer.dense_out.bias": "decoder.core.layers.4.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.4.third_sub_layer.dense_out.weight": "decoder.core.layers.4.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.5.first_sub_layer.out_projection.bias": "decoder.core.layers.5.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.5.first_sub_layer.out_projection.weight": "decoder.core.layers.5.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.5.first_sub_layer.<qkv>.bias": "decoder.core.layers.5.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.5.first_sub_layer.<qkv>.weight": "decoder.core.layers.5.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.5.layer_norm_1.bias": "decoder.core.layers.5.layer_norm_1.bias",
  "transf_decoder._decoder.layers.5.layer_norm_1.weight": "decoder.core.layers.5.layer_norm_1.weight",
  "transf_decoder._decoder.layers.5.layer_norm_2.bias": "decoder.core.layers.5.layer_norm_2.bias",
  "transf_decoder._decoder.layers.5.layer_norm_2.weight": "decoder.core.layers.5.layer_norm_2.weight",
  "transf_decoder._decoder.layers.5.layer_norm_3.bias": "decoder.core.layers.5.layer_norm_3.bias",
  "transf_decoder._decoder.layers.5.layer_norm_3.weight": "decoder.core.layers.5.layer_norm_3.weight",
  "transf_decoder._decoder.layers.5.second_sub_layer.out_projection.bias": "decoder.core.layers.5.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.5.second_sub_layer.out_projection.weight": "decoder.core.layers.5.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.5.second_sub_layer.<qkv>.bias": "decoder.core.layers.5.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.5.second_sub_layer.<qkv>.weight": "decoder.core.layers.5.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.5.third_sub_layer.dense_in.bias": "decoder.core.layers.5.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.5.third_sub_layer.dense_in.weight": "decoder.core.layers.5.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.5.third_sub_layer.dense_out.bias": "decoder.core.layers.5.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.5.third_sub_layer.dense_out.weight": "decoder.core.layers.5.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.6.first_sub_layer.out_projection.bias": "decoder.core.layers.6.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.6.first_sub_layer.out_projection.weight": "decoder.core.layers.6.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.6.first_sub_layer.<qkv>.bias": "decoder.core.layers.6.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.6.first_sub_layer.<qkv>.weight": "decoder.core.layers.6.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.6.layer_norm_1.bias": "decoder.core.layers.6.layer_norm_1.bias",
  "transf_decoder._decoder.layers.6.layer_norm_1.weight": "decoder.core.layers.6.layer_norm_1.weight",
  "transf_decoder._decoder.layers.6.layer_norm_2.bias": "decoder.core.layers.6.layer_norm_2.bias",
  "transf_decoder._decoder.layers.6.layer_norm_2.weight": "decoder.core.layers.6.layer_norm_2.weight",
  "transf_decoder._decoder.layers.6.layer_norm_3.bias": "decoder.core.layers.6.layer_norm_3.bias",
  "transf_decoder._decoder.layers.6.layer_norm_3.weight": "decoder.core.layers.6.layer_norm_3.weight",
  "transf_decoder._decoder.layers.6.second_sub_layer.out_projection.bias": "decoder.core.layers.6.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.6.second_sub_layer.out_projection.weight": "decoder.core.layers.6.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.6.second_sub_layer.<qkv>.bias": "decoder.core.layers.6.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.6.second_sub_layer.<qkv>.weight": "decoder.core.layers.6.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.6.third_sub_layer.dense_in.bias": "decoder.core.layers.6.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.6.third_sub_layer.dense_in.weight": "decoder.core.layers.6.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.6.third_sub_layer.dense_out.bias": "decoder.core.layers.6.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.6.third_sub_layer.dense_out.weight": "decoder.core.layers.6.third_sub_layer.dense_out.weight",
  "transf_decoder._decoder.layers.7.first_sub_layer.out_projection.bias": "decoder.core.layers.7.first_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.7.first_sub_layer.out_projection.weight": "decoder.core.layers.7.first_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.7.first_sub_layer.<qkv>.bias": "decoder.core.layers.7.first_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.7.first_sub_layer.<qkv>.weight": "decoder.core.layers.7.first_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.7.layer_norm_1.bias": "decoder.core.layers.7.layer_norm_1.bias",
  "transf_decoder._decoder.layers.7.layer_norm_1.weight": "decoder.core.layers.7.layer_norm_1.weight",
  "transf_decoder._decoder.layers.7.layer_norm_2.bias": "decoder.core.layers.7.layer_norm_2.bias",
  "transf_decoder._decoder.layers.7.layer_norm_2.weight": "decoder.core.layers.7.layer_norm_2.weight",
  "transf_decoder._decoder.layers.7.layer_norm_3.bias": "decoder.core.layers.7.layer_norm_3.bias",
  "transf_decoder._decoder.layers.7.layer_norm_3.weight": "decoder.core.layers.7.layer_norm_3.weight",
  "transf_decoder._decoder.layers.7.second_sub_layer.out_projection.bias": "decoder.core.layers.7.second_sub_layer.out_proj.bias",
  "transf_decoder._decoder.layers.7.second_sub_layer.out_projection.weight": "decoder.core.layers.7.second_sub_layer.out_proj.weight",
  "transf_decoder._decoder.layers.7.second_sub_layer.<qkv>.bias": "decoder.core.layers.7.second_sub_layer.qkv_proj.bias",
  "transf_decoder._decoder.layers.7.second_sub_layer.<qkv>.weight": "decoder.core.layers.7.second_sub_layer.qkv_proj.weight",
  "transf_decoder._decoder.layers.7.third_sub_layer.dense_in.bias": "decoder.core.layers.7.third_sub_layer.dense_in.bias",
  "transf_decoder._decoder.layers.7.third_sub_layer.dense_in.weight": "decoder.core.layers.7.third_sub_layer.dense_in.weight",
  "transf_decoder._decoder.layers.7.third_sub_layer.dense_out.bias": "decoder.core.layers.7.third_sub_layer.dense_out.bias",
  "transf_decoder._decoder.layers.7.third_sub_layer.dense_out.weight": "decoder.core.layers.7.third_sub_layer.dense_out.weight",
  "transf_decoder._embedding.layer_norm.bias": "decoder.embedding.layer_norm.bias",
  "transf_decoder._embedding.layer_norm.weight": "decoder.embedding.layer_norm.weight",
  "transf_decoder._embedding.position_embedding.pos_enc": "decoder.embedding.position_embedding.pos_enc",
  "transf_decoder._embedding.token_embedding.weight": "decoder.embedding.token_embedding.weight"
}