| { |
| "architectures": [ |
| "Qwen2ForCausalLM" |
| ], |
| "attention_dropout": 0.0, |
| "bos_token_id": 151643, |
| "eos_token_id": 151645, |
| "hidden_act": "silu", |
| "hidden_size": 3584, |
| "initializer_range": 0.02, |
| "intermediate_size": 18944, |
| "max_position_embeddings": 32768, |
| "max_window_layers": 28, |
| "model_type": "qwen2", |
| "num_attention_heads": 28, |
| "num_hidden_layers": 28, |
| "num_key_value_heads": 4, |
| "quantization": { |
| "group_size": 64, |
| "bits": 4, |
| "model.embed_tokens": true, |
| "model.layers.0.self_attn.q_proj": true, |
| "model.layers.0.self_attn.k_proj": true, |
| "model.layers.0.self_attn.v_proj": true, |
| "model.layers.0.self_attn.o_proj": true, |
| "model.layers.0.self_attn.rope": false, |
| "model.layers.0.mlp.gate_proj": true, |
| "model.layers.0.mlp.down_proj": true, |
| "model.layers.0.mlp.up_proj": true, |
| "model.layers.0.input_layernorm": false, |
| "model.layers.0.post_attention_layernorm": false, |
| "model.layers.1.self_attn.q_proj": true, |
| "model.layers.1.self_attn.k_proj": true, |
| "model.layers.1.self_attn.v_proj": true, |
| "model.layers.1.self_attn.o_proj": true, |
| "model.layers.1.self_attn.rope": false, |
| "model.layers.1.mlp.gate_proj": true, |
| "model.layers.1.mlp.down_proj": true, |
| "model.layers.1.mlp.up_proj": true, |
| "model.layers.1.input_layernorm": false, |
| "model.layers.1.post_attention_layernorm": false, |
| "model.layers.2.self_attn.q_proj": true, |
| "model.layers.2.self_attn.k_proj": true, |
| "model.layers.2.self_attn.v_proj": true, |
| "model.layers.2.self_attn.o_proj": true, |
| "model.layers.2.self_attn.rope": false, |
| "model.layers.2.mlp.gate_proj": true, |
| "model.layers.2.mlp.down_proj": true, |
| "model.layers.2.mlp.up_proj": true, |
| "model.layers.2.input_layernorm": false, |
| "model.layers.2.post_attention_layernorm": false, |
| "model.layers.3.self_attn.q_proj": true, |
| "model.layers.3.self_attn.k_proj": true, |
| "model.layers.3.self_attn.v_proj": true, |
| "model.layers.3.self_attn.o_proj": true, |
| "model.layers.3.self_attn.rope": false, |
| "model.layers.3.mlp.gate_proj": true, |
| "model.layers.3.mlp.down_proj": true, |
| "model.layers.3.mlp.up_proj": true, |
| "model.layers.3.input_layernorm": false, |
| "model.layers.3.post_attention_layernorm": false, |
| "model.layers.4.self_attn.q_proj": true, |
| "model.layers.4.self_attn.k_proj": true, |
| "model.layers.4.self_attn.v_proj": true, |
| "model.layers.4.self_attn.o_proj": true, |
| "model.layers.4.self_attn.rope": false, |
| "model.layers.4.mlp.gate_proj": true, |
| "model.layers.4.mlp.down_proj": true, |
| "model.layers.4.mlp.up_proj": true, |
| "model.layers.4.input_layernorm": false, |
| "model.layers.4.post_attention_layernorm": false, |
| "model.layers.5.self_attn.q_proj": true, |
| "model.layers.5.self_attn.k_proj": true, |
| "model.layers.5.self_attn.v_proj": true, |
| "model.layers.5.self_attn.o_proj": true, |
| "model.layers.5.self_attn.rope": false, |
| "model.layers.5.mlp.gate_proj": true, |
| "model.layers.5.mlp.down_proj": true, |
| "model.layers.5.mlp.up_proj": true, |
| "model.layers.5.input_layernorm": false, |
| "model.layers.5.post_attention_layernorm": false, |
| "model.layers.6.self_attn.q_proj": true, |
| "model.layers.6.self_attn.k_proj": true, |
| "model.layers.6.self_attn.v_proj": true, |
| "model.layers.6.self_attn.o_proj": true, |
| "model.layers.6.self_attn.rope": false, |
| "model.layers.6.mlp.gate_proj": true, |
| "model.layers.6.mlp.down_proj": true, |
| "model.layers.6.mlp.up_proj": true, |
| "model.layers.6.input_layernorm": false, |
| "model.layers.6.post_attention_layernorm": false, |
| "model.layers.7.self_attn.q_proj": true, |
| "model.layers.7.self_attn.k_proj": true, |
| "model.layers.7.self_attn.v_proj": true, |
| "model.layers.7.self_attn.o_proj": true, |
| "model.layers.7.self_attn.rope": false, |
| "model.layers.7.mlp.gate_proj": true, |
| "model.layers.7.mlp.down_proj": true, |
| "model.layers.7.mlp.up_proj": true, |
| "model.layers.7.input_layernorm": false, |
| "model.layers.7.post_attention_layernorm": false, |
| "model.layers.8.self_attn.q_proj": true, |
| "model.layers.8.self_attn.k_proj": true, |
| "model.layers.8.self_attn.v_proj": true, |
| "model.layers.8.self_attn.o_proj": true, |
| "model.layers.8.self_attn.rope": false, |
| "model.layers.8.mlp.gate_proj": true, |
| "model.layers.8.mlp.down_proj": true, |
| "model.layers.8.mlp.up_proj": true, |
| "model.layers.8.input_layernorm": false, |
| "model.layers.8.post_attention_layernorm": false, |
| "model.layers.9.self_attn.q_proj": true, |
| "model.layers.9.self_attn.k_proj": true, |
| "model.layers.9.self_attn.v_proj": true, |
| "model.layers.9.self_attn.o_proj": true, |
| "model.layers.9.self_attn.rope": false, |
| "model.layers.9.mlp.gate_proj": true, |
| "model.layers.9.mlp.down_proj": true, |
| "model.layers.9.mlp.up_proj": true, |
| "model.layers.9.input_layernorm": false, |
| "model.layers.9.post_attention_layernorm": false, |
| "model.layers.10.self_attn.q_proj": true, |
| "model.layers.10.self_attn.k_proj": true, |
| "model.layers.10.self_attn.v_proj": true, |
| "model.layers.10.self_attn.o_proj": true, |
| "model.layers.10.self_attn.rope": false, |
| "model.layers.10.mlp.gate_proj": true, |
| "model.layers.10.mlp.down_proj": true, |
| "model.layers.10.mlp.up_proj": true, |
| "model.layers.10.input_layernorm": false, |
| "model.layers.10.post_attention_layernorm": false, |
| "model.layers.11.self_attn.q_proj": true, |
| "model.layers.11.self_attn.k_proj": true, |
| "model.layers.11.self_attn.v_proj": true, |
| "model.layers.11.self_attn.o_proj": true, |
| "model.layers.11.self_attn.rope": false, |
| "model.layers.11.mlp.gate_proj": true, |
| "model.layers.11.mlp.down_proj": true, |
| "model.layers.11.mlp.up_proj": true, |
| "model.layers.11.input_layernorm": false, |
| "model.layers.11.post_attention_layernorm": false, |
| "model.layers.12.self_attn.q_proj": true, |
| "model.layers.12.self_attn.k_proj": true, |
| "model.layers.12.self_attn.v_proj": true, |
| "model.layers.12.self_attn.o_proj": true, |
| "model.layers.12.self_attn.rope": false, |
| "model.layers.12.mlp.gate_proj": true, |
| "model.layers.12.mlp.down_proj": true, |
| "model.layers.12.mlp.up_proj": true, |
| "model.layers.12.input_layernorm": false, |
| "model.layers.12.post_attention_layernorm": false, |
| "model.layers.13.self_attn.q_proj": true, |
| "model.layers.13.self_attn.k_proj": true, |
| "model.layers.13.self_attn.v_proj": true, |
| "model.layers.13.self_attn.o_proj": true, |
| "model.layers.13.self_attn.rope": false, |
| "model.layers.13.mlp.gate_proj": true, |
| "model.layers.13.mlp.down_proj": true, |
| "model.layers.13.mlp.up_proj": true, |
| "model.layers.13.input_layernorm": false, |
| "model.layers.13.post_attention_layernorm": false, |
| "model.layers.14.self_attn.q_proj": true, |
| "model.layers.14.self_attn.k_proj": true, |
| "model.layers.14.self_attn.v_proj": true, |
| "model.layers.14.self_attn.o_proj": true, |
| "model.layers.14.self_attn.rope": false, |
| "model.layers.14.mlp.gate_proj": true, |
| "model.layers.14.mlp.down_proj": true, |
| "model.layers.14.mlp.up_proj": true, |
| "model.layers.14.input_layernorm": false, |
| "model.layers.14.post_attention_layernorm": false, |
| "model.layers.15.self_attn.q_proj": true, |
| "model.layers.15.self_attn.k_proj": true, |
| "model.layers.15.self_attn.v_proj": true, |
| "model.layers.15.self_attn.o_proj": true, |
| "model.layers.15.self_attn.rope": false, |
| "model.layers.15.mlp.gate_proj": true, |
| "model.layers.15.mlp.down_proj": true, |
| "model.layers.15.mlp.up_proj": true, |
| "model.layers.15.input_layernorm": false, |
| "model.layers.15.post_attention_layernorm": false, |
| "model.layers.16.self_attn.q_proj": true, |
| "model.layers.16.self_attn.k_proj": true, |
| "model.layers.16.self_attn.v_proj": true, |
| "model.layers.16.self_attn.o_proj": true, |
| "model.layers.16.self_attn.rope": false, |
| "model.layers.16.mlp.gate_proj": true, |
| "model.layers.16.mlp.down_proj": true, |
| "model.layers.16.mlp.up_proj": true, |
| "model.layers.16.input_layernorm": false, |
| "model.layers.16.post_attention_layernorm": false, |
| "model.layers.17.self_attn.q_proj": true, |
| "model.layers.17.self_attn.k_proj": true, |
| "model.layers.17.self_attn.v_proj": true, |
| "model.layers.17.self_attn.o_proj": true, |
| "model.layers.17.self_attn.rope": false, |
| "model.layers.17.mlp.gate_proj": true, |
| "model.layers.17.mlp.down_proj": true, |
| "model.layers.17.mlp.up_proj": true, |
| "model.layers.17.input_layernorm": false, |
| "model.layers.17.post_attention_layernorm": false, |
| "model.layers.18.self_attn.q_proj": true, |
| "model.layers.18.self_attn.k_proj": true, |
| "model.layers.18.self_attn.v_proj": true, |
| "model.layers.18.self_attn.o_proj": true, |
| "model.layers.18.self_attn.rope": false, |
| "model.layers.18.mlp.gate_proj": true, |
| "model.layers.18.mlp.down_proj": true, |
| "model.layers.18.mlp.up_proj": true, |
| "model.layers.18.input_layernorm": false, |
| "model.layers.18.post_attention_layernorm": false, |
| "model.layers.19.self_attn.q_proj": true, |
| "model.layers.19.self_attn.k_proj": true, |
| "model.layers.19.self_attn.v_proj": true, |
| "model.layers.19.self_attn.o_proj": true, |
| "model.layers.19.self_attn.rope": false, |
| "model.layers.19.mlp.gate_proj": true, |
| "model.layers.19.mlp.down_proj": true, |
| "model.layers.19.mlp.up_proj": true, |
| "model.layers.19.input_layernorm": false, |
| "model.layers.19.post_attention_layernorm": false, |
| "model.layers.20.self_attn.q_proj": true, |
| "model.layers.20.self_attn.k_proj": true, |
| "model.layers.20.self_attn.v_proj": true, |
| "model.layers.20.self_attn.o_proj": true, |
| "model.layers.20.self_attn.rope": false, |
| "model.layers.20.mlp.gate_proj": true, |
| "model.layers.20.mlp.down_proj": true, |
| "model.layers.20.mlp.up_proj": true, |
| "model.layers.20.input_layernorm": false, |
| "model.layers.20.post_attention_layernorm": false, |
| "model.layers.21.self_attn.q_proj": true, |
| "model.layers.21.self_attn.k_proj": true, |
| "model.layers.21.self_attn.v_proj": true, |
| "model.layers.21.self_attn.o_proj": true, |
| "model.layers.21.self_attn.rope": false, |
| "model.layers.21.mlp.gate_proj": true, |
| "model.layers.21.mlp.down_proj": true, |
| "model.layers.21.mlp.up_proj": true, |
| "model.layers.21.input_layernorm": false, |
| "model.layers.21.post_attention_layernorm": false, |
| "model.layers.22.self_attn.q_proj": true, |
| "model.layers.22.self_attn.k_proj": true, |
| "model.layers.22.self_attn.v_proj": true, |
| "model.layers.22.self_attn.o_proj": true, |
| "model.layers.22.self_attn.rope": false, |
| "model.layers.22.mlp.gate_proj": true, |
| "model.layers.22.mlp.down_proj": true, |
| "model.layers.22.mlp.up_proj": true, |
| "model.layers.22.input_layernorm": false, |
| "model.layers.22.post_attention_layernorm": false, |
| "model.layers.23.self_attn.q_proj": true, |
| "model.layers.23.self_attn.k_proj": true, |
| "model.layers.23.self_attn.v_proj": true, |
| "model.layers.23.self_attn.o_proj": true, |
| "model.layers.23.self_attn.rope": false, |
| "model.layers.23.mlp.gate_proj": true, |
| "model.layers.23.mlp.down_proj": true, |
| "model.layers.23.mlp.up_proj": true, |
| "model.layers.23.input_layernorm": false, |
| "model.layers.23.post_attention_layernorm": false, |
| "model.layers.24.self_attn.q_proj": true, |
| "model.layers.24.self_attn.k_proj": true, |
| "model.layers.24.self_attn.v_proj": true, |
| "model.layers.24.self_attn.o_proj": true, |
| "model.layers.24.self_attn.rope": false, |
| "model.layers.24.mlp.gate_proj": true, |
| "model.layers.24.mlp.down_proj": true, |
| "model.layers.24.mlp.up_proj": true, |
| "model.layers.24.input_layernorm": false, |
| "model.layers.24.post_attention_layernorm": false, |
| "model.layers.25.self_attn.q_proj": true, |
| "model.layers.25.self_attn.k_proj": true, |
| "model.layers.25.self_attn.v_proj": true, |
| "model.layers.25.self_attn.o_proj": true, |
| "model.layers.25.self_attn.rope": false, |
| "model.layers.25.mlp.gate_proj": true, |
| "model.layers.25.mlp.down_proj": true, |
| "model.layers.25.mlp.up_proj": true, |
| "model.layers.25.input_layernorm": false, |
| "model.layers.25.post_attention_layernorm": false, |
| "model.layers.26.self_attn.q_proj": true, |
| "model.layers.26.self_attn.k_proj": true, |
| "model.layers.26.self_attn.v_proj": true, |
| "model.layers.26.self_attn.o_proj": true, |
| "model.layers.26.self_attn.rope": false, |
| "model.layers.26.mlp.gate_proj": true, |
| "model.layers.26.mlp.down_proj": true, |
| "model.layers.26.mlp.up_proj": true, |
| "model.layers.26.input_layernorm": false, |
| "model.layers.26.post_attention_layernorm": false, |
| "model.layers.27.self_attn.q_proj": true, |
| "model.layers.27.self_attn.k_proj": true, |
| "model.layers.27.self_attn.v_proj": true, |
| "model.layers.27.self_attn.o_proj": true, |
| "model.layers.27.self_attn.rope": false, |
| "model.layers.27.mlp.gate_proj": true, |
| "model.layers.27.mlp.down_proj": true, |
| "model.layers.27.mlp.up_proj": true, |
| "model.layers.27.input_layernorm": false, |
| "model.layers.27.post_attention_layernorm": false, |
| "model.norm": false, |
| "lm_head": true |
| }, |
| "quantization_config": { |
| "group_size": 64, |
| "bits": 4, |
| "model.embed_tokens": true, |
| "model.layers.0.self_attn.q_proj": true, |
| "model.layers.0.self_attn.k_proj": true, |
| "model.layers.0.self_attn.v_proj": true, |
| "model.layers.0.self_attn.o_proj": true, |
| "model.layers.0.self_attn.rope": false, |
| "model.layers.0.mlp.gate_proj": true, |
| "model.layers.0.mlp.down_proj": true, |
| "model.layers.0.mlp.up_proj": true, |
| "model.layers.0.input_layernorm": false, |
| "model.layers.0.post_attention_layernorm": false, |
| "model.layers.1.self_attn.q_proj": true, |
| "model.layers.1.self_attn.k_proj": true, |
| "model.layers.1.self_attn.v_proj": true, |
| "model.layers.1.self_attn.o_proj": true, |
| "model.layers.1.self_attn.rope": false, |
| "model.layers.1.mlp.gate_proj": true, |
| "model.layers.1.mlp.down_proj": true, |
| "model.layers.1.mlp.up_proj": true, |
| "model.layers.1.input_layernorm": false, |
| "model.layers.1.post_attention_layernorm": false, |
| "model.layers.2.self_attn.q_proj": true, |
| "model.layers.2.self_attn.k_proj": true, |
| "model.layers.2.self_attn.v_proj": true, |
| "model.layers.2.self_attn.o_proj": true, |
| "model.layers.2.self_attn.rope": false, |
| "model.layers.2.mlp.gate_proj": true, |
| "model.layers.2.mlp.down_proj": true, |
| "model.layers.2.mlp.up_proj": true, |
| "model.layers.2.input_layernorm": false, |
| "model.layers.2.post_attention_layernorm": false, |
| "model.layers.3.self_attn.q_proj": true, |
| "model.layers.3.self_attn.k_proj": true, |
| "model.layers.3.self_attn.v_proj": true, |
| "model.layers.3.self_attn.o_proj": true, |
| "model.layers.3.self_attn.rope": false, |
| "model.layers.3.mlp.gate_proj": true, |
| "model.layers.3.mlp.down_proj": true, |
| "model.layers.3.mlp.up_proj": true, |
| "model.layers.3.input_layernorm": false, |
| "model.layers.3.post_attention_layernorm": false, |
| "model.layers.4.self_attn.q_proj": true, |
| "model.layers.4.self_attn.k_proj": true, |
| "model.layers.4.self_attn.v_proj": true, |
| "model.layers.4.self_attn.o_proj": true, |
| "model.layers.4.self_attn.rope": false, |
| "model.layers.4.mlp.gate_proj": true, |
| "model.layers.4.mlp.down_proj": true, |
| "model.layers.4.mlp.up_proj": true, |
| "model.layers.4.input_layernorm": false, |
| "model.layers.4.post_attention_layernorm": false, |
| "model.layers.5.self_attn.q_proj": true, |
| "model.layers.5.self_attn.k_proj": true, |
| "model.layers.5.self_attn.v_proj": true, |
| "model.layers.5.self_attn.o_proj": true, |
| "model.layers.5.self_attn.rope": false, |
| "model.layers.5.mlp.gate_proj": true, |
| "model.layers.5.mlp.down_proj": true, |
| "model.layers.5.mlp.up_proj": true, |
| "model.layers.5.input_layernorm": false, |
| "model.layers.5.post_attention_layernorm": false, |
| "model.layers.6.self_attn.q_proj": true, |
| "model.layers.6.self_attn.k_proj": true, |
| "model.layers.6.self_attn.v_proj": true, |
| "model.layers.6.self_attn.o_proj": true, |
| "model.layers.6.self_attn.rope": false, |
| "model.layers.6.mlp.gate_proj": true, |
| "model.layers.6.mlp.down_proj": true, |
| "model.layers.6.mlp.up_proj": true, |
| "model.layers.6.input_layernorm": false, |
| "model.layers.6.post_attention_layernorm": false, |
| "model.layers.7.self_attn.q_proj": true, |
| "model.layers.7.self_attn.k_proj": true, |
| "model.layers.7.self_attn.v_proj": true, |
| "model.layers.7.self_attn.o_proj": true, |
| "model.layers.7.self_attn.rope": false, |
| "model.layers.7.mlp.gate_proj": true, |
| "model.layers.7.mlp.down_proj": true, |
| "model.layers.7.mlp.up_proj": true, |
| "model.layers.7.input_layernorm": false, |
| "model.layers.7.post_attention_layernorm": false, |
| "model.layers.8.self_attn.q_proj": true, |
| "model.layers.8.self_attn.k_proj": true, |
| "model.layers.8.self_attn.v_proj": true, |
| "model.layers.8.self_attn.o_proj": true, |
| "model.layers.8.self_attn.rope": false, |
| "model.layers.8.mlp.gate_proj": true, |
| "model.layers.8.mlp.down_proj": true, |
| "model.layers.8.mlp.up_proj": true, |
| "model.layers.8.input_layernorm": false, |
| "model.layers.8.post_attention_layernorm": false, |
| "model.layers.9.self_attn.q_proj": true, |
| "model.layers.9.self_attn.k_proj": true, |
| "model.layers.9.self_attn.v_proj": true, |
| "model.layers.9.self_attn.o_proj": true, |
| "model.layers.9.self_attn.rope": false, |
| "model.layers.9.mlp.gate_proj": true, |
| "model.layers.9.mlp.down_proj": true, |
| "model.layers.9.mlp.up_proj": true, |
| "model.layers.9.input_layernorm": false, |
| "model.layers.9.post_attention_layernorm": false, |
| "model.layers.10.self_attn.q_proj": true, |
| "model.layers.10.self_attn.k_proj": true, |
| "model.layers.10.self_attn.v_proj": true, |
| "model.layers.10.self_attn.o_proj": true, |
| "model.layers.10.self_attn.rope": false, |
| "model.layers.10.mlp.gate_proj": true, |
| "model.layers.10.mlp.down_proj": true, |
| "model.layers.10.mlp.up_proj": true, |
| "model.layers.10.input_layernorm": false, |
| "model.layers.10.post_attention_layernorm": false, |
| "model.layers.11.self_attn.q_proj": true, |
| "model.layers.11.self_attn.k_proj": true, |
| "model.layers.11.self_attn.v_proj": true, |
| "model.layers.11.self_attn.o_proj": true, |
| "model.layers.11.self_attn.rope": false, |
| "model.layers.11.mlp.gate_proj": true, |
| "model.layers.11.mlp.down_proj": true, |
| "model.layers.11.mlp.up_proj": true, |
| "model.layers.11.input_layernorm": false, |
| "model.layers.11.post_attention_layernorm": false, |
| "model.layers.12.self_attn.q_proj": true, |
| "model.layers.12.self_attn.k_proj": true, |
| "model.layers.12.self_attn.v_proj": true, |
| "model.layers.12.self_attn.o_proj": true, |
| "model.layers.12.self_attn.rope": false, |
| "model.layers.12.mlp.gate_proj": true, |
| "model.layers.12.mlp.down_proj": true, |
| "model.layers.12.mlp.up_proj": true, |
| "model.layers.12.input_layernorm": false, |
| "model.layers.12.post_attention_layernorm": false, |
| "model.layers.13.self_attn.q_proj": true, |
| "model.layers.13.self_attn.k_proj": true, |
| "model.layers.13.self_attn.v_proj": true, |
| "model.layers.13.self_attn.o_proj": true, |
| "model.layers.13.self_attn.rope": false, |
| "model.layers.13.mlp.gate_proj": true, |
| "model.layers.13.mlp.down_proj": true, |
| "model.layers.13.mlp.up_proj": true, |
| "model.layers.13.input_layernorm": false, |
| "model.layers.13.post_attention_layernorm": false, |
| "model.layers.14.self_attn.q_proj": true, |
| "model.layers.14.self_attn.k_proj": true, |
| "model.layers.14.self_attn.v_proj": true, |
| "model.layers.14.self_attn.o_proj": true, |
| "model.layers.14.self_attn.rope": false, |
| "model.layers.14.mlp.gate_proj": true, |
| "model.layers.14.mlp.down_proj": true, |
| "model.layers.14.mlp.up_proj": true, |
| "model.layers.14.input_layernorm": false, |
| "model.layers.14.post_attention_layernorm": false, |
| "model.layers.15.self_attn.q_proj": true, |
| "model.layers.15.self_attn.k_proj": true, |
| "model.layers.15.self_attn.v_proj": true, |
| "model.layers.15.self_attn.o_proj": true, |
| "model.layers.15.self_attn.rope": false, |
| "model.layers.15.mlp.gate_proj": true, |
| "model.layers.15.mlp.down_proj": true, |
| "model.layers.15.mlp.up_proj": true, |
| "model.layers.15.input_layernorm": false, |
| "model.layers.15.post_attention_layernorm": false, |
| "model.layers.16.self_attn.q_proj": true, |
| "model.layers.16.self_attn.k_proj": true, |
| "model.layers.16.self_attn.v_proj": true, |
| "model.layers.16.self_attn.o_proj": true, |
| "model.layers.16.self_attn.rope": false, |
| "model.layers.16.mlp.gate_proj": true, |
| "model.layers.16.mlp.down_proj": true, |
| "model.layers.16.mlp.up_proj": true, |
| "model.layers.16.input_layernorm": false, |
| "model.layers.16.post_attention_layernorm": false, |
| "model.layers.17.self_attn.q_proj": true, |
| "model.layers.17.self_attn.k_proj": true, |
| "model.layers.17.self_attn.v_proj": true, |
| "model.layers.17.self_attn.o_proj": true, |
| "model.layers.17.self_attn.rope": false, |
| "model.layers.17.mlp.gate_proj": true, |
| "model.layers.17.mlp.down_proj": true, |
| "model.layers.17.mlp.up_proj": true, |
| "model.layers.17.input_layernorm": false, |
| "model.layers.17.post_attention_layernorm": false, |
| "model.layers.18.self_attn.q_proj": true, |
| "model.layers.18.self_attn.k_proj": true, |
| "model.layers.18.self_attn.v_proj": true, |
| "model.layers.18.self_attn.o_proj": true, |
| "model.layers.18.self_attn.rope": false, |
| "model.layers.18.mlp.gate_proj": true, |
| "model.layers.18.mlp.down_proj": true, |
| "model.layers.18.mlp.up_proj": true, |
| "model.layers.18.input_layernorm": false, |
| "model.layers.18.post_attention_layernorm": false, |
| "model.layers.19.self_attn.q_proj": true, |
| "model.layers.19.self_attn.k_proj": true, |
| "model.layers.19.self_attn.v_proj": true, |
| "model.layers.19.self_attn.o_proj": true, |
| "model.layers.19.self_attn.rope": false, |
| "model.layers.19.mlp.gate_proj": true, |
| "model.layers.19.mlp.down_proj": true, |
| "model.layers.19.mlp.up_proj": true, |
| "model.layers.19.input_layernorm": false, |
| "model.layers.19.post_attention_layernorm": false, |
| "model.layers.20.self_attn.q_proj": true, |
| "model.layers.20.self_attn.k_proj": true, |
| "model.layers.20.self_attn.v_proj": true, |
| "model.layers.20.self_attn.o_proj": true, |
| "model.layers.20.self_attn.rope": false, |
| "model.layers.20.mlp.gate_proj": true, |
| "model.layers.20.mlp.down_proj": true, |
| "model.layers.20.mlp.up_proj": true, |
| "model.layers.20.input_layernorm": false, |
| "model.layers.20.post_attention_layernorm": false, |
| "model.layers.21.self_attn.q_proj": true, |
| "model.layers.21.self_attn.k_proj": true, |
| "model.layers.21.self_attn.v_proj": true, |
| "model.layers.21.self_attn.o_proj": true, |
| "model.layers.21.self_attn.rope": false, |
| "model.layers.21.mlp.gate_proj": true, |
| "model.layers.21.mlp.down_proj": true, |
| "model.layers.21.mlp.up_proj": true, |
| "model.layers.21.input_layernorm": false, |
| "model.layers.21.post_attention_layernorm": false, |
| "model.layers.22.self_attn.q_proj": true, |
| "model.layers.22.self_attn.k_proj": true, |
| "model.layers.22.self_attn.v_proj": true, |
| "model.layers.22.self_attn.o_proj": true, |
| "model.layers.22.self_attn.rope": false, |
| "model.layers.22.mlp.gate_proj": true, |
| "model.layers.22.mlp.down_proj": true, |
| "model.layers.22.mlp.up_proj": true, |
| "model.layers.22.input_layernorm": false, |
| "model.layers.22.post_attention_layernorm": false, |
| "model.layers.23.self_attn.q_proj": true, |
| "model.layers.23.self_attn.k_proj": true, |
| "model.layers.23.self_attn.v_proj": true, |
| "model.layers.23.self_attn.o_proj": true, |
| "model.layers.23.self_attn.rope": false, |
| "model.layers.23.mlp.gate_proj": true, |
| "model.layers.23.mlp.down_proj": true, |
| "model.layers.23.mlp.up_proj": true, |
| "model.layers.23.input_layernorm": false, |
| "model.layers.23.post_attention_layernorm": false, |
| "model.layers.24.self_attn.q_proj": true, |
| "model.layers.24.self_attn.k_proj": true, |
| "model.layers.24.self_attn.v_proj": true, |
| "model.layers.24.self_attn.o_proj": true, |
| "model.layers.24.self_attn.rope": false, |
| "model.layers.24.mlp.gate_proj": true, |
| "model.layers.24.mlp.down_proj": true, |
| "model.layers.24.mlp.up_proj": true, |
| "model.layers.24.input_layernorm": false, |
| "model.layers.24.post_attention_layernorm": false, |
| "model.layers.25.self_attn.q_proj": true, |
| "model.layers.25.self_attn.k_proj": true, |
| "model.layers.25.self_attn.v_proj": true, |
| "model.layers.25.self_attn.o_proj": true, |
| "model.layers.25.self_attn.rope": false, |
| "model.layers.25.mlp.gate_proj": true, |
| "model.layers.25.mlp.down_proj": true, |
| "model.layers.25.mlp.up_proj": true, |
| "model.layers.25.input_layernorm": false, |
| "model.layers.25.post_attention_layernorm": false, |
| "model.layers.26.self_attn.q_proj": true, |
| "model.layers.26.self_attn.k_proj": true, |
| "model.layers.26.self_attn.v_proj": true, |
| "model.layers.26.self_attn.o_proj": true, |
| "model.layers.26.self_attn.rope": false, |
| "model.layers.26.mlp.gate_proj": true, |
| "model.layers.26.mlp.down_proj": true, |
| "model.layers.26.mlp.up_proj": true, |
| "model.layers.26.input_layernorm": false, |
| "model.layers.26.post_attention_layernorm": false, |
| "model.layers.27.self_attn.q_proj": true, |
| "model.layers.27.self_attn.k_proj": true, |
| "model.layers.27.self_attn.v_proj": true, |
| "model.layers.27.self_attn.o_proj": true, |
| "model.layers.27.self_attn.rope": false, |
| "model.layers.27.mlp.gate_proj": true, |
| "model.layers.27.mlp.down_proj": true, |
| "model.layers.27.mlp.up_proj": true, |
| "model.layers.27.input_layernorm": false, |
| "model.layers.27.post_attention_layernorm": false, |
| "model.norm": false, |
| "lm_head": true |
| }, |
| "rms_norm_eps": 1e-06, |
| "rope_scaling": null, |
| "rope_theta": 1000000.0, |
| "sliding_window": null, |
| "tie_word_embeddings": false, |
| "torch_dtype": "float32", |
| "transformers_version": "4.46.3", |
| "use_cache": true, |
| "use_sliding_window": false, |
| "vocab_size": 152064 |
| } |
