|
{ |
|
"_name_or_path": "/mnt/dolphinfs/ssd_pool/docker/user/hadoop-vacv/lidengjie/projects/fsvd-llm/cache/llama-7b-hf/models/stage1-bf16-usv", |
|
"architectures": [ |
|
"LlamaForCausalLM" |
|
], |
|
"attention_bias": false, |
|
"attention_dropout": 0.0, |
|
"bos_token_id": 1, |
|
"eos_token_id": 2, |
|
"hidden_act": "silu", |
|
"hidden_size": 4096, |
|
"initializer_range": 0.02, |
|
"intermediate_size": 11008, |
|
"linear_info": { |
|
"model.layers.0.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.0.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 1368 |
|
}, |
|
"model.layers.0.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 1372 |
|
}, |
|
"model.layers.0.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 239 |
|
}, |
|
"model.layers.0.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 794 |
|
}, |
|
"model.layers.0.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 202 |
|
}, |
|
"model.layers.0.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1213 |
|
}, |
|
"model.layers.1.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.1.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 1883 |
|
}, |
|
"model.layers.1.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 1946 |
|
}, |
|
"model.layers.1.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 459 |
|
}, |
|
"model.layers.1.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1356 |
|
}, |
|
"model.layers.1.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 448 |
|
}, |
|
"model.layers.1.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1358 |
|
}, |
|
"model.layers.10.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.10.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 1999 |
|
}, |
|
"model.layers.10.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.10.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 845 |
|
}, |
|
"model.layers.10.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.10.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 882 |
|
}, |
|
"model.layers.10.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.11.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.11.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2079 |
|
}, |
|
"model.layers.11.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.11.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 760 |
|
}, |
|
"model.layers.11.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.11.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 778 |
|
}, |
|
"model.layers.11.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.12.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.12.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2056 |
|
}, |
|
"model.layers.12.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.12.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 832 |
|
}, |
|
"model.layers.12.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.12.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 909 |
|
}, |
|
"model.layers.12.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.13.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.13.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2101 |
|
}, |
|
"model.layers.13.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.13.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 743 |
|
}, |
|
"model.layers.13.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1397 |
|
}, |
|
"model.layers.13.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 836 |
|
}, |
|
"model.layers.13.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.14.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.14.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2190 |
|
}, |
|
"model.layers.14.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.14.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 877 |
|
}, |
|
"model.layers.14.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.14.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 933 |
|
}, |
|
"model.layers.14.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.15.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.15.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2240 |
|
}, |
|
"model.layers.15.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.15.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 780 |
|
}, |
|
"model.layers.15.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.15.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 887 |
|
}, |
|
"model.layers.15.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.16.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.16.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.16.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.16.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 858 |
|
}, |
|
"model.layers.16.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.16.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 885 |
|
}, |
|
"model.layers.16.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.17.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.17.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2261 |
|
}, |
|
"model.layers.17.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.17.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 835 |
|
}, |
|
"model.layers.17.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.17.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 913 |
|
}, |
|
"model.layers.17.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.18.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.18.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2292 |
|
}, |
|
"model.layers.18.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.18.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 743 |
|
}, |
|
"model.layers.18.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.18.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 775 |
|
}, |
|
"model.layers.18.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.19.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.19.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2283 |
|
}, |
|
"model.layers.19.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.19.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 682 |
|
}, |
|
"model.layers.19.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1515 |
|
}, |
|
"model.layers.19.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 758 |
|
}, |
|
"model.layers.19.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1468 |
|
}, |
|
"model.layers.2.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.2.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2008 |
|
}, |
|
"model.layers.2.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.2.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 658 |
|
}, |
|
"model.layers.2.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.2.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 630 |
|
}, |
|
"model.layers.2.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.20.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.20.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2285 |
|
}, |
|
"model.layers.20.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.20.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 762 |
|
}, |
|
"model.layers.20.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1472 |
|
}, |
|
"model.layers.20.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 804 |
|
}, |
|
"model.layers.20.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1495 |
|
}, |
|
"model.layers.21.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.21.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2287 |
|
}, |
|
"model.layers.21.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.21.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 586 |
|
}, |
|
"model.layers.21.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1502 |
|
}, |
|
"model.layers.21.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 609 |
|
}, |
|
"model.layers.21.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1391 |
|
}, |
|
"model.layers.22.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.22.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2306 |
|
}, |
|
"model.layers.22.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.22.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 541 |
|
}, |
|
"model.layers.22.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1008 |
|
}, |
|
"model.layers.22.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 545 |
|
}, |
|
"model.layers.22.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1084 |
|
}, |
|
"model.layers.23.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.23.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2316 |
|
}, |
|
"model.layers.23.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.23.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 575 |
|
}, |
|
"model.layers.23.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1220 |
|
}, |
|
"model.layers.23.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 519 |
|
}, |
|
"model.layers.23.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1169 |
|
}, |
|
"model.layers.24.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.24.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2356 |
|
}, |
|
"model.layers.24.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.24.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 442 |
|
}, |
|
"model.layers.24.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1067 |
|
}, |
|
"model.layers.24.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 470 |
|
}, |
|
"model.layers.24.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1116 |
|
}, |
|
"model.layers.25.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.25.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.25.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.25.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 485 |
|
}, |
|
"model.layers.25.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1042 |
|
}, |
|
"model.layers.25.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 533 |
|
}, |
|
"model.layers.25.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1179 |
|
}, |
|
"model.layers.26.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.26.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.26.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.26.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 481 |
|
}, |
|
"model.layers.26.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1006 |
|
}, |
|
"model.layers.26.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 572 |
|
}, |
|
"model.layers.26.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1154 |
|
}, |
|
"model.layers.27.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.27.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.27.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.27.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 370 |
|
}, |
|
"model.layers.27.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 790 |
|
}, |
|
"model.layers.27.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 414 |
|
}, |
|
"model.layers.27.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1019 |
|
}, |
|
"model.layers.28.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.28.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.28.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.28.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 490 |
|
}, |
|
"model.layers.28.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1084 |
|
}, |
|
"model.layers.28.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 494 |
|
}, |
|
"model.layers.28.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1161 |
|
}, |
|
"model.layers.29.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.29.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.29.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.29.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 374 |
|
}, |
|
"model.layers.29.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 955 |
|
}, |
|
"model.layers.29.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 411 |
|
}, |
|
"model.layers.29.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1123 |
|
}, |
|
"model.layers.3.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.3.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.3.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.3.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 704 |
|
}, |
|
"model.layers.3.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1196 |
|
}, |
|
"model.layers.3.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 698 |
|
}, |
|
"model.layers.3.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.30.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.30.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.30.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.30.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 282 |
|
}, |
|
"model.layers.30.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 967 |
|
}, |
|
"model.layers.30.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 297 |
|
}, |
|
"model.layers.30.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1146 |
|
}, |
|
"model.layers.31.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.31.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.31.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.31.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 337 |
|
}, |
|
"model.layers.31.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 852 |
|
}, |
|
"model.layers.31.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 341 |
|
}, |
|
"model.layers.31.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1060 |
|
}, |
|
"model.layers.4.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.4.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.4.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.4.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 834 |
|
}, |
|
"model.layers.4.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.4.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 815 |
|
}, |
|
"model.layers.4.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.5.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.5.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2259 |
|
}, |
|
"model.layers.5.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.5.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 810 |
|
}, |
|
"model.layers.5.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1354 |
|
}, |
|
"model.layers.5.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 843 |
|
}, |
|
"model.layers.5.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.6.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.6.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2239 |
|
}, |
|
"model.layers.6.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.6.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 793 |
|
}, |
|
"model.layers.6.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.6.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 890 |
|
}, |
|
"model.layers.6.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.7.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.7.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2060 |
|
}, |
|
"model.layers.7.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.7.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 727 |
|
}, |
|
"model.layers.7.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.7.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 736 |
|
}, |
|
"model.layers.7.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.8.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.8.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 2007 |
|
}, |
|
"model.layers.8.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.8.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 831 |
|
}, |
|
"model.layers.8.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1477 |
|
}, |
|
"model.layers.8.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 823 |
|
}, |
|
"model.layers.8.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.9.mlp.down_proj": { |
|
"bias": true, |
|
"in_features": 11008, |
|
"out_features": 4096, |
|
"rank": 0 |
|
}, |
|
"model.layers.9.mlp.gate_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 1995 |
|
}, |
|
"model.layers.9.mlp.up_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 11008, |
|
"rank": 0 |
|
}, |
|
"model.layers.9.self_attn.k_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 822 |
|
}, |
|
"model.layers.9.self_attn.o_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1403 |
|
}, |
|
"model.layers.9.self_attn.q_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 873 |
|
}, |
|
"model.layers.9.self_attn.v_proj": { |
|
"bias": true, |
|
"in_features": 4096, |
|
"out_features": 4096, |
|
"rank": 1321 |
|
} |
|
}, |
|
"max_position_embeddings": 2048, |
|
"mlp_bias": false, |
|
"model_type": "llama", |
|
"num_attention_heads": 32, |
|
"num_hidden_layers": 32, |
|
"num_key_value_heads": 32, |
|
"pad_token_id": 0, |
|
"pretraining_tp": 1, |
|
"rms_norm_eps": 1e-06, |
|
"rope_scaling": null, |
|
"rope_theta": 10000.0, |
|
"tie_word_embeddings": false, |
|
"torch_dtype": "bfloat16", |
|
"transformers_version": "4.43.2", |
|
"use_cache": true, |
|
"vocab_size": 32000 |
|
} |
|
|