Automatic Speech Recognition
MLX
Safetensors
cohere_asr
speech-recognition
transcription
audio
apple
macos
on-device
quantized
mixed-precision
3bit
4bit
custom_code
4-bit precision
Instructions to use MarkChen1214/cohere-transcribe-03-2026-MLX-Mixed-3bit4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use MarkChen1214/cohere-transcribe-03-2026-MLX-Mixed-3bit4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir cohere-transcribe-03-2026-MLX-Mixed-3bit4bit MarkChen1214/cohere-transcribe-03-2026-MLX-Mixed-3bit4bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
| { | |
| "architectures": [ | |
| "CohereAsrForConditionalGeneration" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_cohere_asr.CohereAsrConfig", | |
| "AutoFeatureExtractor": "processing_cohere_asr.CohereAsrFeatureExtractor", | |
| "AutoModel": "modeling_cohere_asr.CohereAsrModel", | |
| "AutoModelForSpeechSeq2Seq": "modeling_cohere_asr.CohereAsrForConditionalGeneration", | |
| "AutoProcessor": "processing_cohere_asr.CohereAsrProcessor", | |
| "AutoTokenizer": "tokenization_cohere_asr.CohereAsrTokenizer" | |
| }, | |
| "batch_size": 128, | |
| "codex_pointwise_linearized": true, | |
| "decoding": { | |
| "beam": { | |
| "beam_size": 1, | |
| "len_pen": 0.0, | |
| "max_generation_delta": 50 | |
| }, | |
| "return_best_hypothesis": true, | |
| "strategy": "beam" | |
| }, | |
| "encoder": { | |
| "att_context_size": [ | |
| -1, | |
| -1 | |
| ], | |
| "causal_downsampling": false, | |
| "conv_context_size": null, | |
| "conv_kernel_size": 9, | |
| "conv_norm_type": "batch_norm", | |
| "d_model": 1280, | |
| "dropout": 0, | |
| "dropout_att": 0, | |
| "dropout_emb": 0, | |
| "dropout_pre_encoder": 0, | |
| "feat_in": 128, | |
| "feat_out": -1, | |
| "ff_expansion_factor": 4, | |
| "n_heads": 8, | |
| "n_layers": 48, | |
| "pos_emb_max_len": 5000, | |
| "reduction": null, | |
| "reduction_factor": 1, | |
| "reduction_position": null, | |
| "self_attention_model": "rel_pos", | |
| "subsampling": "dw_striding", | |
| "subsampling_conv_channels": 256, | |
| "subsampling_factor": 8, | |
| "untie_biases": true, | |
| "xscaling": false | |
| }, | |
| "head": { | |
| "activation": "relu", | |
| "dropout": 0, | |
| "hidden_size": 1024, | |
| "log_softmax": true, | |
| "num_classes": 16384, | |
| "num_layers": 1, | |
| "use_transformer_init": true | |
| }, | |
| "is_encoder_decoder": true, | |
| "log_batch_stats": false, | |
| "log_prediction": true, | |
| "max_audio_clip_s": 35, | |
| "max_seq_len": 1024, | |
| "model_defaults": { | |
| "asr_enc_hidden": 1280, | |
| "lm_dec_hidden": 1024, | |
| "lm_enc_hidden": 1024 | |
| }, | |
| "model_type": "cohere_asr", | |
| "multitask_metrics_cfg": { | |
| "log_predictions": true, | |
| "metrics": { | |
| "wer": { | |
| "constraint": ".source_lang==.target_lang" | |
| } | |
| } | |
| }, | |
| "overlap_chunk_second": 5, | |
| "preprocessor": { | |
| "dither": 1e-05, | |
| "features": 128, | |
| "frame_splicing": 1, | |
| "log": true, | |
| "n_fft": 512, | |
| "normalize": "per_feature", | |
| "pad_to": 0, | |
| "pad_value": 0.0, | |
| "sample_rate": 16000, | |
| "window": "hann", | |
| "window_size": 0.025, | |
| "window_stride": 0.01 | |
| }, | |
| "prompt_defaults": [ | |
| { | |
| "role": "user", | |
| "slots": { | |
| "decodercontext": "", | |
| "diarize": "<|nodiarize|>", | |
| "emotion": "<|emo:undefined|>", | |
| "itn": "<|noitn|>", | |
| "pnc": "<|pnc|>", | |
| "source_lang": "<|en|>", | |
| "target_lang": "<|en|>", | |
| "timestamp": "<|notimestamp|>" | |
| } | |
| }, | |
| { | |
| "role": "user_partial", | |
| "slots": { | |
| "decodercontext": "" | |
| } | |
| } | |
| ], | |
| "prompt_format": "cohere_asr", | |
| "quantization": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine", | |
| "encoder.pre_encode.out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.embedding.token_embedding": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "encoder_decoder_proj": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "log_softmax.mlp.layer0": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| } | |
| }, | |
| "quantization_config": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine", | |
| "encoder.pre_encode.out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.0.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.1.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.2.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.3.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.4.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.5.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.6.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.7.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.8.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.9.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.10.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.11.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.12.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.13.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.14.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.15.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.16.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.17.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.18.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.19.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.20.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.21.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.22.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.23.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.24.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.25.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.26.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.27.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.28.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.29.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.30.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.31.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.32.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.33.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.34.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.35.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.36.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.37.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.38.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.39.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.40.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.41.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.42.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.43.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.44.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.45.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.46.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward1.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward1.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_q": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_k": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_v": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_pos": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.self_attn.linear_out": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.conv.pointwise_conv1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.conv.pointwise_conv2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward2.linear1": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "encoder.layers.47.feed_forward2.linear2": { | |
| "group_size": 64, | |
| "bits": 3, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.embedding.token_embedding": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.0.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.1.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.2.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.3.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.4.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.5.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.6.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.first_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.query_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.key_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.value_net": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.second_sub_layer.out_projection": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.third_sub_layer.dense_in": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "transf_decoder.decoder.layers.7.third_sub_layer.dense_out": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "encoder_decoder_proj": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| }, | |
| "log_softmax.mlp.layer0": { | |
| "group_size": 64, | |
| "bits": 4, | |
| "mode": "affine" | |
| } | |
| }, | |
| "sample_rate": 16000, | |
| "supported_languages": [ | |
| "en", | |
| "fr", | |
| "de", | |
| "es", | |
| "it", | |
| "pt", | |
| "nl", | |
| "pl", | |
| "el", | |
| "ar", | |
| "ja", | |
| "zh", | |
| "vi", | |
| "ko" | |
| ], | |
| "transf_decoder": { | |
| "config_dict": { | |
| "attn_layer_dropout": 0, | |
| "attn_score_dropout": 0, | |
| "embedding_dropout": 0, | |
| "ffn_dropout": 0, | |
| "hidden_act": "relu", | |
| "hidden_size": 1024, | |
| "inner_size": 4096, | |
| "learn_positional_encodings": false, | |
| "lm_dec_hidden": 1280, | |
| "max_sequence_length": 1024, | |
| "num_attention_heads": 8, | |
| "num_layers": 8, | |
| "num_token_types": 0, | |
| "pre_ln": true, | |
| "vocab_size": "None" | |
| }, | |
| "encoder": null, | |
| "model_name": null, | |
| "pre_ln_final_layer_norm": true, | |
| "pretrained": false | |
| }, | |
| "transf_encoder": { | |
| "attn_layer_dropout": 0, | |
| "attn_score_dropout": 0, | |
| "ffn_dropout": 0, | |
| "hidden_size": 1024, | |
| "inner_size": 4096, | |
| "mask_future": false, | |
| "num_attention_heads": 8, | |
| "num_layers": 0, | |
| "pre_ln": true, | |
| "pre_ln_final_layer_norm": true | |
| }, | |
| "use_loss_mask_for_prompt": false, | |
| "vocab_size": 16384 | |
| } |