glennmatlin's picture
download
raw
40.7 kB
{
"num_grads": 17971,
"dtype": {
"names": [
"layers.0.self_attn.q_proj",
"layers.0.self_attn.k_proj",
"layers.0.self_attn.v_proj",
"layers.0.self_attn.o_proj",
"layers.0.mlp.gate_proj",
"layers.0.mlp.up_proj",
"layers.0.mlp.down_proj",
"layers.1.self_attn.q_proj",
"layers.1.self_attn.k_proj",
"layers.1.self_attn.v_proj",
"layers.1.self_attn.o_proj",
"layers.1.mlp.gate_proj",
"layers.1.mlp.up_proj",
"layers.1.mlp.down_proj",
"layers.2.self_attn.q_proj",
"layers.2.self_attn.k_proj",
"layers.2.self_attn.v_proj",
"layers.2.self_attn.o_proj",
"layers.2.mlp.gate_proj",
"layers.2.mlp.up_proj",
"layers.2.mlp.down_proj",
"layers.3.self_attn.q_proj",
"layers.3.self_attn.k_proj",
"layers.3.self_attn.v_proj",
"layers.3.self_attn.o_proj",
"layers.3.mlp.gate_proj",
"layers.3.mlp.up_proj",
"layers.3.mlp.down_proj",
"layers.4.self_attn.q_proj",
"layers.4.self_attn.k_proj",
"layers.4.self_attn.v_proj",
"layers.4.self_attn.o_proj",
"layers.4.mlp.gate_proj",
"layers.4.mlp.up_proj",
"layers.4.mlp.down_proj",
"layers.5.self_attn.q_proj",
"layers.5.self_attn.k_proj",
"layers.5.self_attn.v_proj",
"layers.5.self_attn.o_proj",
"layers.5.mlp.gate_proj",
"layers.5.mlp.up_proj",
"layers.5.mlp.down_proj",
"layers.6.self_attn.q_proj",
"layers.6.self_attn.k_proj",
"layers.6.self_attn.v_proj",
"layers.6.self_attn.o_proj",
"layers.6.mlp.gate_proj",
"layers.6.mlp.up_proj",
"layers.6.mlp.down_proj",
"layers.7.self_attn.q_proj",
"layers.7.self_attn.k_proj",
"layers.7.self_attn.v_proj",
"layers.7.self_attn.o_proj",
"layers.7.mlp.gate_proj",
"layers.7.mlp.up_proj",
"layers.7.mlp.down_proj",
"layers.8.self_attn.q_proj",
"layers.8.self_attn.k_proj",
"layers.8.self_attn.v_proj",
"layers.8.self_attn.o_proj",
"layers.8.mlp.gate_proj",
"layers.8.mlp.up_proj",
"layers.8.mlp.down_proj",
"layers.9.self_attn.q_proj",
"layers.9.self_attn.k_proj",
"layers.9.self_attn.v_proj",
"layers.9.self_attn.o_proj",
"layers.9.mlp.gate_proj",
"layers.9.mlp.up_proj",
"layers.9.mlp.down_proj",
"layers.10.self_attn.q_proj",
"layers.10.self_attn.k_proj",
"layers.10.self_attn.v_proj",
"layers.10.self_attn.o_proj",
"layers.10.mlp.gate_proj",
"layers.10.mlp.up_proj",
"layers.10.mlp.down_proj",
"layers.11.self_attn.q_proj",
"layers.11.self_attn.k_proj",
"layers.11.self_attn.v_proj",
"layers.11.self_attn.o_proj",
"layers.11.mlp.gate_proj",
"layers.11.mlp.up_proj",
"layers.11.mlp.down_proj",
"layers.12.self_attn.q_proj",
"layers.12.self_attn.k_proj",
"layers.12.self_attn.v_proj",
"layers.12.self_attn.o_proj",
"layers.12.mlp.gate_proj",
"layers.12.mlp.up_proj",
"layers.12.mlp.down_proj",
"layers.13.self_attn.q_proj",
"layers.13.self_attn.k_proj",
"layers.13.self_attn.v_proj",
"layers.13.self_attn.o_proj",
"layers.13.mlp.gate_proj",
"layers.13.mlp.up_proj",
"layers.13.mlp.down_proj",
"layers.14.self_attn.q_proj",
"layers.14.self_attn.k_proj",
"layers.14.self_attn.v_proj",
"layers.14.self_attn.o_proj",
"layers.14.mlp.gate_proj",
"layers.14.mlp.up_proj",
"layers.14.mlp.down_proj",
"layers.15.self_attn.q_proj",
"layers.15.self_attn.k_proj",
"layers.15.self_attn.v_proj",
"layers.15.self_attn.o_proj",
"layers.15.mlp.gate_proj",
"layers.15.mlp.up_proj",
"layers.15.mlp.down_proj",
"layers.16.self_attn.q_proj",
"layers.16.self_attn.k_proj",
"layers.16.self_attn.v_proj",
"layers.16.self_attn.o_proj",
"layers.16.mlp.gate_proj",
"layers.16.mlp.up_proj",
"layers.16.mlp.down_proj",
"layers.17.self_attn.q_proj",
"layers.17.self_attn.k_proj",
"layers.17.self_attn.v_proj",
"layers.17.self_attn.o_proj",
"layers.17.mlp.gate_proj",
"layers.17.mlp.up_proj",
"layers.17.mlp.down_proj",
"layers.18.self_attn.q_proj",
"layers.18.self_attn.k_proj",
"layers.18.self_attn.v_proj",
"layers.18.self_attn.o_proj",
"layers.18.mlp.gate_proj",
"layers.18.mlp.up_proj",
"layers.18.mlp.down_proj",
"layers.19.self_attn.q_proj",
"layers.19.self_attn.k_proj",
"layers.19.self_attn.v_proj",
"layers.19.self_attn.o_proj",
"layers.19.mlp.gate_proj",
"layers.19.mlp.up_proj",
"layers.19.mlp.down_proj",
"layers.20.self_attn.q_proj",
"layers.20.self_attn.k_proj",
"layers.20.self_attn.v_proj",
"layers.20.self_attn.o_proj",
"layers.20.mlp.gate_proj",
"layers.20.mlp.up_proj",
"layers.20.mlp.down_proj",
"layers.21.self_attn.q_proj",
"layers.21.self_attn.k_proj",
"layers.21.self_attn.v_proj",
"layers.21.self_attn.o_proj",
"layers.21.mlp.gate_proj",
"layers.21.mlp.up_proj",
"layers.21.mlp.down_proj",
"layers.22.self_attn.q_proj",
"layers.22.self_attn.k_proj",
"layers.22.self_attn.v_proj",
"layers.22.self_attn.o_proj",
"layers.22.mlp.gate_proj",
"layers.22.mlp.up_proj",
"layers.22.mlp.down_proj",
"layers.23.self_attn.q_proj",
"layers.23.self_attn.k_proj",
"layers.23.self_attn.v_proj",
"layers.23.self_attn.o_proj",
"layers.23.mlp.gate_proj",
"layers.23.mlp.up_proj",
"layers.23.mlp.down_proj",
"layers.24.self_attn.q_proj",
"layers.24.self_attn.k_proj",
"layers.24.self_attn.v_proj",
"layers.24.self_attn.o_proj",
"layers.24.mlp.gate_proj",
"layers.24.mlp.up_proj",
"layers.24.mlp.down_proj",
"layers.25.self_attn.q_proj",
"layers.25.self_attn.k_proj",
"layers.25.self_attn.v_proj",
"layers.25.self_attn.o_proj",
"layers.25.mlp.gate_proj",
"layers.25.mlp.up_proj",
"layers.25.mlp.down_proj",
"layers.26.self_attn.q_proj",
"layers.26.self_attn.k_proj",
"layers.26.self_attn.v_proj",
"layers.26.self_attn.o_proj",
"layers.26.mlp.gate_proj",
"layers.26.mlp.up_proj",
"layers.26.mlp.down_proj",
"layers.27.self_attn.q_proj",
"layers.27.self_attn.k_proj",
"layers.27.self_attn.v_proj",
"layers.27.self_attn.o_proj",
"layers.27.mlp.gate_proj",
"layers.27.mlp.up_proj",
"layers.27.mlp.down_proj",
"layers.28.self_attn.q_proj",
"layers.28.self_attn.k_proj",
"layers.28.self_attn.v_proj",
"layers.28.self_attn.o_proj",
"layers.28.mlp.gate_proj",
"layers.28.mlp.up_proj",
"layers.28.mlp.down_proj",
"layers.29.self_attn.q_proj",
"layers.29.self_attn.k_proj",
"layers.29.self_attn.v_proj",
"layers.29.self_attn.o_proj",
"layers.29.mlp.gate_proj",
"layers.29.mlp.up_proj",
"layers.29.mlp.down_proj",
"layers.30.self_attn.q_proj",
"layers.30.self_attn.k_proj",
"layers.30.self_attn.v_proj",
"layers.30.self_attn.o_proj",
"layers.30.mlp.gate_proj",
"layers.30.mlp.up_proj",
"layers.30.mlp.down_proj",
"layers.31.self_attn.q_proj",
"layers.31.self_attn.k_proj",
"layers.31.self_attn.v_proj",
"layers.31.self_attn.o_proj",
"layers.31.mlp.gate_proj",
"layers.31.mlp.up_proj",
"layers.31.mlp.down_proj",
"layers.32.self_attn.q_proj",
"layers.32.self_attn.k_proj",
"layers.32.self_attn.v_proj",
"layers.32.self_attn.o_proj",
"layers.32.mlp.gate_proj",
"layers.32.mlp.up_proj",
"layers.32.mlp.down_proj",
"layers.33.self_attn.q_proj",
"layers.33.self_attn.k_proj",
"layers.33.self_attn.v_proj",
"layers.33.self_attn.o_proj",
"layers.33.mlp.gate_proj",
"layers.33.mlp.up_proj",
"layers.33.mlp.down_proj",
"layers.34.self_attn.q_proj",
"layers.34.self_attn.k_proj",
"layers.34.self_attn.v_proj",
"layers.34.self_attn.o_proj",
"layers.34.mlp.gate_proj",
"layers.34.mlp.up_proj",
"layers.34.mlp.down_proj",
"layers.35.self_attn.q_proj",
"layers.35.self_attn.k_proj",
"layers.35.self_attn.v_proj",
"layers.35.self_attn.o_proj",
"layers.35.mlp.gate_proj",
"layers.35.mlp.up_proj",
"layers.35.mlp.down_proj",
"layers.36.self_attn.q_proj",
"layers.36.self_attn.k_proj",
"layers.36.self_attn.v_proj",
"layers.36.self_attn.o_proj",
"layers.36.mlp.gate_proj",
"layers.36.mlp.up_proj",
"layers.36.mlp.down_proj",
"layers.37.self_attn.q_proj",
"layers.37.self_attn.k_proj",
"layers.37.self_attn.v_proj",
"layers.37.self_attn.o_proj",
"layers.37.mlp.gate_proj",
"layers.37.mlp.up_proj",
"layers.37.mlp.down_proj",
"layers.38.self_attn.q_proj",
"layers.38.self_attn.k_proj",
"layers.38.self_attn.v_proj",
"layers.38.self_attn.o_proj",
"layers.38.mlp.gate_proj",
"layers.38.mlp.up_proj",
"layers.38.mlp.down_proj",
"layers.39.self_attn.q_proj",
"layers.39.self_attn.k_proj",
"layers.39.self_attn.v_proj",
"layers.39.self_attn.o_proj",
"layers.39.mlp.gate_proj",
"layers.39.mlp.up_proj",
"layers.39.mlp.down_proj",
"layers.40.self_attn.q_proj",
"layers.40.self_attn.k_proj",
"layers.40.self_attn.v_proj",
"layers.40.self_attn.o_proj",
"layers.40.mlp.gate_proj",
"layers.40.mlp.up_proj",
"layers.40.mlp.down_proj",
"layers.41.self_attn.q_proj",
"layers.41.self_attn.k_proj",
"layers.41.self_attn.v_proj",
"layers.41.self_attn.o_proj",
"layers.41.mlp.gate_proj",
"layers.41.mlp.up_proj",
"layers.41.mlp.down_proj",
"layers.42.self_attn.q_proj",
"layers.42.self_attn.k_proj",
"layers.42.self_attn.v_proj",
"layers.42.self_attn.o_proj",
"layers.42.mlp.gate_proj",
"layers.42.mlp.up_proj",
"layers.42.mlp.down_proj",
"layers.43.self_attn.q_proj",
"layers.43.self_attn.k_proj",
"layers.43.self_attn.v_proj",
"layers.43.self_attn.o_proj",
"layers.43.mlp.gate_proj",
"layers.43.mlp.up_proj",
"layers.43.mlp.down_proj",
"layers.44.self_attn.q_proj",
"layers.44.self_attn.k_proj",
"layers.44.self_attn.v_proj",
"layers.44.self_attn.o_proj",
"layers.44.mlp.gate_proj",
"layers.44.mlp.up_proj",
"layers.44.mlp.down_proj",
"layers.45.self_attn.q_proj",
"layers.45.self_attn.k_proj",
"layers.45.self_attn.v_proj",
"layers.45.self_attn.o_proj",
"layers.45.mlp.gate_proj",
"layers.45.mlp.up_proj",
"layers.45.mlp.down_proj",
"layers.46.self_attn.q_proj",
"layers.46.self_attn.k_proj",
"layers.46.self_attn.v_proj",
"layers.46.self_attn.o_proj",
"layers.46.mlp.gate_proj",
"layers.46.mlp.up_proj",
"layers.46.mlp.down_proj",
"layers.47.self_attn.q_proj",
"layers.47.self_attn.k_proj",
"layers.47.self_attn.v_proj",
"layers.47.self_attn.o_proj",
"layers.47.mlp.gate_proj",
"layers.47.mlp.up_proj",
"layers.47.mlp.down_proj",
"layers.48.self_attn.q_proj",
"layers.48.self_attn.k_proj",
"layers.48.self_attn.v_proj",
"layers.48.self_attn.o_proj",
"layers.48.mlp.gate_proj",
"layers.48.mlp.up_proj",
"layers.48.mlp.down_proj",
"layers.49.self_attn.q_proj",
"layers.49.self_attn.k_proj",
"layers.49.self_attn.v_proj",
"layers.49.self_attn.o_proj",
"layers.49.mlp.gate_proj",
"layers.49.mlp.up_proj",
"layers.49.mlp.down_proj",
"layers.50.self_attn.q_proj",
"layers.50.self_attn.k_proj",
"layers.50.self_attn.v_proj",
"layers.50.self_attn.o_proj",
"layers.50.mlp.gate_proj",
"layers.50.mlp.up_proj",
"layers.50.mlp.down_proj",
"layers.51.self_attn.q_proj",
"layers.51.self_attn.k_proj",
"layers.51.self_attn.v_proj",
"layers.51.self_attn.o_proj",
"layers.51.mlp.gate_proj",
"layers.51.mlp.up_proj",
"layers.51.mlp.down_proj",
"layers.52.self_attn.q_proj",
"layers.52.self_attn.k_proj",
"layers.52.self_attn.v_proj",
"layers.52.self_attn.o_proj",
"layers.52.mlp.gate_proj",
"layers.52.mlp.up_proj",
"layers.52.mlp.down_proj",
"layers.53.self_attn.q_proj",
"layers.53.self_attn.k_proj",
"layers.53.self_attn.v_proj",
"layers.53.self_attn.o_proj",
"layers.53.mlp.gate_proj",
"layers.53.mlp.up_proj",
"layers.53.mlp.down_proj",
"layers.54.self_attn.q_proj",
"layers.54.self_attn.k_proj",
"layers.54.self_attn.v_proj",
"layers.54.self_attn.o_proj",
"layers.54.mlp.gate_proj",
"layers.54.mlp.up_proj",
"layers.54.mlp.down_proj",
"layers.55.self_attn.q_proj",
"layers.55.self_attn.k_proj",
"layers.55.self_attn.v_proj",
"layers.55.self_attn.o_proj",
"layers.55.mlp.gate_proj",
"layers.55.mlp.up_proj",
"layers.55.mlp.down_proj",
"layers.56.self_attn.q_proj",
"layers.56.self_attn.k_proj",
"layers.56.self_attn.v_proj",
"layers.56.self_attn.o_proj",
"layers.56.mlp.gate_proj",
"layers.56.mlp.up_proj",
"layers.56.mlp.down_proj",
"layers.57.self_attn.q_proj",
"layers.57.self_attn.k_proj",
"layers.57.self_attn.v_proj",
"layers.57.self_attn.o_proj",
"layers.57.mlp.gate_proj",
"layers.57.mlp.up_proj",
"layers.57.mlp.down_proj",
"layers.58.self_attn.q_proj",
"layers.58.self_attn.k_proj",
"layers.58.self_attn.v_proj",
"layers.58.self_attn.o_proj",
"layers.58.mlp.gate_proj",
"layers.58.mlp.up_proj",
"layers.58.mlp.down_proj",
"layers.59.self_attn.q_proj",
"layers.59.self_attn.k_proj",
"layers.59.self_attn.v_proj",
"layers.59.self_attn.o_proj",
"layers.59.mlp.gate_proj",
"layers.59.mlp.up_proj",
"layers.59.mlp.down_proj",
"layers.60.self_attn.q_proj",
"layers.60.self_attn.k_proj",
"layers.60.self_attn.v_proj",
"layers.60.self_attn.o_proj",
"layers.60.mlp.gate_proj",
"layers.60.mlp.up_proj",
"layers.60.mlp.down_proj",
"layers.61.self_attn.q_proj",
"layers.61.self_attn.k_proj",
"layers.61.self_attn.v_proj",
"layers.61.self_attn.o_proj",
"layers.61.mlp.gate_proj",
"layers.61.mlp.up_proj",
"layers.61.mlp.down_proj",
"layers.62.self_attn.q_proj",
"layers.62.self_attn.k_proj",
"layers.62.self_attn.v_proj",
"layers.62.self_attn.o_proj",
"layers.62.mlp.gate_proj",
"layers.62.mlp.up_proj",
"layers.62.mlp.down_proj",
"layers.63.self_attn.q_proj",
"layers.63.self_attn.k_proj",
"layers.63.self_attn.v_proj",
"layers.63.self_attn.o_proj",
"layers.63.mlp.gate_proj",
"layers.63.mlp.up_proj",
"layers.63.mlp.down_proj"
],
"formats": [
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4",
"(256,)<f4"
],
"itemsize": 458752
},
"grad_sizes": {
"layers.0.self_attn.q_proj": 256,
"layers.0.self_attn.k_proj": 256,
"layers.0.self_attn.v_proj": 256,
"layers.0.self_attn.o_proj": 256,
"layers.0.mlp.gate_proj": 256,
"layers.0.mlp.up_proj": 256,
"layers.0.mlp.down_proj": 256,
"layers.1.self_attn.q_proj": 256,
"layers.1.self_attn.k_proj": 256,
"layers.1.self_attn.v_proj": 256,
"layers.1.self_attn.o_proj": 256,
"layers.1.mlp.gate_proj": 256,
"layers.1.mlp.up_proj": 256,
"layers.1.mlp.down_proj": 256,
"layers.2.self_attn.q_proj": 256,
"layers.2.self_attn.k_proj": 256,
"layers.2.self_attn.v_proj": 256,
"layers.2.self_attn.o_proj": 256,
"layers.2.mlp.gate_proj": 256,
"layers.2.mlp.up_proj": 256,
"layers.2.mlp.down_proj": 256,
"layers.3.self_attn.q_proj": 256,
"layers.3.self_attn.k_proj": 256,
"layers.3.self_attn.v_proj": 256,
"layers.3.self_attn.o_proj": 256,
"layers.3.mlp.gate_proj": 256,
"layers.3.mlp.up_proj": 256,
"layers.3.mlp.down_proj": 256,
"layers.4.self_attn.q_proj": 256,
"layers.4.self_attn.k_proj": 256,
"layers.4.self_attn.v_proj": 256,
"layers.4.self_attn.o_proj": 256,
"layers.4.mlp.gate_proj": 256,
"layers.4.mlp.up_proj": 256,
"layers.4.mlp.down_proj": 256,
"layers.5.self_attn.q_proj": 256,
"layers.5.self_attn.k_proj": 256,
"layers.5.self_attn.v_proj": 256,
"layers.5.self_attn.o_proj": 256,
"layers.5.mlp.gate_proj": 256,
"layers.5.mlp.up_proj": 256,
"layers.5.mlp.down_proj": 256,
"layers.6.self_attn.q_proj": 256,
"layers.6.self_attn.k_proj": 256,
"layers.6.self_attn.v_proj": 256,
"layers.6.self_attn.o_proj": 256,
"layers.6.mlp.gate_proj": 256,
"layers.6.mlp.up_proj": 256,
"layers.6.mlp.down_proj": 256,
"layers.7.self_attn.q_proj": 256,
"layers.7.self_attn.k_proj": 256,
"layers.7.self_attn.v_proj": 256,
"layers.7.self_attn.o_proj": 256,
"layers.7.mlp.gate_proj": 256,
"layers.7.mlp.up_proj": 256,
"layers.7.mlp.down_proj": 256,
"layers.8.self_attn.q_proj": 256,
"layers.8.self_attn.k_proj": 256,
"layers.8.self_attn.v_proj": 256,
"layers.8.self_attn.o_proj": 256,
"layers.8.mlp.gate_proj": 256,
"layers.8.mlp.up_proj": 256,
"layers.8.mlp.down_proj": 256,
"layers.9.self_attn.q_proj": 256,
"layers.9.self_attn.k_proj": 256,
"layers.9.self_attn.v_proj": 256,
"layers.9.self_attn.o_proj": 256,
"layers.9.mlp.gate_proj": 256,
"layers.9.mlp.up_proj": 256,
"layers.9.mlp.down_proj": 256,
"layers.10.self_attn.q_proj": 256,
"layers.10.self_attn.k_proj": 256,
"layers.10.self_attn.v_proj": 256,
"layers.10.self_attn.o_proj": 256,
"layers.10.mlp.gate_proj": 256,
"layers.10.mlp.up_proj": 256,
"layers.10.mlp.down_proj": 256,
"layers.11.self_attn.q_proj": 256,
"layers.11.self_attn.k_proj": 256,
"layers.11.self_attn.v_proj": 256,
"layers.11.self_attn.o_proj": 256,
"layers.11.mlp.gate_proj": 256,
"layers.11.mlp.up_proj": 256,
"layers.11.mlp.down_proj": 256,
"layers.12.self_attn.q_proj": 256,
"layers.12.self_attn.k_proj": 256,
"layers.12.self_attn.v_proj": 256,
"layers.12.self_attn.o_proj": 256,
"layers.12.mlp.gate_proj": 256,
"layers.12.mlp.up_proj": 256,
"layers.12.mlp.down_proj": 256,
"layers.13.self_attn.q_proj": 256,
"layers.13.self_attn.k_proj": 256,
"layers.13.self_attn.v_proj": 256,
"layers.13.self_attn.o_proj": 256,
"layers.13.mlp.gate_proj": 256,
"layers.13.mlp.up_proj": 256,
"layers.13.mlp.down_proj": 256,
"layers.14.self_attn.q_proj": 256,
"layers.14.self_attn.k_proj": 256,
"layers.14.self_attn.v_proj": 256,
"layers.14.self_attn.o_proj": 256,
"layers.14.mlp.gate_proj": 256,
"layers.14.mlp.up_proj": 256,
"layers.14.mlp.down_proj": 256,
"layers.15.self_attn.q_proj": 256,
"layers.15.self_attn.k_proj": 256,
"layers.15.self_attn.v_proj": 256,
"layers.15.self_attn.o_proj": 256,
"layers.15.mlp.gate_proj": 256,
"layers.15.mlp.up_proj": 256,
"layers.15.mlp.down_proj": 256,
"layers.16.self_attn.q_proj": 256,
"layers.16.self_attn.k_proj": 256,
"layers.16.self_attn.v_proj": 256,
"layers.16.self_attn.o_proj": 256,
"layers.16.mlp.gate_proj": 256,
"layers.16.mlp.up_proj": 256,
"layers.16.mlp.down_proj": 256,
"layers.17.self_attn.q_proj": 256,
"layers.17.self_attn.k_proj": 256,
"layers.17.self_attn.v_proj": 256,
"layers.17.self_attn.o_proj": 256,
"layers.17.mlp.gate_proj": 256,
"layers.17.mlp.up_proj": 256,
"layers.17.mlp.down_proj": 256,
"layers.18.self_attn.q_proj": 256,
"layers.18.self_attn.k_proj": 256,
"layers.18.self_attn.v_proj": 256,
"layers.18.self_attn.o_proj": 256,
"layers.18.mlp.gate_proj": 256,
"layers.18.mlp.up_proj": 256,
"layers.18.mlp.down_proj": 256,
"layers.19.self_attn.q_proj": 256,
"layers.19.self_attn.k_proj": 256,
"layers.19.self_attn.v_proj": 256,
"layers.19.self_attn.o_proj": 256,
"layers.19.mlp.gate_proj": 256,
"layers.19.mlp.up_proj": 256,
"layers.19.mlp.down_proj": 256,
"layers.20.self_attn.q_proj": 256,
"layers.20.self_attn.k_proj": 256,
"layers.20.self_attn.v_proj": 256,
"layers.20.self_attn.o_proj": 256,
"layers.20.mlp.gate_proj": 256,
"layers.20.mlp.up_proj": 256,
"layers.20.mlp.down_proj": 256,
"layers.21.self_attn.q_proj": 256,
"layers.21.self_attn.k_proj": 256,
"layers.21.self_attn.v_proj": 256,
"layers.21.self_attn.o_proj": 256,
"layers.21.mlp.gate_proj": 256,
"layers.21.mlp.up_proj": 256,
"layers.21.mlp.down_proj": 256,
"layers.22.self_attn.q_proj": 256,
"layers.22.self_attn.k_proj": 256,
"layers.22.self_attn.v_proj": 256,
"layers.22.self_attn.o_proj": 256,
"layers.22.mlp.gate_proj": 256,
"layers.22.mlp.up_proj": 256,
"layers.22.mlp.down_proj": 256,
"layers.23.self_attn.q_proj": 256,
"layers.23.self_attn.k_proj": 256,
"layers.23.self_attn.v_proj": 256,
"layers.23.self_attn.o_proj": 256,
"layers.23.mlp.gate_proj": 256,
"layers.23.mlp.up_proj": 256,
"layers.23.mlp.down_proj": 256,
"layers.24.self_attn.q_proj": 256,
"layers.24.self_attn.k_proj": 256,
"layers.24.self_attn.v_proj": 256,
"layers.24.self_attn.o_proj": 256,
"layers.24.mlp.gate_proj": 256,
"layers.24.mlp.up_proj": 256,
"layers.24.mlp.down_proj": 256,
"layers.25.self_attn.q_proj": 256,
"layers.25.self_attn.k_proj": 256,
"layers.25.self_attn.v_proj": 256,
"layers.25.self_attn.o_proj": 256,
"layers.25.mlp.gate_proj": 256,
"layers.25.mlp.up_proj": 256,
"layers.25.mlp.down_proj": 256,
"layers.26.self_attn.q_proj": 256,
"layers.26.self_attn.k_proj": 256,
"layers.26.self_attn.v_proj": 256,
"layers.26.self_attn.o_proj": 256,
"layers.26.mlp.gate_proj": 256,
"layers.26.mlp.up_proj": 256,
"layers.26.mlp.down_proj": 256,
"layers.27.self_attn.q_proj": 256,
"layers.27.self_attn.k_proj": 256,
"layers.27.self_attn.v_proj": 256,
"layers.27.self_attn.o_proj": 256,
"layers.27.mlp.gate_proj": 256,
"layers.27.mlp.up_proj": 256,
"layers.27.mlp.down_proj": 256,
"layers.28.self_attn.q_proj": 256,
"layers.28.self_attn.k_proj": 256,
"layers.28.self_attn.v_proj": 256,
"layers.28.self_attn.o_proj": 256,
"layers.28.mlp.gate_proj": 256,
"layers.28.mlp.up_proj": 256,
"layers.28.mlp.down_proj": 256,
"layers.29.self_attn.q_proj": 256,
"layers.29.self_attn.k_proj": 256,
"layers.29.self_attn.v_proj": 256,
"layers.29.self_attn.o_proj": 256,
"layers.29.mlp.gate_proj": 256,
"layers.29.mlp.up_proj": 256,
"layers.29.mlp.down_proj": 256,
"layers.30.self_attn.q_proj": 256,
"layers.30.self_attn.k_proj": 256,
"layers.30.self_attn.v_proj": 256,
"layers.30.self_attn.o_proj": 256,
"layers.30.mlp.gate_proj": 256,
"layers.30.mlp.up_proj": 256,
"layers.30.mlp.down_proj": 256,
"layers.31.self_attn.q_proj": 256,
"layers.31.self_attn.k_proj": 256,
"layers.31.self_attn.v_proj": 256,
"layers.31.self_attn.o_proj": 256,
"layers.31.mlp.gate_proj": 256,
"layers.31.mlp.up_proj": 256,
"layers.31.mlp.down_proj": 256,
"layers.32.self_attn.q_proj": 256,
"layers.32.self_attn.k_proj": 256,
"layers.32.self_attn.v_proj": 256,
"layers.32.self_attn.o_proj": 256,
"layers.32.mlp.gate_proj": 256,
"layers.32.mlp.up_proj": 256,
"layers.32.mlp.down_proj": 256,
"layers.33.self_attn.q_proj": 256,
"layers.33.self_attn.k_proj": 256,
"layers.33.self_attn.v_proj": 256,
"layers.33.self_attn.o_proj": 256,
"layers.33.mlp.gate_proj": 256,
"layers.33.mlp.up_proj": 256,
"layers.33.mlp.down_proj": 256,
"layers.34.self_attn.q_proj": 256,
"layers.34.self_attn.k_proj": 256,
"layers.34.self_attn.v_proj": 256,
"layers.34.self_attn.o_proj": 256,
"layers.34.mlp.gate_proj": 256,
"layers.34.mlp.up_proj": 256,
"layers.34.mlp.down_proj": 256,
"layers.35.self_attn.q_proj": 256,
"layers.35.self_attn.k_proj": 256,
"layers.35.self_attn.v_proj": 256,
"layers.35.self_attn.o_proj": 256,
"layers.35.mlp.gate_proj": 256,
"layers.35.mlp.up_proj": 256,
"layers.35.mlp.down_proj": 256,
"layers.36.self_attn.q_proj": 256,
"layers.36.self_attn.k_proj": 256,
"layers.36.self_attn.v_proj": 256,
"layers.36.self_attn.o_proj": 256,
"layers.36.mlp.gate_proj": 256,
"layers.36.mlp.up_proj": 256,
"layers.36.mlp.down_proj": 256,
"layers.37.self_attn.q_proj": 256,
"layers.37.self_attn.k_proj": 256,
"layers.37.self_attn.v_proj": 256,
"layers.37.self_attn.o_proj": 256,
"layers.37.mlp.gate_proj": 256,
"layers.37.mlp.up_proj": 256,
"layers.37.mlp.down_proj": 256,
"layers.38.self_attn.q_proj": 256,
"layers.38.self_attn.k_proj": 256,
"layers.38.self_attn.v_proj": 256,
"layers.38.self_attn.o_proj": 256,
"layers.38.mlp.gate_proj": 256,
"layers.38.mlp.up_proj": 256,
"layers.38.mlp.down_proj": 256,
"layers.39.self_attn.q_proj": 256,
"layers.39.self_attn.k_proj": 256,
"layers.39.self_attn.v_proj": 256,
"layers.39.self_attn.o_proj": 256,
"layers.39.mlp.gate_proj": 256,
"layers.39.mlp.up_proj": 256,
"layers.39.mlp.down_proj": 256,
"layers.40.self_attn.q_proj": 256,
"layers.40.self_attn.k_proj": 256,
"layers.40.self_attn.v_proj": 256,
"layers.40.self_attn.o_proj": 256,
"layers.40.mlp.gate_proj": 256,
"layers.40.mlp.up_proj": 256,
"layers.40.mlp.down_proj": 256,
"layers.41.self_attn.q_proj": 256,
"layers.41.self_attn.k_proj": 256,
"layers.41.self_attn.v_proj": 256,
"layers.41.self_attn.o_proj": 256,
"layers.41.mlp.gate_proj": 256,
"layers.41.mlp.up_proj": 256,
"layers.41.mlp.down_proj": 256,
"layers.42.self_attn.q_proj": 256,
"layers.42.self_attn.k_proj": 256,
"layers.42.self_attn.v_proj": 256,
"layers.42.self_attn.o_proj": 256,
"layers.42.mlp.gate_proj": 256,
"layers.42.mlp.up_proj": 256,
"layers.42.mlp.down_proj": 256,
"layers.43.self_attn.q_proj": 256,
"layers.43.self_attn.k_proj": 256,
"layers.43.self_attn.v_proj": 256,
"layers.43.self_attn.o_proj": 256,
"layers.43.mlp.gate_proj": 256,
"layers.43.mlp.up_proj": 256,
"layers.43.mlp.down_proj": 256,
"layers.44.self_attn.q_proj": 256,
"layers.44.self_attn.k_proj": 256,
"layers.44.self_attn.v_proj": 256,
"layers.44.self_attn.o_proj": 256,
"layers.44.mlp.gate_proj": 256,
"layers.44.mlp.up_proj": 256,
"layers.44.mlp.down_proj": 256,
"layers.45.self_attn.q_proj": 256,
"layers.45.self_attn.k_proj": 256,
"layers.45.self_attn.v_proj": 256,
"layers.45.self_attn.o_proj": 256,
"layers.45.mlp.gate_proj": 256,
"layers.45.mlp.up_proj": 256,
"layers.45.mlp.down_proj": 256,
"layers.46.self_attn.q_proj": 256,
"layers.46.self_attn.k_proj": 256,
"layers.46.self_attn.v_proj": 256,
"layers.46.self_attn.o_proj": 256,
"layers.46.mlp.gate_proj": 256,
"layers.46.mlp.up_proj": 256,
"layers.46.mlp.down_proj": 256,
"layers.47.self_attn.q_proj": 256,
"layers.47.self_attn.k_proj": 256,
"layers.47.self_attn.v_proj": 256,
"layers.47.self_attn.o_proj": 256,
"layers.47.mlp.gate_proj": 256,
"layers.47.mlp.up_proj": 256,
"layers.47.mlp.down_proj": 256,
"layers.48.self_attn.q_proj": 256,
"layers.48.self_attn.k_proj": 256,
"layers.48.self_attn.v_proj": 256,
"layers.48.self_attn.o_proj": 256,
"layers.48.mlp.gate_proj": 256,
"layers.48.mlp.up_proj": 256,
"layers.48.mlp.down_proj": 256,
"layers.49.self_attn.q_proj": 256,
"layers.49.self_attn.k_proj": 256,
"layers.49.self_attn.v_proj": 256,
"layers.49.self_attn.o_proj": 256,
"layers.49.mlp.gate_proj": 256,
"layers.49.mlp.up_proj": 256,
"layers.49.mlp.down_proj": 256,
"layers.50.self_attn.q_proj": 256,
"layers.50.self_attn.k_proj": 256,
"layers.50.self_attn.v_proj": 256,
"layers.50.self_attn.o_proj": 256,
"layers.50.mlp.gate_proj": 256,
"layers.50.mlp.up_proj": 256,
"layers.50.mlp.down_proj": 256,
"layers.51.self_attn.q_proj": 256,
"layers.51.self_attn.k_proj": 256,
"layers.51.self_attn.v_proj": 256,
"layers.51.self_attn.o_proj": 256,
"layers.51.mlp.gate_proj": 256,
"layers.51.mlp.up_proj": 256,
"layers.51.mlp.down_proj": 256,
"layers.52.self_attn.q_proj": 256,
"layers.52.self_attn.k_proj": 256,
"layers.52.self_attn.v_proj": 256,
"layers.52.self_attn.o_proj": 256,
"layers.52.mlp.gate_proj": 256,
"layers.52.mlp.up_proj": 256,
"layers.52.mlp.down_proj": 256,
"layers.53.self_attn.q_proj": 256,
"layers.53.self_attn.k_proj": 256,
"layers.53.self_attn.v_proj": 256,
"layers.53.self_attn.o_proj": 256,
"layers.53.mlp.gate_proj": 256,
"layers.53.mlp.up_proj": 256,
"layers.53.mlp.down_proj": 256,
"layers.54.self_attn.q_proj": 256,
"layers.54.self_attn.k_proj": 256,
"layers.54.self_attn.v_proj": 256,
"layers.54.self_attn.o_proj": 256,
"layers.54.mlp.gate_proj": 256,
"layers.54.mlp.up_proj": 256,
"layers.54.mlp.down_proj": 256,
"layers.55.self_attn.q_proj": 256,
"layers.55.self_attn.k_proj": 256,
"layers.55.self_attn.v_proj": 256,
"layers.55.self_attn.o_proj": 256,
"layers.55.mlp.gate_proj": 256,
"layers.55.mlp.up_proj": 256,
"layers.55.mlp.down_proj": 256,
"layers.56.self_attn.q_proj": 256,
"layers.56.self_attn.k_proj": 256,
"layers.56.self_attn.v_proj": 256,
"layers.56.self_attn.o_proj": 256,
"layers.56.mlp.gate_proj": 256,
"layers.56.mlp.up_proj": 256,
"layers.56.mlp.down_proj": 256,
"layers.57.self_attn.q_proj": 256,
"layers.57.self_attn.k_proj": 256,
"layers.57.self_attn.v_proj": 256,
"layers.57.self_attn.o_proj": 256,
"layers.57.mlp.gate_proj": 256,
"layers.57.mlp.up_proj": 256,
"layers.57.mlp.down_proj": 256,
"layers.58.self_attn.q_proj": 256,
"layers.58.self_attn.k_proj": 256,
"layers.58.self_attn.v_proj": 256,
"layers.58.self_attn.o_proj": 256,
"layers.58.mlp.gate_proj": 256,
"layers.58.mlp.up_proj": 256,
"layers.58.mlp.down_proj": 256,
"layers.59.self_attn.q_proj": 256,
"layers.59.self_attn.k_proj": 256,
"layers.59.self_attn.v_proj": 256,
"layers.59.self_attn.o_proj": 256,
"layers.59.mlp.gate_proj": 256,
"layers.59.mlp.up_proj": 256,
"layers.59.mlp.down_proj": 256,
"layers.60.self_attn.q_proj": 256,
"layers.60.self_attn.k_proj": 256,
"layers.60.self_attn.v_proj": 256,
"layers.60.self_attn.o_proj": 256,
"layers.60.mlp.gate_proj": 256,
"layers.60.mlp.up_proj": 256,
"layers.60.mlp.down_proj": 256,
"layers.61.self_attn.q_proj": 256,
"layers.61.self_attn.k_proj": 256,
"layers.61.self_attn.v_proj": 256,
"layers.61.self_attn.o_proj": 256,
"layers.61.mlp.gate_proj": 256,
"layers.61.mlp.up_proj": 256,
"layers.61.mlp.down_proj": 256,
"layers.62.self_attn.q_proj": 256,
"layers.62.self_attn.k_proj": 256,
"layers.62.self_attn.v_proj": 256,
"layers.62.self_attn.o_proj": 256,
"layers.62.mlp.gate_proj": 256,
"layers.62.mlp.up_proj": 256,
"layers.62.mlp.down_proj": 256,
"layers.63.self_attn.q_proj": 256,
"layers.63.self_attn.k_proj": 256,
"layers.63.self_attn.v_proj": 256,
"layers.63.self_attn.o_proj": 256,
"layers.63.mlp.gate_proj": 256,
"layers.63.mlp.up_proj": 256,
"layers.63.mlp.down_proj": 256
},
"base_dtype": "float32"
}

Xet Storage Details

Size:
40.7 kB
·
Xet hash:
efd2204f9481523ff5a1729887b87dfe312b9c77dc47dc56a2e9ade8a98a6e62

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.