{ "architectures": [ "VoxtralForConditionalGeneration" ], "audio_config": { "activation_dropout": 0.0, "activation_function": "gelu", "attention_dropout": 0.0, "dropout": 0.0, "head_dim": 64, "hidden_size": 1280, "initializer_range": 0.02, "intermediate_size": 5120, "layerdrop": 0.0, "max_source_positions": 1500, "model_type": "voxtral_encoder", "num_attention_heads": 20, "num_hidden_layers": 32, "num_key_value_heads": 20, "num_mel_bins": 128, "scale_embedding": false, "vocab_size": 51866 }, "audio_token_id": 24, "hidden_size": 3072, "model_type": "voxtral", "projector_hidden_act": "gelu", "quantization": { "group_size": 64, "bits": 8, "audio_tower.conv1": false, "audio_tower.conv2": false, "audio_tower.embed_positions": false, "audio_tower.layers.0.self_attn.k_proj": true, "audio_tower.layers.0.self_attn.v_proj": true, "audio_tower.layers.0.self_attn.q_proj": true, "audio_tower.layers.0.self_attn.out_proj": true, "audio_tower.layers.0.self_attn_layer_norm": false, "audio_tower.layers.0.activation_fn": false, "audio_tower.layers.0.fc1": true, "audio_tower.layers.0.fc2": true, "audio_tower.layers.0.final_layer_norm": false, "audio_tower.layers.1.self_attn.k_proj": true, "audio_tower.layers.1.self_attn.v_proj": true, "audio_tower.layers.1.self_attn.q_proj": true, "audio_tower.layers.1.self_attn.out_proj": true, "audio_tower.layers.1.self_attn_layer_norm": false, "audio_tower.layers.1.activation_fn": false, "audio_tower.layers.1.fc1": true, "audio_tower.layers.1.fc2": true, "audio_tower.layers.1.final_layer_norm": false, "audio_tower.layers.2.self_attn.k_proj": true, "audio_tower.layers.2.self_attn.v_proj": true, "audio_tower.layers.2.self_attn.q_proj": true, "audio_tower.layers.2.self_attn.out_proj": true, "audio_tower.layers.2.self_attn_layer_norm": false, "audio_tower.layers.2.activation_fn": false, "audio_tower.layers.2.fc1": true, "audio_tower.layers.2.fc2": true, "audio_tower.layers.2.final_layer_norm": false, "audio_tower.layers.3.self_attn.k_proj": true, "audio_tower.layers.3.self_attn.v_proj": true, "audio_tower.layers.3.self_attn.q_proj": true, "audio_tower.layers.3.self_attn.out_proj": true, "audio_tower.layers.3.self_attn_layer_norm": false, "audio_tower.layers.3.activation_fn": false, "audio_tower.layers.3.fc1": true, "audio_tower.layers.3.fc2": true, "audio_tower.layers.3.final_layer_norm": false, "audio_tower.layers.4.self_attn.k_proj": true, "audio_tower.layers.4.self_attn.v_proj": true, "audio_tower.layers.4.self_attn.q_proj": true, "audio_tower.layers.4.self_attn.out_proj": true, "audio_tower.layers.4.self_attn_layer_norm": false, "audio_tower.layers.4.activation_fn": false, "audio_tower.layers.4.fc1": true, "audio_tower.layers.4.fc2": true, "audio_tower.layers.4.final_layer_norm": false, "audio_tower.layers.5.self_attn.k_proj": true, "audio_tower.layers.5.self_attn.v_proj": true, "audio_tower.layers.5.self_attn.q_proj": true, "audio_tower.layers.5.self_attn.out_proj": true, "audio_tower.layers.5.self_attn_layer_norm": false, "audio_tower.layers.5.activation_fn": false, "audio_tower.layers.5.fc1": true, "audio_tower.layers.5.fc2": true, "audio_tower.layers.5.final_layer_norm": false, "audio_tower.layers.6.self_attn.k_proj": true, "audio_tower.layers.6.self_attn.v_proj": true, "audio_tower.layers.6.self_attn.q_proj": true, "audio_tower.layers.6.self_attn.out_proj": true, "audio_tower.layers.6.self_attn_layer_norm": false, "audio_tower.layers.6.activation_fn": false, "audio_tower.layers.6.fc1": true, "audio_tower.layers.6.fc2": true, "audio_tower.layers.6.final_layer_norm": false, "audio_tower.layers.7.self_attn.k_proj": true, "audio_tower.layers.7.self_attn.v_proj": true, "audio_tower.layers.7.self_attn.q_proj": true, "audio_tower.layers.7.self_attn.out_proj": true, "audio_tower.layers.7.self_attn_layer_norm": false, "audio_tower.layers.7.activation_fn": false, "audio_tower.layers.7.fc1": true, "audio_tower.layers.7.fc2": true, "audio_tower.layers.7.final_layer_norm": false, "audio_tower.layers.8.self_attn.k_proj": true, "audio_tower.layers.8.self_attn.v_proj": true, "audio_tower.layers.8.self_attn.q_proj": true, "audio_tower.layers.8.self_attn.out_proj": true, "audio_tower.layers.8.self_attn_layer_norm": false, "audio_tower.layers.8.activation_fn": false, "audio_tower.layers.8.fc1": true, "audio_tower.layers.8.fc2": true, "audio_tower.layers.8.final_layer_norm": false, "audio_tower.layers.9.self_attn.k_proj": true, "audio_tower.layers.9.self_attn.v_proj": true, "audio_tower.layers.9.self_attn.q_proj": true, "audio_tower.layers.9.self_attn.out_proj": true, "audio_tower.layers.9.self_attn_layer_norm": false, "audio_tower.layers.9.activation_fn": false, "audio_tower.layers.9.fc1": true, "audio_tower.layers.9.fc2": true, "audio_tower.layers.9.final_layer_norm": false, "audio_tower.layers.10.self_attn.k_proj": true, "audio_tower.layers.10.self_attn.v_proj": true, "audio_tower.layers.10.self_attn.q_proj": true, "audio_tower.layers.10.self_attn.out_proj": true, "audio_tower.layers.10.self_attn_layer_norm": false, "audio_tower.layers.10.activation_fn": false, "audio_tower.layers.10.fc1": true, "audio_tower.layers.10.fc2": true, "audio_tower.layers.10.final_layer_norm": false, "audio_tower.layers.11.self_attn.k_proj": true, "audio_tower.layers.11.self_attn.v_proj": true, "audio_tower.layers.11.self_attn.q_proj": true, "audio_tower.layers.11.self_attn.out_proj": true, "audio_tower.layers.11.self_attn_layer_norm": false, "audio_tower.layers.11.activation_fn": false, "audio_tower.layers.11.fc1": true, "audio_tower.layers.11.fc2": true, "audio_tower.layers.11.final_layer_norm": false, "audio_tower.layers.12.self_attn.k_proj": true, "audio_tower.layers.12.self_attn.v_proj": true, "audio_tower.layers.12.self_attn.q_proj": true, "audio_tower.layers.12.self_attn.out_proj": true, "audio_tower.layers.12.self_attn_layer_norm": false, "audio_tower.layers.12.activation_fn": false, "audio_tower.layers.12.fc1": true, "audio_tower.layers.12.fc2": true, "audio_tower.layers.12.final_layer_norm": false, "audio_tower.layers.13.self_attn.k_proj": true, "audio_tower.layers.13.self_attn.v_proj": true, "audio_tower.layers.13.self_attn.q_proj": true, "audio_tower.layers.13.self_attn.out_proj": true, "audio_tower.layers.13.self_attn_layer_norm": false, "audio_tower.layers.13.activation_fn": false, "audio_tower.layers.13.fc1": true, "audio_tower.layers.13.fc2": true, "audio_tower.layers.13.final_layer_norm": false, "audio_tower.layers.14.self_attn.k_proj": true, "audio_tower.layers.14.self_attn.v_proj": true, "audio_tower.layers.14.self_attn.q_proj": true, "audio_tower.layers.14.self_attn.out_proj": true, "audio_tower.layers.14.self_attn_layer_norm": false, "audio_tower.layers.14.activation_fn": false, "audio_tower.layers.14.fc1": true, "audio_tower.layers.14.fc2": true, "audio_tower.layers.14.final_layer_norm": false, "audio_tower.layers.15.self_attn.k_proj": true, "audio_tower.layers.15.self_attn.v_proj": true, "audio_tower.layers.15.self_attn.q_proj": true, "audio_tower.layers.15.self_attn.out_proj": true, "audio_tower.layers.15.self_attn_layer_norm": false, "audio_tower.layers.15.activation_fn": false, "audio_tower.layers.15.fc1": true, "audio_tower.layers.15.fc2": true, "audio_tower.layers.15.final_layer_norm": false, "audio_tower.layers.16.self_attn.k_proj": true, "audio_tower.layers.16.self_attn.v_proj": true, "audio_tower.layers.16.self_attn.q_proj": true, "audio_tower.layers.16.self_attn.out_proj": true, "audio_tower.layers.16.self_attn_layer_norm": false, "audio_tower.layers.16.activation_fn": false, "audio_tower.layers.16.fc1": true, "audio_tower.layers.16.fc2": true, "audio_tower.layers.16.final_layer_norm": false, "audio_tower.layers.17.self_attn.k_proj": true, "audio_tower.layers.17.self_attn.v_proj": true, "audio_tower.layers.17.self_attn.q_proj": true, "audio_tower.layers.17.self_attn.out_proj": true, "audio_tower.layers.17.self_attn_layer_norm": false, "audio_tower.layers.17.activation_fn": false, "audio_tower.layers.17.fc1": true, "audio_tower.layers.17.fc2": true, "audio_tower.layers.17.final_layer_norm": false, "audio_tower.layers.18.self_attn.k_proj": true, "audio_tower.layers.18.self_attn.v_proj": true, "audio_tower.layers.18.self_attn.q_proj": true, "audio_tower.layers.18.self_attn.out_proj": true, "audio_tower.layers.18.self_attn_layer_norm": false, "audio_tower.layers.18.activation_fn": false, "audio_tower.layers.18.fc1": true, "audio_tower.layers.18.fc2": true, "audio_tower.layers.18.final_layer_norm": false, "audio_tower.layers.19.self_attn.k_proj": true, "audio_tower.layers.19.self_attn.v_proj": true, "audio_tower.layers.19.self_attn.q_proj": true, "audio_tower.layers.19.self_attn.out_proj": true, "audio_tower.layers.19.self_attn_layer_norm": false, "audio_tower.layers.19.activation_fn": false, "audio_tower.layers.19.fc1": true, "audio_tower.layers.19.fc2": true, "audio_tower.layers.19.final_layer_norm": false, "audio_tower.layers.20.self_attn.k_proj": true, "audio_tower.layers.20.self_attn.v_proj": true, "audio_tower.layers.20.self_attn.q_proj": true, "audio_tower.layers.20.self_attn.out_proj": true, "audio_tower.layers.20.self_attn_layer_norm": false, "audio_tower.layers.20.activation_fn": false, "audio_tower.layers.20.fc1": true, "audio_tower.layers.20.fc2": true, "audio_tower.layers.20.final_layer_norm": false, "audio_tower.layers.21.self_attn.k_proj": true, "audio_tower.layers.21.self_attn.v_proj": true, "audio_tower.layers.21.self_attn.q_proj": true, "audio_tower.layers.21.self_attn.out_proj": true, "audio_tower.layers.21.self_attn_layer_norm": false, "audio_tower.layers.21.activation_fn": false, "audio_tower.layers.21.fc1": true, "audio_tower.layers.21.fc2": true, "audio_tower.layers.21.final_layer_norm": false, "audio_tower.layers.22.self_attn.k_proj": true, "audio_tower.layers.22.self_attn.v_proj": true, "audio_tower.layers.22.self_attn.q_proj": true, "audio_tower.layers.22.self_attn.out_proj": true, "audio_tower.layers.22.self_attn_layer_norm": false, "audio_tower.layers.22.activation_fn": false, "audio_tower.layers.22.fc1": true, "audio_tower.layers.22.fc2": true, "audio_tower.layers.22.final_layer_norm": false, "audio_tower.layers.23.self_attn.k_proj": true, "audio_tower.layers.23.self_attn.v_proj": true, "audio_tower.layers.23.self_attn.q_proj": true, "audio_tower.layers.23.self_attn.out_proj": true, "audio_tower.layers.23.self_attn_layer_norm": false, "audio_tower.layers.23.activation_fn": false, "audio_tower.layers.23.fc1": true, "audio_tower.layers.23.fc2": true, "audio_tower.layers.23.final_layer_norm": false, "audio_tower.layers.24.self_attn.k_proj": true, "audio_tower.layers.24.self_attn.v_proj": true, "audio_tower.layers.24.self_attn.q_proj": true, "audio_tower.layers.24.self_attn.out_proj": true, "audio_tower.layers.24.self_attn_layer_norm": false, "audio_tower.layers.24.activation_fn": false, "audio_tower.layers.24.fc1": true, "audio_tower.layers.24.fc2": true, "audio_tower.layers.24.final_layer_norm": false, "audio_tower.layers.25.self_attn.k_proj": true, "audio_tower.layers.25.self_attn.v_proj": true, "audio_tower.layers.25.self_attn.q_proj": true, "audio_tower.layers.25.self_attn.out_proj": true, "audio_tower.layers.25.self_attn_layer_norm": false, "audio_tower.layers.25.activation_fn": false, "audio_tower.layers.25.fc1": true, "audio_tower.layers.25.fc2": true, "audio_tower.layers.25.final_layer_norm": false, "audio_tower.layers.26.self_attn.k_proj": true, "audio_tower.layers.26.self_attn.v_proj": true, "audio_tower.layers.26.self_attn.q_proj": true, "audio_tower.layers.26.self_attn.out_proj": true, "audio_tower.layers.26.self_attn_layer_norm": false, "audio_tower.layers.26.activation_fn": false, "audio_tower.layers.26.fc1": true, "audio_tower.layers.26.fc2": true, "audio_tower.layers.26.final_layer_norm": false, "audio_tower.layers.27.self_attn.k_proj": true, "audio_tower.layers.27.self_attn.v_proj": true, "audio_tower.layers.27.self_attn.q_proj": true, "audio_tower.layers.27.self_attn.out_proj": true, "audio_tower.layers.27.self_attn_layer_norm": false, "audio_tower.layers.27.activation_fn": false, "audio_tower.layers.27.fc1": true, "audio_tower.layers.27.fc2": true, "audio_tower.layers.27.final_layer_norm": false, "audio_tower.layers.28.self_attn.k_proj": true, "audio_tower.layers.28.self_attn.v_proj": true, "audio_tower.layers.28.self_attn.q_proj": true, "audio_tower.layers.28.self_attn.out_proj": true, "audio_tower.layers.28.self_attn_layer_norm": false, "audio_tower.layers.28.activation_fn": false, "audio_tower.layers.28.fc1": true, "audio_tower.layers.28.fc2": true, "audio_tower.layers.28.final_layer_norm": false, "audio_tower.layers.29.self_attn.k_proj": true, "audio_tower.layers.29.self_attn.v_proj": true, "audio_tower.layers.29.self_attn.q_proj": true, "audio_tower.layers.29.self_attn.out_proj": true, "audio_tower.layers.29.self_attn_layer_norm": false, "audio_tower.layers.29.activation_fn": false, "audio_tower.layers.29.fc1": true, "audio_tower.layers.29.fc2": true, "audio_tower.layers.29.final_layer_norm": false, "audio_tower.layers.30.self_attn.k_proj": true, "audio_tower.layers.30.self_attn.v_proj": true, "audio_tower.layers.30.self_attn.q_proj": true, "audio_tower.layers.30.self_attn.out_proj": true, "audio_tower.layers.30.self_attn_layer_norm": false, "audio_tower.layers.30.activation_fn": false, "audio_tower.layers.30.fc1": true, "audio_tower.layers.30.fc2": true, "audio_tower.layers.30.final_layer_norm": false, "audio_tower.layers.31.self_attn.k_proj": true, "audio_tower.layers.31.self_attn.v_proj": true, "audio_tower.layers.31.self_attn.q_proj": true, "audio_tower.layers.31.self_attn.out_proj": true, "audio_tower.layers.31.self_attn_layer_norm": false, "audio_tower.layers.31.activation_fn": false, "audio_tower.layers.31.fc1": true, "audio_tower.layers.31.fc2": true, "audio_tower.layers.31.final_layer_norm": false, "audio_tower.layer_norm": false, "multi_modal_projector.linear_1": true, "multi_modal_projector.act": false, "multi_modal_projector.linear_2": true, "language_model.embed_tokens": true, "language_model.layers.0.self_attn.q_proj": true, "language_model.layers.0.self_attn.k_proj": true, "language_model.layers.0.self_attn.v_proj": true, "language_model.layers.0.self_attn.o_proj": true, "language_model.layers.0.self_attn.rope": false, "language_model.layers.0.mlp.gate_proj": true, "language_model.layers.0.mlp.up_proj": true, "language_model.layers.0.mlp.down_proj": true, "language_model.layers.0.input_layernorm": false, "language_model.layers.0.post_attention_layernorm": false, "language_model.layers.1.self_attn.q_proj": true, "language_model.layers.1.self_attn.k_proj": true, "language_model.layers.1.self_attn.v_proj": true, "language_model.layers.1.self_attn.o_proj": true, "language_model.layers.1.self_attn.rope": false, "language_model.layers.1.mlp.gate_proj": true, "language_model.layers.1.mlp.up_proj": true, "language_model.layers.1.mlp.down_proj": true, "language_model.layers.1.input_layernorm": false, "language_model.layers.1.post_attention_layernorm": false, "language_model.layers.2.self_attn.q_proj": true, "language_model.layers.2.self_attn.k_proj": true, "language_model.layers.2.self_attn.v_proj": true, "language_model.layers.2.self_attn.o_proj": true, "language_model.layers.2.self_attn.rope": false, "language_model.layers.2.mlp.gate_proj": true, "language_model.layers.2.mlp.up_proj": true, "language_model.layers.2.mlp.down_proj": true, "language_model.layers.2.input_layernorm": false, "language_model.layers.2.post_attention_layernorm": false, "language_model.layers.3.self_attn.q_proj": true, "language_model.layers.3.self_attn.k_proj": true, "language_model.layers.3.self_attn.v_proj": true, "language_model.layers.3.self_attn.o_proj": true, "language_model.layers.3.self_attn.rope": false, "language_model.layers.3.mlp.gate_proj": true, "language_model.layers.3.mlp.up_proj": true, "language_model.layers.3.mlp.down_proj": true, "language_model.layers.3.input_layernorm": false, "language_model.layers.3.post_attention_layernorm": false, "language_model.layers.4.self_attn.q_proj": true, "language_model.layers.4.self_attn.k_proj": true, "language_model.layers.4.self_attn.v_proj": true, "language_model.layers.4.self_attn.o_proj": true, "language_model.layers.4.self_attn.rope": false, "language_model.layers.4.mlp.gate_proj": true, "language_model.layers.4.mlp.up_proj": true, "language_model.layers.4.mlp.down_proj": true, "language_model.layers.4.input_layernorm": false, "language_model.layers.4.post_attention_layernorm": false, "language_model.layers.5.self_attn.q_proj": true, "language_model.layers.5.self_attn.k_proj": true, "language_model.layers.5.self_attn.v_proj": true, "language_model.layers.5.self_attn.o_proj": true, "language_model.layers.5.self_attn.rope": false, "language_model.layers.5.mlp.gate_proj": true, "language_model.layers.5.mlp.up_proj": true, "language_model.layers.5.mlp.down_proj": true, "language_model.layers.5.input_layernorm": false, "language_model.layers.5.post_attention_layernorm": false, "language_model.layers.6.self_attn.q_proj": true, "language_model.layers.6.self_attn.k_proj": true, "language_model.layers.6.self_attn.v_proj": true, "language_model.layers.6.self_attn.o_proj": true, "language_model.layers.6.self_attn.rope": false, "language_model.layers.6.mlp.gate_proj": true, "language_model.layers.6.mlp.up_proj": true, "language_model.layers.6.mlp.down_proj": true, "language_model.layers.6.input_layernorm": false, "language_model.layers.6.post_attention_layernorm": false, "language_model.layers.7.self_attn.q_proj": true, "language_model.layers.7.self_attn.k_proj": true, "language_model.layers.7.self_attn.v_proj": true, "language_model.layers.7.self_attn.o_proj": true, "language_model.layers.7.self_attn.rope": false, "language_model.layers.7.mlp.gate_proj": true, "language_model.layers.7.mlp.up_proj": true, "language_model.layers.7.mlp.down_proj": true, "language_model.layers.7.input_layernorm": false, "language_model.layers.7.post_attention_layernorm": false, "language_model.layers.8.self_attn.q_proj": true, "language_model.layers.8.self_attn.k_proj": true, "language_model.layers.8.self_attn.v_proj": true, "language_model.layers.8.self_attn.o_proj": true, "language_model.layers.8.self_attn.rope": false, "language_model.layers.8.mlp.gate_proj": true, "language_model.layers.8.mlp.up_proj": true, "language_model.layers.8.mlp.down_proj": true, "language_model.layers.8.input_layernorm": false, "language_model.layers.8.post_attention_layernorm": false, "language_model.layers.9.self_attn.q_proj": true, "language_model.layers.9.self_attn.k_proj": true, "language_model.layers.9.self_attn.v_proj": true, "language_model.layers.9.self_attn.o_proj": true, "language_model.layers.9.self_attn.rope": false, "language_model.layers.9.mlp.gate_proj": true, "language_model.layers.9.mlp.up_proj": true, "language_model.layers.9.mlp.down_proj": true, "language_model.layers.9.input_layernorm": false, "language_model.layers.9.post_attention_layernorm": false, "language_model.layers.10.self_attn.q_proj": true, "language_model.layers.10.self_attn.k_proj": true, "language_model.layers.10.self_attn.v_proj": true, "language_model.layers.10.self_attn.o_proj": true, "language_model.layers.10.self_attn.rope": false, "language_model.layers.10.mlp.gate_proj": true, "language_model.layers.10.mlp.up_proj": true, "language_model.layers.10.mlp.down_proj": true, "language_model.layers.10.input_layernorm": false, "language_model.layers.10.post_attention_layernorm": false, "language_model.layers.11.self_attn.q_proj": true, "language_model.layers.11.self_attn.k_proj": true, "language_model.layers.11.self_attn.v_proj": true, "language_model.layers.11.self_attn.o_proj": true, "language_model.layers.11.self_attn.rope": false, "language_model.layers.11.mlp.gate_proj": true, "language_model.layers.11.mlp.up_proj": true, "language_model.layers.11.mlp.down_proj": true, "language_model.layers.11.input_layernorm": false, "language_model.layers.11.post_attention_layernorm": false, "language_model.layers.12.self_attn.q_proj": true, "language_model.layers.12.self_attn.k_proj": true, "language_model.layers.12.self_attn.v_proj": true, "language_model.layers.12.self_attn.o_proj": true, "language_model.layers.12.self_attn.rope": false, "language_model.layers.12.mlp.gate_proj": true, "language_model.layers.12.mlp.up_proj": true, "language_model.layers.12.mlp.down_proj": true, "language_model.layers.12.input_layernorm": false, "language_model.layers.12.post_attention_layernorm": false, "language_model.layers.13.self_attn.q_proj": true, "language_model.layers.13.self_attn.k_proj": true, "language_model.layers.13.self_attn.v_proj": true, "language_model.layers.13.self_attn.o_proj": true, "language_model.layers.13.self_attn.rope": false, "language_model.layers.13.mlp.gate_proj": true, "language_model.layers.13.mlp.up_proj": true, "language_model.layers.13.mlp.down_proj": true, "language_model.layers.13.input_layernorm": false, "language_model.layers.13.post_attention_layernorm": false, "language_model.layers.14.self_attn.q_proj": true, "language_model.layers.14.self_attn.k_proj": true, "language_model.layers.14.self_attn.v_proj": true, "language_model.layers.14.self_attn.o_proj": true, "language_model.layers.14.self_attn.rope": false, "language_model.layers.14.mlp.gate_proj": true, "language_model.layers.14.mlp.up_proj": true, "language_model.layers.14.mlp.down_proj": true, "language_model.layers.14.input_layernorm": false, "language_model.layers.14.post_attention_layernorm": false, "language_model.layers.15.self_attn.q_proj": true, "language_model.layers.15.self_attn.k_proj": true, "language_model.layers.15.self_attn.v_proj": true, "language_model.layers.15.self_attn.o_proj": true, "language_model.layers.15.self_attn.rope": false, "language_model.layers.15.mlp.gate_proj": true, "language_model.layers.15.mlp.up_proj": true, "language_model.layers.15.mlp.down_proj": true, "language_model.layers.15.input_layernorm": false, "language_model.layers.15.post_attention_layernorm": false, "language_model.layers.16.self_attn.q_proj": true, "language_model.layers.16.self_attn.k_proj": true, "language_model.layers.16.self_attn.v_proj": true, "language_model.layers.16.self_attn.o_proj": true, "language_model.layers.16.self_attn.rope": false, "language_model.layers.16.mlp.gate_proj": true, "language_model.layers.16.mlp.up_proj": true, "language_model.layers.16.mlp.down_proj": true, "language_model.layers.16.input_layernorm": false, "language_model.layers.16.post_attention_layernorm": false, "language_model.layers.17.self_attn.q_proj": true, "language_model.layers.17.self_attn.k_proj": true, "language_model.layers.17.self_attn.v_proj": true, "language_model.layers.17.self_attn.o_proj": true, "language_model.layers.17.self_attn.rope": false, "language_model.layers.17.mlp.gate_proj": true, "language_model.layers.17.mlp.up_proj": true, "language_model.layers.17.mlp.down_proj": true, "language_model.layers.17.input_layernorm": false, "language_model.layers.17.post_attention_layernorm": false, "language_model.layers.18.self_attn.q_proj": true, "language_model.layers.18.self_attn.k_proj": true, "language_model.layers.18.self_attn.v_proj": true, "language_model.layers.18.self_attn.o_proj": true, "language_model.layers.18.self_attn.rope": false, "language_model.layers.18.mlp.gate_proj": true, "language_model.layers.18.mlp.up_proj": true, "language_model.layers.18.mlp.down_proj": true, "language_model.layers.18.input_layernorm": false, "language_model.layers.18.post_attention_layernorm": false, "language_model.layers.19.self_attn.q_proj": true, "language_model.layers.19.self_attn.k_proj": true, "language_model.layers.19.self_attn.v_proj": true, "language_model.layers.19.self_attn.o_proj": true, "language_model.layers.19.self_attn.rope": false, "language_model.layers.19.mlp.gate_proj": true, "language_model.layers.19.mlp.up_proj": true, "language_model.layers.19.mlp.down_proj": true, "language_model.layers.19.input_layernorm": false, "language_model.layers.19.post_attention_layernorm": false, "language_model.layers.20.self_attn.q_proj": true, "language_model.layers.20.self_attn.k_proj": true, "language_model.layers.20.self_attn.v_proj": true, "language_model.layers.20.self_attn.o_proj": true, "language_model.layers.20.self_attn.rope": false, "language_model.layers.20.mlp.gate_proj": true, "language_model.layers.20.mlp.up_proj": true, "language_model.layers.20.mlp.down_proj": true, "language_model.layers.20.input_layernorm": false, "language_model.layers.20.post_attention_layernorm": false, "language_model.layers.21.self_attn.q_proj": true, "language_model.layers.21.self_attn.k_proj": true, "language_model.layers.21.self_attn.v_proj": true, "language_model.layers.21.self_attn.o_proj": true, "language_model.layers.21.self_attn.rope": false, "language_model.layers.21.mlp.gate_proj": true, "language_model.layers.21.mlp.up_proj": true, "language_model.layers.21.mlp.down_proj": true, "language_model.layers.21.input_layernorm": false, "language_model.layers.21.post_attention_layernorm": false, "language_model.layers.22.self_attn.q_proj": true, "language_model.layers.22.self_attn.k_proj": true, "language_model.layers.22.self_attn.v_proj": true, "language_model.layers.22.self_attn.o_proj": true, "language_model.layers.22.self_attn.rope": false, "language_model.layers.22.mlp.gate_proj": true, "language_model.layers.22.mlp.up_proj": true, "language_model.layers.22.mlp.down_proj": true, "language_model.layers.22.input_layernorm": false, "language_model.layers.22.post_attention_layernorm": false, "language_model.layers.23.self_attn.q_proj": true, "language_model.layers.23.self_attn.k_proj": true, "language_model.layers.23.self_attn.v_proj": true, "language_model.layers.23.self_attn.o_proj": true, "language_model.layers.23.self_attn.rope": false, "language_model.layers.23.mlp.gate_proj": true, "language_model.layers.23.mlp.up_proj": true, "language_model.layers.23.mlp.down_proj": true, "language_model.layers.23.input_layernorm": false, "language_model.layers.23.post_attention_layernorm": false, "language_model.layers.24.self_attn.q_proj": true, "language_model.layers.24.self_attn.k_proj": true, "language_model.layers.24.self_attn.v_proj": true, "language_model.layers.24.self_attn.o_proj": true, "language_model.layers.24.self_attn.rope": false, "language_model.layers.24.mlp.gate_proj": true, "language_model.layers.24.mlp.up_proj": true, "language_model.layers.24.mlp.down_proj": true, "language_model.layers.24.input_layernorm": false, "language_model.layers.24.post_attention_layernorm": false, "language_model.layers.25.self_attn.q_proj": true, "language_model.layers.25.self_attn.k_proj": true, "language_model.layers.25.self_attn.v_proj": true, "language_model.layers.25.self_attn.o_proj": true, "language_model.layers.25.self_attn.rope": false, "language_model.layers.25.mlp.gate_proj": true, "language_model.layers.25.mlp.up_proj": true, "language_model.layers.25.mlp.down_proj": true, "language_model.layers.25.input_layernorm": false, "language_model.layers.25.post_attention_layernorm": false, "language_model.layers.26.self_attn.q_proj": true, "language_model.layers.26.self_attn.k_proj": true, "language_model.layers.26.self_attn.v_proj": true, "language_model.layers.26.self_attn.o_proj": true, "language_model.layers.26.self_attn.rope": false, "language_model.layers.26.mlp.gate_proj": true, "language_model.layers.26.mlp.up_proj": true, "language_model.layers.26.mlp.down_proj": true, "language_model.layers.26.input_layernorm": false, "language_model.layers.26.post_attention_layernorm": false, "language_model.layers.27.self_attn.q_proj": true, "language_model.layers.27.self_attn.k_proj": true, "language_model.layers.27.self_attn.v_proj": true, "language_model.layers.27.self_attn.o_proj": true, "language_model.layers.27.self_attn.rope": false, "language_model.layers.27.mlp.gate_proj": true, "language_model.layers.27.mlp.up_proj": true, "language_model.layers.27.mlp.down_proj": true, "language_model.layers.27.input_layernorm": false, "language_model.layers.27.post_attention_layernorm": false, "language_model.layers.28.self_attn.q_proj": true, "language_model.layers.28.self_attn.k_proj": true, "language_model.layers.28.self_attn.v_proj": true, "language_model.layers.28.self_attn.o_proj": true, "language_model.layers.28.self_attn.rope": false, "language_model.layers.28.mlp.gate_proj": true, "language_model.layers.28.mlp.up_proj": true, "language_model.layers.28.mlp.down_proj": true, "language_model.layers.28.input_layernorm": false, "language_model.layers.28.post_attention_layernorm": false, "language_model.layers.29.self_attn.q_proj": true, "language_model.layers.29.self_attn.k_proj": true, "language_model.layers.29.self_attn.v_proj": true, "language_model.layers.29.self_attn.o_proj": true, "language_model.layers.29.self_attn.rope": false, "language_model.layers.29.mlp.gate_proj": true, "language_model.layers.29.mlp.up_proj": true, "language_model.layers.29.mlp.down_proj": true, "language_model.layers.29.input_layernorm": false, "language_model.layers.29.post_attention_layernorm": false, "language_model.norm": false, "embed_tokens": true, "lm_head": true }, "quantization_config": { "group_size": 64, "bits": 8, "audio_tower.conv1": false, "audio_tower.conv2": false, "audio_tower.embed_positions": false, "audio_tower.layers.0.self_attn.k_proj": true, "audio_tower.layers.0.self_attn.v_proj": true, "audio_tower.layers.0.self_attn.q_proj": true, "audio_tower.layers.0.self_attn.out_proj": true, "audio_tower.layers.0.self_attn_layer_norm": false, "audio_tower.layers.0.activation_fn": false, "audio_tower.layers.0.fc1": true, "audio_tower.layers.0.fc2": true, "audio_tower.layers.0.final_layer_norm": false, "audio_tower.layers.1.self_attn.k_proj": true, "audio_tower.layers.1.self_attn.v_proj": true, "audio_tower.layers.1.self_attn.q_proj": true, "audio_tower.layers.1.self_attn.out_proj": true, "audio_tower.layers.1.self_attn_layer_norm": false, "audio_tower.layers.1.activation_fn": false, "audio_tower.layers.1.fc1": true, "audio_tower.layers.1.fc2": true, "audio_tower.layers.1.final_layer_norm": false, "audio_tower.layers.2.self_attn.k_proj": true, "audio_tower.layers.2.self_attn.v_proj": true, "audio_tower.layers.2.self_attn.q_proj": true, "audio_tower.layers.2.self_attn.out_proj": true, "audio_tower.layers.2.self_attn_layer_norm": false, "audio_tower.layers.2.activation_fn": false, "audio_tower.layers.2.fc1": true, "audio_tower.layers.2.fc2": true, "audio_tower.layers.2.final_layer_norm": false, "audio_tower.layers.3.self_attn.k_proj": true, "audio_tower.layers.3.self_attn.v_proj": true, "audio_tower.layers.3.self_attn.q_proj": true, "audio_tower.layers.3.self_attn.out_proj": true, "audio_tower.layers.3.self_attn_layer_norm": false, "audio_tower.layers.3.activation_fn": false, "audio_tower.layers.3.fc1": true, "audio_tower.layers.3.fc2": true, "audio_tower.layers.3.final_layer_norm": false, "audio_tower.layers.4.self_attn.k_proj": true, "audio_tower.layers.4.self_attn.v_proj": true, "audio_tower.layers.4.self_attn.q_proj": true, "audio_tower.layers.4.self_attn.out_proj": true, "audio_tower.layers.4.self_attn_layer_norm": false, "audio_tower.layers.4.activation_fn": false, "audio_tower.layers.4.fc1": true, "audio_tower.layers.4.fc2": true, "audio_tower.layers.4.final_layer_norm": false, "audio_tower.layers.5.self_attn.k_proj": true, "audio_tower.layers.5.self_attn.v_proj": true, "audio_tower.layers.5.self_attn.q_proj": true, "audio_tower.layers.5.self_attn.out_proj": true, "audio_tower.layers.5.self_attn_layer_norm": false, "audio_tower.layers.5.activation_fn": false, "audio_tower.layers.5.fc1": true, "audio_tower.layers.5.fc2": true, "audio_tower.layers.5.final_layer_norm": false, "audio_tower.layers.6.self_attn.k_proj": true, "audio_tower.layers.6.self_attn.v_proj": true, "audio_tower.layers.6.self_attn.q_proj": true, "audio_tower.layers.6.self_attn.out_proj": true, "audio_tower.layers.6.self_attn_layer_norm": false, "audio_tower.layers.6.activation_fn": false, "audio_tower.layers.6.fc1": true, "audio_tower.layers.6.fc2": true, "audio_tower.layers.6.final_layer_norm": false, "audio_tower.layers.7.self_attn.k_proj": true, "audio_tower.layers.7.self_attn.v_proj": true, "audio_tower.layers.7.self_attn.q_proj": true, "audio_tower.layers.7.self_attn.out_proj": true, "audio_tower.layers.7.self_attn_layer_norm": false, "audio_tower.layers.7.activation_fn": false, "audio_tower.layers.7.fc1": true, "audio_tower.layers.7.fc2": true, "audio_tower.layers.7.final_layer_norm": false, "audio_tower.layers.8.self_attn.k_proj": true, "audio_tower.layers.8.self_attn.v_proj": true, "audio_tower.layers.8.self_attn.q_proj": true, "audio_tower.layers.8.self_attn.out_proj": true, "audio_tower.layers.8.self_attn_layer_norm": false, "audio_tower.layers.8.activation_fn": false, "audio_tower.layers.8.fc1": true, "audio_tower.layers.8.fc2": true, "audio_tower.layers.8.final_layer_norm": false, "audio_tower.layers.9.self_attn.k_proj": true, "audio_tower.layers.9.self_attn.v_proj": true, "audio_tower.layers.9.self_attn.q_proj": true, "audio_tower.layers.9.self_attn.out_proj": true, "audio_tower.layers.9.self_attn_layer_norm": false, "audio_tower.layers.9.activation_fn": false, "audio_tower.layers.9.fc1": true, "audio_tower.layers.9.fc2": true, "audio_tower.layers.9.final_layer_norm": false, "audio_tower.layers.10.self_attn.k_proj": true, "audio_tower.layers.10.self_attn.v_proj": true, "audio_tower.layers.10.self_attn.q_proj": true, "audio_tower.layers.10.self_attn.out_proj": true, "audio_tower.layers.10.self_attn_layer_norm": false, "audio_tower.layers.10.activation_fn": false, "audio_tower.layers.10.fc1": true, "audio_tower.layers.10.fc2": true, "audio_tower.layers.10.final_layer_norm": false, "audio_tower.layers.11.self_attn.k_proj": true, "audio_tower.layers.11.self_attn.v_proj": true, "audio_tower.layers.11.self_attn.q_proj": true, "audio_tower.layers.11.self_attn.out_proj": true, "audio_tower.layers.11.self_attn_layer_norm": false, "audio_tower.layers.11.activation_fn": false, "audio_tower.layers.11.fc1": true, "audio_tower.layers.11.fc2": true, "audio_tower.layers.11.final_layer_norm": false, "audio_tower.layers.12.self_attn.k_proj": true, "audio_tower.layers.12.self_attn.v_proj": true, "audio_tower.layers.12.self_attn.q_proj": true, "audio_tower.layers.12.self_attn.out_proj": true, "audio_tower.layers.12.self_attn_layer_norm": false, "audio_tower.layers.12.activation_fn": false, "audio_tower.layers.12.fc1": true, "audio_tower.layers.12.fc2": true, "audio_tower.layers.12.final_layer_norm": false, "audio_tower.layers.13.self_attn.k_proj": true, "audio_tower.layers.13.self_attn.v_proj": true, "audio_tower.layers.13.self_attn.q_proj": true, "audio_tower.layers.13.self_attn.out_proj": true, "audio_tower.layers.13.self_attn_layer_norm": false, "audio_tower.layers.13.activation_fn": false, "audio_tower.layers.13.fc1": true, "audio_tower.layers.13.fc2": true, "audio_tower.layers.13.final_layer_norm": false, "audio_tower.layers.14.self_attn.k_proj": true, "audio_tower.layers.14.self_attn.v_proj": true, "audio_tower.layers.14.self_attn.q_proj": true, "audio_tower.layers.14.self_attn.out_proj": true, "audio_tower.layers.14.self_attn_layer_norm": false, "audio_tower.layers.14.activation_fn": false, "audio_tower.layers.14.fc1": true, "audio_tower.layers.14.fc2": true, "audio_tower.layers.14.final_layer_norm": false, "audio_tower.layers.15.self_attn.k_proj": true, "audio_tower.layers.15.self_attn.v_proj": true, "audio_tower.layers.15.self_attn.q_proj": true, "audio_tower.layers.15.self_attn.out_proj": true, "audio_tower.layers.15.self_attn_layer_norm": false, "audio_tower.layers.15.activation_fn": false, "audio_tower.layers.15.fc1": true, "audio_tower.layers.15.fc2": true, "audio_tower.layers.15.final_layer_norm": false, "audio_tower.layers.16.self_attn.k_proj": true, "audio_tower.layers.16.self_attn.v_proj": true, "audio_tower.layers.16.self_attn.q_proj": true, "audio_tower.layers.16.self_attn.out_proj": true, "audio_tower.layers.16.self_attn_layer_norm": false, "audio_tower.layers.16.activation_fn": false, "audio_tower.layers.16.fc1": true, "audio_tower.layers.16.fc2": true, "audio_tower.layers.16.final_layer_norm": false, "audio_tower.layers.17.self_attn.k_proj": true, "audio_tower.layers.17.self_attn.v_proj": true, "audio_tower.layers.17.self_attn.q_proj": true, "audio_tower.layers.17.self_attn.out_proj": true, "audio_tower.layers.17.self_attn_layer_norm": false, "audio_tower.layers.17.activation_fn": false, "audio_tower.layers.17.fc1": true, "audio_tower.layers.17.fc2": true, "audio_tower.layers.17.final_layer_norm": false, "audio_tower.layers.18.self_attn.k_proj": true, "audio_tower.layers.18.self_attn.v_proj": true, "audio_tower.layers.18.self_attn.q_proj": true, "audio_tower.layers.18.self_attn.out_proj": true, "audio_tower.layers.18.self_attn_layer_norm": false, "audio_tower.layers.18.activation_fn": false, "audio_tower.layers.18.fc1": true, "audio_tower.layers.18.fc2": true, "audio_tower.layers.18.final_layer_norm": false, "audio_tower.layers.19.self_attn.k_proj": true, "audio_tower.layers.19.self_attn.v_proj": true, "audio_tower.layers.19.self_attn.q_proj": true, "audio_tower.layers.19.self_attn.out_proj": true, "audio_tower.layers.19.self_attn_layer_norm": false, "audio_tower.layers.19.activation_fn": false, "audio_tower.layers.19.fc1": true, "audio_tower.layers.19.fc2": true, "audio_tower.layers.19.final_layer_norm": false, "audio_tower.layers.20.self_attn.k_proj": true, "audio_tower.layers.20.self_attn.v_proj": true, "audio_tower.layers.20.self_attn.q_proj": true, "audio_tower.layers.20.self_attn.out_proj": true, "audio_tower.layers.20.self_attn_layer_norm": false, "audio_tower.layers.20.activation_fn": false, "audio_tower.layers.20.fc1": true, "audio_tower.layers.20.fc2": true, "audio_tower.layers.20.final_layer_norm": false, "audio_tower.layers.21.self_attn.k_proj": true, "audio_tower.layers.21.self_attn.v_proj": true, "audio_tower.layers.21.self_attn.q_proj": true, "audio_tower.layers.21.self_attn.out_proj": true, "audio_tower.layers.21.self_attn_layer_norm": false, "audio_tower.layers.21.activation_fn": false, "audio_tower.layers.21.fc1": true, "audio_tower.layers.21.fc2": true, "audio_tower.layers.21.final_layer_norm": false, "audio_tower.layers.22.self_attn.k_proj": true, "audio_tower.layers.22.self_attn.v_proj": true, "audio_tower.layers.22.self_attn.q_proj": true, "audio_tower.layers.22.self_attn.out_proj": true, "audio_tower.layers.22.self_attn_layer_norm": false, "audio_tower.layers.22.activation_fn": false, "audio_tower.layers.22.fc1": true, "audio_tower.layers.22.fc2": true, "audio_tower.layers.22.final_layer_norm": false, "audio_tower.layers.23.self_attn.k_proj": true, "audio_tower.layers.23.self_attn.v_proj": true, "audio_tower.layers.23.self_attn.q_proj": true, "audio_tower.layers.23.self_attn.out_proj": true, "audio_tower.layers.23.self_attn_layer_norm": false, "audio_tower.layers.23.activation_fn": false, "audio_tower.layers.23.fc1": true, "audio_tower.layers.23.fc2": true, "audio_tower.layers.23.final_layer_norm": false, "audio_tower.layers.24.self_attn.k_proj": true, "audio_tower.layers.24.self_attn.v_proj": true, "audio_tower.layers.24.self_attn.q_proj": true, "audio_tower.layers.24.self_attn.out_proj": true, "audio_tower.layers.24.self_attn_layer_norm": false, "audio_tower.layers.24.activation_fn": false, "audio_tower.layers.24.fc1": true, "audio_tower.layers.24.fc2": true, "audio_tower.layers.24.final_layer_norm": false, "audio_tower.layers.25.self_attn.k_proj": true, "audio_tower.layers.25.self_attn.v_proj": true, "audio_tower.layers.25.self_attn.q_proj": true, "audio_tower.layers.25.self_attn.out_proj": true, "audio_tower.layers.25.self_attn_layer_norm": false, "audio_tower.layers.25.activation_fn": false, "audio_tower.layers.25.fc1": true, "audio_tower.layers.25.fc2": true, "audio_tower.layers.25.final_layer_norm": false, "audio_tower.layers.26.self_attn.k_proj": true, "audio_tower.layers.26.self_attn.v_proj": true, "audio_tower.layers.26.self_attn.q_proj": true, "audio_tower.layers.26.self_attn.out_proj": true, "audio_tower.layers.26.self_attn_layer_norm": false, "audio_tower.layers.26.activation_fn": false, "audio_tower.layers.26.fc1": true, "audio_tower.layers.26.fc2": true, "audio_tower.layers.26.final_layer_norm": false, "audio_tower.layers.27.self_attn.k_proj": true, "audio_tower.layers.27.self_attn.v_proj": true, "audio_tower.layers.27.self_attn.q_proj": true, "audio_tower.layers.27.self_attn.out_proj": true, "audio_tower.layers.27.self_attn_layer_norm": false, "audio_tower.layers.27.activation_fn": false, "audio_tower.layers.27.fc1": true, "audio_tower.layers.27.fc2": true, "audio_tower.layers.27.final_layer_norm": false, "audio_tower.layers.28.self_attn.k_proj": true, "audio_tower.layers.28.self_attn.v_proj": true, "audio_tower.layers.28.self_attn.q_proj": true, "audio_tower.layers.28.self_attn.out_proj": true, "audio_tower.layers.28.self_attn_layer_norm": false, "audio_tower.layers.28.activation_fn": false, "audio_tower.layers.28.fc1": true, "audio_tower.layers.28.fc2": true, "audio_tower.layers.28.final_layer_norm": false, "audio_tower.layers.29.self_attn.k_proj": true, "audio_tower.layers.29.self_attn.v_proj": true, "audio_tower.layers.29.self_attn.q_proj": true, "audio_tower.layers.29.self_attn.out_proj": true, "audio_tower.layers.29.self_attn_layer_norm": false, "audio_tower.layers.29.activation_fn": false, "audio_tower.layers.29.fc1": true, "audio_tower.layers.29.fc2": true, "audio_tower.layers.29.final_layer_norm": false, "audio_tower.layers.30.self_attn.k_proj": true, "audio_tower.layers.30.self_attn.v_proj": true, "audio_tower.layers.30.self_attn.q_proj": true, "audio_tower.layers.30.self_attn.out_proj": true, "audio_tower.layers.30.self_attn_layer_norm": false, "audio_tower.layers.30.activation_fn": false, "audio_tower.layers.30.fc1": true, "audio_tower.layers.30.fc2": true, "audio_tower.layers.30.final_layer_norm": false, "audio_tower.layers.31.self_attn.k_proj": true, "audio_tower.layers.31.self_attn.v_proj": true, "audio_tower.layers.31.self_attn.q_proj": true, "audio_tower.layers.31.self_attn.out_proj": true, "audio_tower.layers.31.self_attn_layer_norm": false, "audio_tower.layers.31.activation_fn": false, "audio_tower.layers.31.fc1": true, "audio_tower.layers.31.fc2": true, "audio_tower.layers.31.final_layer_norm": false, "audio_tower.layer_norm": false, "multi_modal_projector.linear_1": true, "multi_modal_projector.act": false, "multi_modal_projector.linear_2": true, "language_model.embed_tokens": true, "language_model.layers.0.self_attn.q_proj": true, "language_model.layers.0.self_attn.k_proj": true, "language_model.layers.0.self_attn.v_proj": true, "language_model.layers.0.self_attn.o_proj": true, "language_model.layers.0.self_attn.rope": false, "language_model.layers.0.mlp.gate_proj": true, "language_model.layers.0.mlp.up_proj": true, "language_model.layers.0.mlp.down_proj": true, "language_model.layers.0.input_layernorm": false, "language_model.layers.0.post_attention_layernorm": false, "language_model.layers.1.self_attn.q_proj": true, "language_model.layers.1.self_attn.k_proj": true, "language_model.layers.1.self_attn.v_proj": true, "language_model.layers.1.self_attn.o_proj": true, "language_model.layers.1.self_attn.rope": false, "language_model.layers.1.mlp.gate_proj": true, "language_model.layers.1.mlp.up_proj": true, "language_model.layers.1.mlp.down_proj": true, "language_model.layers.1.input_layernorm": false, "language_model.layers.1.post_attention_layernorm": false, "language_model.layers.2.self_attn.q_proj": true, "language_model.layers.2.self_attn.k_proj": true, "language_model.layers.2.self_attn.v_proj": true, "language_model.layers.2.self_attn.o_proj": true, "language_model.layers.2.self_attn.rope": false, "language_model.layers.2.mlp.gate_proj": true, "language_model.layers.2.mlp.up_proj": true, "language_model.layers.2.mlp.down_proj": true, "language_model.layers.2.input_layernorm": false, "language_model.layers.2.post_attention_layernorm": false, "language_model.layers.3.self_attn.q_proj": true, "language_model.layers.3.self_attn.k_proj": true, "language_model.layers.3.self_attn.v_proj": true, "language_model.layers.3.self_attn.o_proj": true, "language_model.layers.3.self_attn.rope": false, "language_model.layers.3.mlp.gate_proj": true, "language_model.layers.3.mlp.up_proj": true, "language_model.layers.3.mlp.down_proj": true, "language_model.layers.3.input_layernorm": false, "language_model.layers.3.post_attention_layernorm": false, "language_model.layers.4.self_attn.q_proj": true, "language_model.layers.4.self_attn.k_proj": true, "language_model.layers.4.self_attn.v_proj": true, "language_model.layers.4.self_attn.o_proj": true, "language_model.layers.4.self_attn.rope": false, "language_model.layers.4.mlp.gate_proj": true, "language_model.layers.4.mlp.up_proj": true, "language_model.layers.4.mlp.down_proj": true, "language_model.layers.4.input_layernorm": false, "language_model.layers.4.post_attention_layernorm": false, "language_model.layers.5.self_attn.q_proj": true, "language_model.layers.5.self_attn.k_proj": true, "language_model.layers.5.self_attn.v_proj": true, "language_model.layers.5.self_attn.o_proj": true, "language_model.layers.5.self_attn.rope": false, "language_model.layers.5.mlp.gate_proj": true, "language_model.layers.5.mlp.up_proj": true, "language_model.layers.5.mlp.down_proj": true, "language_model.layers.5.input_layernorm": false, "language_model.layers.5.post_attention_layernorm": false, "language_model.layers.6.self_attn.q_proj": true, "language_model.layers.6.self_attn.k_proj": true, "language_model.layers.6.self_attn.v_proj": true, "language_model.layers.6.self_attn.o_proj": true, "language_model.layers.6.self_attn.rope": false, "language_model.layers.6.mlp.gate_proj": true, "language_model.layers.6.mlp.up_proj": true, "language_model.layers.6.mlp.down_proj": true, "language_model.layers.6.input_layernorm": false, "language_model.layers.6.post_attention_layernorm": false, "language_model.layers.7.self_attn.q_proj": true, "language_model.layers.7.self_attn.k_proj": true, "language_model.layers.7.self_attn.v_proj": true, "language_model.layers.7.self_attn.o_proj": true, "language_model.layers.7.self_attn.rope": false, "language_model.layers.7.mlp.gate_proj": true, "language_model.layers.7.mlp.up_proj": true, "language_model.layers.7.mlp.down_proj": true, "language_model.layers.7.input_layernorm": false, "language_model.layers.7.post_attention_layernorm": false, "language_model.layers.8.self_attn.q_proj": true, "language_model.layers.8.self_attn.k_proj": true, "language_model.layers.8.self_attn.v_proj": true, "language_model.layers.8.self_attn.o_proj": true, "language_model.layers.8.self_attn.rope": false, "language_model.layers.8.mlp.gate_proj": true, "language_model.layers.8.mlp.up_proj": true, "language_model.layers.8.mlp.down_proj": true, "language_model.layers.8.input_layernorm": false, "language_model.layers.8.post_attention_layernorm": false, "language_model.layers.9.self_attn.q_proj": true, "language_model.layers.9.self_attn.k_proj": true, "language_model.layers.9.self_attn.v_proj": true, "language_model.layers.9.self_attn.o_proj": true, "language_model.layers.9.self_attn.rope": false, "language_model.layers.9.mlp.gate_proj": true, "language_model.layers.9.mlp.up_proj": true, "language_model.layers.9.mlp.down_proj": true, "language_model.layers.9.input_layernorm": false, "language_model.layers.9.post_attention_layernorm": false, "language_model.layers.10.self_attn.q_proj": true, "language_model.layers.10.self_attn.k_proj": true, "language_model.layers.10.self_attn.v_proj": true, "language_model.layers.10.self_attn.o_proj": true, "language_model.layers.10.self_attn.rope": false, "language_model.layers.10.mlp.gate_proj": true, "language_model.layers.10.mlp.up_proj": true, "language_model.layers.10.mlp.down_proj": true, "language_model.layers.10.input_layernorm": false, "language_model.layers.10.post_attention_layernorm": false, "language_model.layers.11.self_attn.q_proj": true, "language_model.layers.11.self_attn.k_proj": true, "language_model.layers.11.self_attn.v_proj": true, "language_model.layers.11.self_attn.o_proj": true, "language_model.layers.11.self_attn.rope": false, "language_model.layers.11.mlp.gate_proj": true, "language_model.layers.11.mlp.up_proj": true, "language_model.layers.11.mlp.down_proj": true, "language_model.layers.11.input_layernorm": false, "language_model.layers.11.post_attention_layernorm": false, "language_model.layers.12.self_attn.q_proj": true, "language_model.layers.12.self_attn.k_proj": true, "language_model.layers.12.self_attn.v_proj": true, "language_model.layers.12.self_attn.o_proj": true, "language_model.layers.12.self_attn.rope": false, "language_model.layers.12.mlp.gate_proj": true, "language_model.layers.12.mlp.up_proj": true, "language_model.layers.12.mlp.down_proj": true, "language_model.layers.12.input_layernorm": false, "language_model.layers.12.post_attention_layernorm": false, "language_model.layers.13.self_attn.q_proj": true, "language_model.layers.13.self_attn.k_proj": true, "language_model.layers.13.self_attn.v_proj": true, "language_model.layers.13.self_attn.o_proj": true, "language_model.layers.13.self_attn.rope": false, "language_model.layers.13.mlp.gate_proj": true, "language_model.layers.13.mlp.up_proj": true, "language_model.layers.13.mlp.down_proj": true, "language_model.layers.13.input_layernorm": false, "language_model.layers.13.post_attention_layernorm": false, "language_model.layers.14.self_attn.q_proj": true, "language_model.layers.14.self_attn.k_proj": true, "language_model.layers.14.self_attn.v_proj": true, "language_model.layers.14.self_attn.o_proj": true, "language_model.layers.14.self_attn.rope": false, "language_model.layers.14.mlp.gate_proj": true, "language_model.layers.14.mlp.up_proj": true, "language_model.layers.14.mlp.down_proj": true, "language_model.layers.14.input_layernorm": false, "language_model.layers.14.post_attention_layernorm": false, "language_model.layers.15.self_attn.q_proj": true, "language_model.layers.15.self_attn.k_proj": true, "language_model.layers.15.self_attn.v_proj": true, "language_model.layers.15.self_attn.o_proj": true, "language_model.layers.15.self_attn.rope": false, "language_model.layers.15.mlp.gate_proj": true, "language_model.layers.15.mlp.up_proj": true, "language_model.layers.15.mlp.down_proj": true, "language_model.layers.15.input_layernorm": false, "language_model.layers.15.post_attention_layernorm": false, "language_model.layers.16.self_attn.q_proj": true, "language_model.layers.16.self_attn.k_proj": true, "language_model.layers.16.self_attn.v_proj": true, "language_model.layers.16.self_attn.o_proj": true, "language_model.layers.16.self_attn.rope": false, "language_model.layers.16.mlp.gate_proj": true, "language_model.layers.16.mlp.up_proj": true, "language_model.layers.16.mlp.down_proj": true, "language_model.layers.16.input_layernorm": false, "language_model.layers.16.post_attention_layernorm": false, "language_model.layers.17.self_attn.q_proj": true, "language_model.layers.17.self_attn.k_proj": true, "language_model.layers.17.self_attn.v_proj": true, "language_model.layers.17.self_attn.o_proj": true, "language_model.layers.17.self_attn.rope": false, "language_model.layers.17.mlp.gate_proj": true, "language_model.layers.17.mlp.up_proj": true, "language_model.layers.17.mlp.down_proj": true, "language_model.layers.17.input_layernorm": false, "language_model.layers.17.post_attention_layernorm": false, "language_model.layers.18.self_attn.q_proj": true, "language_model.layers.18.self_attn.k_proj": true, "language_model.layers.18.self_attn.v_proj": true, "language_model.layers.18.self_attn.o_proj": true, "language_model.layers.18.self_attn.rope": false, "language_model.layers.18.mlp.gate_proj": true, "language_model.layers.18.mlp.up_proj": true, "language_model.layers.18.mlp.down_proj": true, "language_model.layers.18.input_layernorm": false, "language_model.layers.18.post_attention_layernorm": false, "language_model.layers.19.self_attn.q_proj": true, "language_model.layers.19.self_attn.k_proj": true, "language_model.layers.19.self_attn.v_proj": true, "language_model.layers.19.self_attn.o_proj": true, "language_model.layers.19.self_attn.rope": false, "language_model.layers.19.mlp.gate_proj": true, "language_model.layers.19.mlp.up_proj": true, "language_model.layers.19.mlp.down_proj": true, "language_model.layers.19.input_layernorm": false, "language_model.layers.19.post_attention_layernorm": false, "language_model.layers.20.self_attn.q_proj": true, "language_model.layers.20.self_attn.k_proj": true, "language_model.layers.20.self_attn.v_proj": true, "language_model.layers.20.self_attn.o_proj": true, "language_model.layers.20.self_attn.rope": false, "language_model.layers.20.mlp.gate_proj": true, "language_model.layers.20.mlp.up_proj": true, "language_model.layers.20.mlp.down_proj": true, "language_model.layers.20.input_layernorm": false, "language_model.layers.20.post_attention_layernorm": false, "language_model.layers.21.self_attn.q_proj": true, "language_model.layers.21.self_attn.k_proj": true, "language_model.layers.21.self_attn.v_proj": true, "language_model.layers.21.self_attn.o_proj": true, "language_model.layers.21.self_attn.rope": false, "language_model.layers.21.mlp.gate_proj": true, "language_model.layers.21.mlp.up_proj": true, "language_model.layers.21.mlp.down_proj": true, "language_model.layers.21.input_layernorm": false, "language_model.layers.21.post_attention_layernorm": false, "language_model.layers.22.self_attn.q_proj": true, "language_model.layers.22.self_attn.k_proj": true, "language_model.layers.22.self_attn.v_proj": true, "language_model.layers.22.self_attn.o_proj": true, "language_model.layers.22.self_attn.rope": false, "language_model.layers.22.mlp.gate_proj": true, "language_model.layers.22.mlp.up_proj": true, "language_model.layers.22.mlp.down_proj": true, "language_model.layers.22.input_layernorm": false, "language_model.layers.22.post_attention_layernorm": false, "language_model.layers.23.self_attn.q_proj": true, "language_model.layers.23.self_attn.k_proj": true, "language_model.layers.23.self_attn.v_proj": true, "language_model.layers.23.self_attn.o_proj": true, "language_model.layers.23.self_attn.rope": false, "language_model.layers.23.mlp.gate_proj": true, "language_model.layers.23.mlp.up_proj": true, "language_model.layers.23.mlp.down_proj": true, "language_model.layers.23.input_layernorm": false, "language_model.layers.23.post_attention_layernorm": false, "language_model.layers.24.self_attn.q_proj": true, "language_model.layers.24.self_attn.k_proj": true, "language_model.layers.24.self_attn.v_proj": true, "language_model.layers.24.self_attn.o_proj": true, "language_model.layers.24.self_attn.rope": false, "language_model.layers.24.mlp.gate_proj": true, "language_model.layers.24.mlp.up_proj": true, "language_model.layers.24.mlp.down_proj": true, "language_model.layers.24.input_layernorm": false, "language_model.layers.24.post_attention_layernorm": false, "language_model.layers.25.self_attn.q_proj": true, "language_model.layers.25.self_attn.k_proj": true, "language_model.layers.25.self_attn.v_proj": true, "language_model.layers.25.self_attn.o_proj": true, "language_model.layers.25.self_attn.rope": false, "language_model.layers.25.mlp.gate_proj": true, "language_model.layers.25.mlp.up_proj": true, "language_model.layers.25.mlp.down_proj": true, "language_model.layers.25.input_layernorm": false, "language_model.layers.25.post_attention_layernorm": false, "language_model.layers.26.self_attn.q_proj": true, "language_model.layers.26.self_attn.k_proj": true, "language_model.layers.26.self_attn.v_proj": true, "language_model.layers.26.self_attn.o_proj": true, "language_model.layers.26.self_attn.rope": false, "language_model.layers.26.mlp.gate_proj": true, "language_model.layers.26.mlp.up_proj": true, "language_model.layers.26.mlp.down_proj": true, "language_model.layers.26.input_layernorm": false, "language_model.layers.26.post_attention_layernorm": false, "language_model.layers.27.self_attn.q_proj": true, "language_model.layers.27.self_attn.k_proj": true, "language_model.layers.27.self_attn.v_proj": true, "language_model.layers.27.self_attn.o_proj": true, "language_model.layers.27.self_attn.rope": false, "language_model.layers.27.mlp.gate_proj": true, "language_model.layers.27.mlp.up_proj": true, "language_model.layers.27.mlp.down_proj": true, "language_model.layers.27.input_layernorm": false, "language_model.layers.27.post_attention_layernorm": false, "language_model.layers.28.self_attn.q_proj": true, "language_model.layers.28.self_attn.k_proj": true, "language_model.layers.28.self_attn.v_proj": true, "language_model.layers.28.self_attn.o_proj": true, "language_model.layers.28.self_attn.rope": false, "language_model.layers.28.mlp.gate_proj": true, "language_model.layers.28.mlp.up_proj": true, "language_model.layers.28.mlp.down_proj": true, "language_model.layers.28.input_layernorm": false, "language_model.layers.28.post_attention_layernorm": false, "language_model.layers.29.self_attn.q_proj": true, "language_model.layers.29.self_attn.k_proj": true, "language_model.layers.29.self_attn.v_proj": true, "language_model.layers.29.self_attn.o_proj": true, "language_model.layers.29.self_attn.rope": false, "language_model.layers.29.mlp.gate_proj": true, "language_model.layers.29.mlp.up_proj": true, "language_model.layers.29.mlp.down_proj": true, "language_model.layers.29.input_layernorm": false, "language_model.layers.29.post_attention_layernorm": false, "language_model.norm": false, "embed_tokens": true, "lm_head": true }, "text_config": { "attention_bias": false, "attention_dropout": 0.0, "head_dim": 128, "hidden_act": "silu", "hidden_size": 3072, "initializer_range": 0.02, "intermediate_size": 8192, "max_position_embeddings": 131072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 30, "num_key_value_heads": 8, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 100000000.0, "sliding_window": null, "use_cache": true, "vocab_size": 131072 }, "torch_dtype": "bfloat16", "transformers_version": "4.54.0.dev0", "vocab_size": 131072 }