| { | |
| "metadata": { | |
| "total_size": 11962152320 | |
| }, | |
| "weight_map": { | |
| "transformer.h.0.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.0.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.1.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.10.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.10.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.11.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.12.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.13.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.14.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.15.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.15.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.16.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.17.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.18.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.19.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.2.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.2.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.20.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.20.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.21.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.resid_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.resid_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.resid_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.resid_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.ln_2.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.ln_2.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.mlp.c_fc.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.mlp.c_fc.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.mlp.c_proj.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.22.mlp.c_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.attn_dropout.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.attn_dropout.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.attn_dropout.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.attn_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.attn_dropout.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_proj.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_proj.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_proj.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_proj.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.c_proj.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.dimension_reduce.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.k_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.o_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.attn.resid_dropout.q_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.attn.v_proj.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.ln_1.bias": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.ln_1.weight": "model-00003-of-00006.safetensors", | |
| "transformer.h.23.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.23.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.24.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.25.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.26.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.27.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.28.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.29.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.3.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.3.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.30.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.mlp.c_fc.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.mlp.c_fc.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.mlp.c_proj.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.30.mlp.c_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.attn_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.attn_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.attn_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.attn_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.attn_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_proj.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_proj.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_proj.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_proj.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.c_proj.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.resid_dropout.dimension_reduce.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.resid_dropout.k_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.resid_dropout.o_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.resid_dropout.q_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.resid_dropout.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.attn.v_proj.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.ln_1.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.ln_1.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.ln_2.bias": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.ln_2.weight": "model-00004-of-00006.safetensors", | |
| "transformer.h.31.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.31.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.31.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.31.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.32.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.33.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.34.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.35.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.36.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.37.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.mlp.c_proj.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.38.mlp.c_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.attn_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.attn_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.attn_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.attn_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.attn_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_proj.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_proj.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_proj.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_proj.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.c_proj.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.resid_dropout.dimension_reduce.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.resid_dropout.k_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.resid_dropout.o_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.resid_dropout.q_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.resid_dropout.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.attn.v_proj.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.ln_1.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.ln_1.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.ln_2.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.ln_2.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.mlp.c_fc.bias": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.mlp.c_fc.weight": "model-00005-of-00006.safetensors", | |
| "transformer.h.39.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.39.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.4.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.4.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.40.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.40.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.41.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.42.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.43.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.44.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.45.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.46.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.attn_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.attn_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.attn_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.attn_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.attn_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_proj.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_proj.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_proj.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_proj.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.c_proj.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.resid_dropout.dimension_reduce.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.resid_dropout.k_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.resid_dropout.o_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.resid_dropout.q_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.resid_dropout.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.attn.v_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.ln_1.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.ln_1.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.ln_2.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.ln_2.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.mlp.c_fc.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.mlp.c_fc.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.mlp.c_proj.bias": "model-00006-of-00006.safetensors", | |
| "transformer.h.47.mlp.c_proj.weight": "model-00006-of-00006.safetensors", | |
| "transformer.h.5.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.5.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_proj.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_proj.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_proj.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_proj.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.c_proj.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.resid_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.resid_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.resid_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.resid_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.resid_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.ln_2.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.ln_2.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.mlp.c_fc.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.mlp.c_fc.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.mlp.c_proj.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.6.mlp.c_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.attn_dropout.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.attn_dropout.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.attn_dropout.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.attn_dropout.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.attn_dropout.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.c_attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.c_attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.c_attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.dimension_reduce.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.k_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.o_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.q_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.attn.v_proj.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.ln_1.bias": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.ln_1.weight": "model-00001-of-00006.safetensors", | |
| "transformer.h.7.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.7.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.8.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.attn_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.attn_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.attn_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.attn_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.attn_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_proj.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_proj.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_proj.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_proj.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.c_proj.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.resid_dropout.dimension_reduce.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.resid_dropout.k_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.resid_dropout.o_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.resid_dropout.q_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.resid_dropout.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.attn.v_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.ln_1.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.ln_1.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.ln_2.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.ln_2.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.mlp.c_fc.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.mlp.c_fc.weight": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.mlp.c_proj.bias": "model-00002-of-00006.safetensors", | |
| "transformer.h.9.mlp.c_proj.weight": "model-00002-of-00006.safetensors", | |
| "transformer.ln_f.bias": "model-00006-of-00006.safetensors", | |
| "transformer.ln_f.weight": "model-00006-of-00006.safetensors", | |
| "transformer.wpe.weight": "model-00001-of-00006.safetensors", | |
| "transformer.wte.weight": "model-00001-of-00006.safetensors" | |
| } | |
| } | |