{ "add_cross_attention": false, "architectures": [ "Gemma3ForConditionalGeneration" ], "bad_words_ids": null, "begin_suppress_tokens": null, "boi_token_index": 255999, "bos_token_id": null, "chunk_size_feed_forward": 0, "cross_attention_hidden_size": null, "decoder_start_token_id": null, "diversity_penalty": 0.0, "do_sample": false, "early_stopping": false, "encoder_no_repeat_ngram_size": 0, "eoi_token_index": 256000, "eos_token_id": [ 1, 106 ], "exponential_decay_length_penalty": null, "finetuning_task": null, "forced_bos_token_id": null, "forced_eos_token_id": null, "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "image_token_index": 262144, "initializer_range": 0.02, "is_decoder": false, "is_encoder_decoder": false, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "length_penalty": 1.0, "max_length": 20, "min_length": 0, "mm_tokens_per_image": 256, "model_type": "gemma3", "no_repeat_ngram_size": 0, "num_beam_groups": 1, "num_beams": 1, "num_return_sequences": 1, "output_attentions": false, "output_hidden_states": false, "output_scores": false, "pad_token_id": null, "prefix": null, "problem_type": null, "pruned_heads": {}, "quantization": { "group_size": 64, "bits": 6, "vision_tower.vision_model.embeddings.patch_embedding": false, "vision_tower.vision_model.embeddings.position_embedding": false, "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.0.layer_norm1": false, "vision_tower.vision_model.encoder.layers.0.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.0.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.0.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.0.layer_norm2": false, "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.1.layer_norm1": false, "vision_tower.vision_model.encoder.layers.1.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.1.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.1.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.1.layer_norm2": false, "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.2.layer_norm1": false, "vision_tower.vision_model.encoder.layers.2.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.2.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.2.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.2.layer_norm2": false, "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.3.layer_norm1": false, "vision_tower.vision_model.encoder.layers.3.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.3.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.3.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.3.layer_norm2": false, "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.4.layer_norm1": false, "vision_tower.vision_model.encoder.layers.4.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.4.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.4.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.4.layer_norm2": false, "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.5.layer_norm1": false, "vision_tower.vision_model.encoder.layers.5.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.5.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.5.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.5.layer_norm2": false, "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.6.layer_norm1": false, "vision_tower.vision_model.encoder.layers.6.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.6.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.6.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.6.layer_norm2": false, "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.7.layer_norm1": false, "vision_tower.vision_model.encoder.layers.7.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.7.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.7.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.7.layer_norm2": false, "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.8.layer_norm1": false, "vision_tower.vision_model.encoder.layers.8.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.8.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.8.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.8.layer_norm2": false, "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.9.layer_norm1": false, "vision_tower.vision_model.encoder.layers.9.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.9.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.9.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.9.layer_norm2": false, "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.10.layer_norm1": false, "vision_tower.vision_model.encoder.layers.10.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.10.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.10.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.10.layer_norm2": false, "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.11.layer_norm1": false, "vision_tower.vision_model.encoder.layers.11.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.11.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.11.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.11.layer_norm2": false, "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.12.layer_norm1": false, "vision_tower.vision_model.encoder.layers.12.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.12.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.12.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.12.layer_norm2": false, "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.13.layer_norm1": false, "vision_tower.vision_model.encoder.layers.13.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.13.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.13.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.13.layer_norm2": false, "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.14.layer_norm1": false, "vision_tower.vision_model.encoder.layers.14.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.14.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.14.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.14.layer_norm2": false, "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.15.layer_norm1": false, "vision_tower.vision_model.encoder.layers.15.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.15.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.15.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.15.layer_norm2": false, "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.16.layer_norm1": false, "vision_tower.vision_model.encoder.layers.16.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.16.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.16.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.16.layer_norm2": false, "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.17.layer_norm1": false, "vision_tower.vision_model.encoder.layers.17.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.17.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.17.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.17.layer_norm2": false, "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.18.layer_norm1": false, "vision_tower.vision_model.encoder.layers.18.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.18.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.18.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.18.layer_norm2": false, "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.19.layer_norm1": false, "vision_tower.vision_model.encoder.layers.19.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.19.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.19.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.19.layer_norm2": false, "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.20.layer_norm1": false, "vision_tower.vision_model.encoder.layers.20.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.20.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.20.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.20.layer_norm2": false, "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.21.layer_norm1": false, "vision_tower.vision_model.encoder.layers.21.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.21.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.21.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.21.layer_norm2": false, "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.22.layer_norm1": false, "vision_tower.vision_model.encoder.layers.22.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.22.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.22.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.22.layer_norm2": false, "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.23.layer_norm1": false, "vision_tower.vision_model.encoder.layers.23.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.23.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.23.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.23.layer_norm2": false, "vision_tower.vision_model.encoder.layers.24.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.24.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.24.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.24.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.24.layer_norm1": false, "vision_tower.vision_model.encoder.layers.24.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.24.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.24.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.24.layer_norm2": false, "vision_tower.vision_model.encoder.layers.25.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.25.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.25.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.25.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.25.layer_norm1": false, "vision_tower.vision_model.encoder.layers.25.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.25.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.25.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.25.layer_norm2": false, "vision_tower.vision_model.encoder.layers.26.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.26.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.26.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.26.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.26.layer_norm1": false, "vision_tower.vision_model.encoder.layers.26.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.26.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.26.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.26.layer_norm2": false, "vision_tower.vision_model.post_layernorm": false, "language_model.model.embed_tokens": true, "language_model.model.layers.0.self_attn.q_proj": true, "language_model.model.layers.0.self_attn.k_proj": true, "language_model.model.layers.0.self_attn.v_proj": true, "language_model.model.layers.0.self_attn.o_proj": true, "language_model.model.layers.0.self_attn.q_norm": false, "language_model.model.layers.0.self_attn.k_norm": false, "language_model.model.layers.0.self_attn.rope": false, "language_model.model.layers.0.mlp.gate_proj": true, "language_model.model.layers.0.mlp.down_proj": true, "language_model.model.layers.0.mlp.up_proj": true, "language_model.model.layers.0.input_layernorm": false, "language_model.model.layers.0.post_attention_layernorm": false, "language_model.model.layers.0.pre_feedforward_layernorm": false, "language_model.model.layers.0.post_feedforward_layernorm": false, "language_model.model.layers.1.self_attn.q_proj": true, "language_model.model.layers.1.self_attn.k_proj": true, "language_model.model.layers.1.self_attn.v_proj": true, "language_model.model.layers.1.self_attn.o_proj": true, "language_model.model.layers.1.self_attn.q_norm": false, "language_model.model.layers.1.self_attn.k_norm": false, "language_model.model.layers.1.self_attn.rope": false, "language_model.model.layers.1.mlp.gate_proj": true, "language_model.model.layers.1.mlp.down_proj": true, "language_model.model.layers.1.mlp.up_proj": true, "language_model.model.layers.1.input_layernorm": false, "language_model.model.layers.1.post_attention_layernorm": false, "language_model.model.layers.1.pre_feedforward_layernorm": false, "language_model.model.layers.1.post_feedforward_layernorm": false, "language_model.model.layers.2.self_attn.q_proj": true, "language_model.model.layers.2.self_attn.k_proj": true, "language_model.model.layers.2.self_attn.v_proj": true, "language_model.model.layers.2.self_attn.o_proj": true, "language_model.model.layers.2.self_attn.q_norm": false, "language_model.model.layers.2.self_attn.k_norm": false, "language_model.model.layers.2.self_attn.rope": false, "language_model.model.layers.2.mlp.gate_proj": true, "language_model.model.layers.2.mlp.down_proj": true, "language_model.model.layers.2.mlp.up_proj": true, "language_model.model.layers.2.input_layernorm": false, "language_model.model.layers.2.post_attention_layernorm": false, "language_model.model.layers.2.pre_feedforward_layernorm": false, "language_model.model.layers.2.post_feedforward_layernorm": false, "language_model.model.layers.3.self_attn.q_proj": true, "language_model.model.layers.3.self_attn.k_proj": true, "language_model.model.layers.3.self_attn.v_proj": true, "language_model.model.layers.3.self_attn.o_proj": true, "language_model.model.layers.3.self_attn.q_norm": false, "language_model.model.layers.3.self_attn.k_norm": false, "language_model.model.layers.3.self_attn.rope": false, "language_model.model.layers.3.mlp.gate_proj": true, "language_model.model.layers.3.mlp.down_proj": true, "language_model.model.layers.3.mlp.up_proj": true, "language_model.model.layers.3.input_layernorm": false, "language_model.model.layers.3.post_attention_layernorm": false, "language_model.model.layers.3.pre_feedforward_layernorm": false, "language_model.model.layers.3.post_feedforward_layernorm": false, "language_model.model.layers.4.self_attn.q_proj": true, "language_model.model.layers.4.self_attn.k_proj": true, "language_model.model.layers.4.self_attn.v_proj": true, "language_model.model.layers.4.self_attn.o_proj": true, "language_model.model.layers.4.self_attn.q_norm": false, "language_model.model.layers.4.self_attn.k_norm": false, "language_model.model.layers.4.self_attn.rope": false, "language_model.model.layers.4.mlp.gate_proj": true, "language_model.model.layers.4.mlp.down_proj": true, "language_model.model.layers.4.mlp.up_proj": true, "language_model.model.layers.4.input_layernorm": false, "language_model.model.layers.4.post_attention_layernorm": false, "language_model.model.layers.4.pre_feedforward_layernorm": false, "language_model.model.layers.4.post_feedforward_layernorm": false, "language_model.model.layers.5.self_attn.q_proj": true, "language_model.model.layers.5.self_attn.k_proj": true, "language_model.model.layers.5.self_attn.v_proj": true, "language_model.model.layers.5.self_attn.o_proj": true, "language_model.model.layers.5.self_attn.q_norm": false, "language_model.model.layers.5.self_attn.k_norm": false, "language_model.model.layers.5.self_attn.rope": false, "language_model.model.layers.5.mlp.gate_proj": true, "language_model.model.layers.5.mlp.down_proj": true, "language_model.model.layers.5.mlp.up_proj": true, "language_model.model.layers.5.input_layernorm": false, "language_model.model.layers.5.post_attention_layernorm": false, "language_model.model.layers.5.pre_feedforward_layernorm": false, "language_model.model.layers.5.post_feedforward_layernorm": false, "language_model.model.layers.6.self_attn.q_proj": true, "language_model.model.layers.6.self_attn.k_proj": true, "language_model.model.layers.6.self_attn.v_proj": true, "language_model.model.layers.6.self_attn.o_proj": true, "language_model.model.layers.6.self_attn.q_norm": false, "language_model.model.layers.6.self_attn.k_norm": false, "language_model.model.layers.6.self_attn.rope": false, "language_model.model.layers.6.mlp.gate_proj": true, "language_model.model.layers.6.mlp.down_proj": true, "language_model.model.layers.6.mlp.up_proj": true, "language_model.model.layers.6.input_layernorm": false, "language_model.model.layers.6.post_attention_layernorm": false, "language_model.model.layers.6.pre_feedforward_layernorm": false, "language_model.model.layers.6.post_feedforward_layernorm": false, "language_model.model.layers.7.self_attn.q_proj": true, "language_model.model.layers.7.self_attn.k_proj": true, "language_model.model.layers.7.self_attn.v_proj": true, "language_model.model.layers.7.self_attn.o_proj": true, "language_model.model.layers.7.self_attn.q_norm": false, "language_model.model.layers.7.self_attn.k_norm": false, "language_model.model.layers.7.self_attn.rope": false, "language_model.model.layers.7.mlp.gate_proj": true, "language_model.model.layers.7.mlp.down_proj": true, "language_model.model.layers.7.mlp.up_proj": true, "language_model.model.layers.7.input_layernorm": false, "language_model.model.layers.7.post_attention_layernorm": false, "language_model.model.layers.7.pre_feedforward_layernorm": false, "language_model.model.layers.7.post_feedforward_layernorm": false, "language_model.model.layers.8.self_attn.q_proj": true, "language_model.model.layers.8.self_attn.k_proj": true, "language_model.model.layers.8.self_attn.v_proj": true, "language_model.model.layers.8.self_attn.o_proj": true, "language_model.model.layers.8.self_attn.q_norm": false, "language_model.model.layers.8.self_attn.k_norm": false, "language_model.model.layers.8.self_attn.rope": false, "language_model.model.layers.8.mlp.gate_proj": true, "language_model.model.layers.8.mlp.down_proj": true, "language_model.model.layers.8.mlp.up_proj": true, "language_model.model.layers.8.input_layernorm": false, "language_model.model.layers.8.post_attention_layernorm": false, "language_model.model.layers.8.pre_feedforward_layernorm": false, "language_model.model.layers.8.post_feedforward_layernorm": false, "language_model.model.layers.9.self_attn.q_proj": true, "language_model.model.layers.9.self_attn.k_proj": true, "language_model.model.layers.9.self_attn.v_proj": true, "language_model.model.layers.9.self_attn.o_proj": true, "language_model.model.layers.9.self_attn.q_norm": false, "language_model.model.layers.9.self_attn.k_norm": false, "language_model.model.layers.9.self_attn.rope": false, "language_model.model.layers.9.mlp.gate_proj": true, "language_model.model.layers.9.mlp.down_proj": true, "language_model.model.layers.9.mlp.up_proj": true, "language_model.model.layers.9.input_layernorm": false, "language_model.model.layers.9.post_attention_layernorm": false, "language_model.model.layers.9.pre_feedforward_layernorm": false, "language_model.model.layers.9.post_feedforward_layernorm": false, "language_model.model.layers.10.self_attn.q_proj": true, "language_model.model.layers.10.self_attn.k_proj": true, "language_model.model.layers.10.self_attn.v_proj": true, "language_model.model.layers.10.self_attn.o_proj": true, "language_model.model.layers.10.self_attn.q_norm": false, "language_model.model.layers.10.self_attn.k_norm": false, "language_model.model.layers.10.self_attn.rope": false, "language_model.model.layers.10.mlp.gate_proj": true, "language_model.model.layers.10.mlp.down_proj": true, "language_model.model.layers.10.mlp.up_proj": true, "language_model.model.layers.10.input_layernorm": false, "language_model.model.layers.10.post_attention_layernorm": false, "language_model.model.layers.10.pre_feedforward_layernorm": false, "language_model.model.layers.10.post_feedforward_layernorm": false, "language_model.model.layers.11.self_attn.q_proj": true, "language_model.model.layers.11.self_attn.k_proj": true, "language_model.model.layers.11.self_attn.v_proj": true, "language_model.model.layers.11.self_attn.o_proj": true, "language_model.model.layers.11.self_attn.q_norm": false, "language_model.model.layers.11.self_attn.k_norm": false, "language_model.model.layers.11.self_attn.rope": false, "language_model.model.layers.11.mlp.gate_proj": true, "language_model.model.layers.11.mlp.down_proj": true, "language_model.model.layers.11.mlp.up_proj": true, "language_model.model.layers.11.input_layernorm": false, "language_model.model.layers.11.post_attention_layernorm": false, "language_model.model.layers.11.pre_feedforward_layernorm": false, "language_model.model.layers.11.post_feedforward_layernorm": false, "language_model.model.layers.12.self_attn.q_proj": true, "language_model.model.layers.12.self_attn.k_proj": true, "language_model.model.layers.12.self_attn.v_proj": true, "language_model.model.layers.12.self_attn.o_proj": true, "language_model.model.layers.12.self_attn.q_norm": false, "language_model.model.layers.12.self_attn.k_norm": false, "language_model.model.layers.12.self_attn.rope": false, "language_model.model.layers.12.mlp.gate_proj": true, "language_model.model.layers.12.mlp.down_proj": true, "language_model.model.layers.12.mlp.up_proj": true, "language_model.model.layers.12.input_layernorm": false, "language_model.model.layers.12.post_attention_layernorm": false, "language_model.model.layers.12.pre_feedforward_layernorm": false, "language_model.model.layers.12.post_feedforward_layernorm": false, "language_model.model.layers.13.self_attn.q_proj": true, "language_model.model.layers.13.self_attn.k_proj": true, "language_model.model.layers.13.self_attn.v_proj": true, "language_model.model.layers.13.self_attn.o_proj": true, "language_model.model.layers.13.self_attn.q_norm": false, "language_model.model.layers.13.self_attn.k_norm": false, "language_model.model.layers.13.self_attn.rope": false, "language_model.model.layers.13.mlp.gate_proj": true, "language_model.model.layers.13.mlp.down_proj": true, "language_model.model.layers.13.mlp.up_proj": true, "language_model.model.layers.13.input_layernorm": false, "language_model.model.layers.13.post_attention_layernorm": false, "language_model.model.layers.13.pre_feedforward_layernorm": false, "language_model.model.layers.13.post_feedforward_layernorm": false, "language_model.model.layers.14.self_attn.q_proj": true, "language_model.model.layers.14.self_attn.k_proj": true, "language_model.model.layers.14.self_attn.v_proj": true, "language_model.model.layers.14.self_attn.o_proj": true, "language_model.model.layers.14.self_attn.q_norm": false, "language_model.model.layers.14.self_attn.k_norm": false, "language_model.model.layers.14.self_attn.rope": false, "language_model.model.layers.14.mlp.gate_proj": true, "language_model.model.layers.14.mlp.down_proj": true, "language_model.model.layers.14.mlp.up_proj": true, "language_model.model.layers.14.input_layernorm": false, "language_model.model.layers.14.post_attention_layernorm": false, "language_model.model.layers.14.pre_feedforward_layernorm": false, "language_model.model.layers.14.post_feedforward_layernorm": false, "language_model.model.layers.15.self_attn.q_proj": true, "language_model.model.layers.15.self_attn.k_proj": true, "language_model.model.layers.15.self_attn.v_proj": true, "language_model.model.layers.15.self_attn.o_proj": true, "language_model.model.layers.15.self_attn.q_norm": false, "language_model.model.layers.15.self_attn.k_norm": false, "language_model.model.layers.15.self_attn.rope": false, "language_model.model.layers.15.mlp.gate_proj": true, "language_model.model.layers.15.mlp.down_proj": true, "language_model.model.layers.15.mlp.up_proj": true, "language_model.model.layers.15.input_layernorm": false, "language_model.model.layers.15.post_attention_layernorm": false, "language_model.model.layers.15.pre_feedforward_layernorm": false, "language_model.model.layers.15.post_feedforward_layernorm": false, "language_model.model.layers.16.self_attn.q_proj": true, "language_model.model.layers.16.self_attn.k_proj": true, "language_model.model.layers.16.self_attn.v_proj": true, "language_model.model.layers.16.self_attn.o_proj": true, "language_model.model.layers.16.self_attn.q_norm": false, "language_model.model.layers.16.self_attn.k_norm": false, "language_model.model.layers.16.self_attn.rope": false, "language_model.model.layers.16.mlp.gate_proj": true, "language_model.model.layers.16.mlp.down_proj": true, "language_model.model.layers.16.mlp.up_proj": true, "language_model.model.layers.16.input_layernorm": false, "language_model.model.layers.16.post_attention_layernorm": false, "language_model.model.layers.16.pre_feedforward_layernorm": false, "language_model.model.layers.16.post_feedforward_layernorm": false, "language_model.model.layers.17.self_attn.q_proj": true, "language_model.model.layers.17.self_attn.k_proj": true, "language_model.model.layers.17.self_attn.v_proj": true, "language_model.model.layers.17.self_attn.o_proj": true, "language_model.model.layers.17.self_attn.q_norm": false, "language_model.model.layers.17.self_attn.k_norm": false, "language_model.model.layers.17.self_attn.rope": false, "language_model.model.layers.17.mlp.gate_proj": true, "language_model.model.layers.17.mlp.down_proj": true, "language_model.model.layers.17.mlp.up_proj": true, "language_model.model.layers.17.input_layernorm": false, "language_model.model.layers.17.post_attention_layernorm": false, "language_model.model.layers.17.pre_feedforward_layernorm": false, "language_model.model.layers.17.post_feedforward_layernorm": false, "language_model.model.layers.18.self_attn.q_proj": true, "language_model.model.layers.18.self_attn.k_proj": true, "language_model.model.layers.18.self_attn.v_proj": true, "language_model.model.layers.18.self_attn.o_proj": true, "language_model.model.layers.18.self_attn.q_norm": false, "language_model.model.layers.18.self_attn.k_norm": false, "language_model.model.layers.18.self_attn.rope": false, "language_model.model.layers.18.mlp.gate_proj": true, "language_model.model.layers.18.mlp.down_proj": true, "language_model.model.layers.18.mlp.up_proj": true, "language_model.model.layers.18.input_layernorm": false, "language_model.model.layers.18.post_attention_layernorm": false, "language_model.model.layers.18.pre_feedforward_layernorm": false, "language_model.model.layers.18.post_feedforward_layernorm": false, "language_model.model.layers.19.self_attn.q_proj": true, "language_model.model.layers.19.self_attn.k_proj": true, "language_model.model.layers.19.self_attn.v_proj": true, "language_model.model.layers.19.self_attn.o_proj": true, "language_model.model.layers.19.self_attn.q_norm": false, "language_model.model.layers.19.self_attn.k_norm": false, "language_model.model.layers.19.self_attn.rope": false, "language_model.model.layers.19.mlp.gate_proj": true, "language_model.model.layers.19.mlp.down_proj": true, "language_model.model.layers.19.mlp.up_proj": true, "language_model.model.layers.19.input_layernorm": false, "language_model.model.layers.19.post_attention_layernorm": false, "language_model.model.layers.19.pre_feedforward_layernorm": false, "language_model.model.layers.19.post_feedforward_layernorm": false, "language_model.model.layers.20.self_attn.q_proj": true, "language_model.model.layers.20.self_attn.k_proj": true, "language_model.model.layers.20.self_attn.v_proj": true, "language_model.model.layers.20.self_attn.o_proj": true, "language_model.model.layers.20.self_attn.q_norm": false, "language_model.model.layers.20.self_attn.k_norm": false, "language_model.model.layers.20.self_attn.rope": false, "language_model.model.layers.20.mlp.gate_proj": true, "language_model.model.layers.20.mlp.down_proj": true, "language_model.model.layers.20.mlp.up_proj": true, "language_model.model.layers.20.input_layernorm": false, "language_model.model.layers.20.post_attention_layernorm": false, "language_model.model.layers.20.pre_feedforward_layernorm": false, "language_model.model.layers.20.post_feedforward_layernorm": false, "language_model.model.layers.21.self_attn.q_proj": true, "language_model.model.layers.21.self_attn.k_proj": true, "language_model.model.layers.21.self_attn.v_proj": true, "language_model.model.layers.21.self_attn.o_proj": true, "language_model.model.layers.21.self_attn.q_norm": false, "language_model.model.layers.21.self_attn.k_norm": false, "language_model.model.layers.21.self_attn.rope": false, "language_model.model.layers.21.mlp.gate_proj": true, "language_model.model.layers.21.mlp.down_proj": true, "language_model.model.layers.21.mlp.up_proj": true, "language_model.model.layers.21.input_layernorm": false, "language_model.model.layers.21.post_attention_layernorm": false, "language_model.model.layers.21.pre_feedforward_layernorm": false, "language_model.model.layers.21.post_feedforward_layernorm": false, "language_model.model.layers.22.self_attn.q_proj": true, "language_model.model.layers.22.self_attn.k_proj": true, "language_model.model.layers.22.self_attn.v_proj": true, "language_model.model.layers.22.self_attn.o_proj": true, "language_model.model.layers.22.self_attn.q_norm": false, "language_model.model.layers.22.self_attn.k_norm": false, "language_model.model.layers.22.self_attn.rope": false, "language_model.model.layers.22.mlp.gate_proj": true, "language_model.model.layers.22.mlp.down_proj": true, "language_model.model.layers.22.mlp.up_proj": true, "language_model.model.layers.22.input_layernorm": false, "language_model.model.layers.22.post_attention_layernorm": false, "language_model.model.layers.22.pre_feedforward_layernorm": false, "language_model.model.layers.22.post_feedforward_layernorm": false, "language_model.model.layers.23.self_attn.q_proj": true, "language_model.model.layers.23.self_attn.k_proj": true, "language_model.model.layers.23.self_attn.v_proj": true, "language_model.model.layers.23.self_attn.o_proj": true, "language_model.model.layers.23.self_attn.q_norm": false, "language_model.model.layers.23.self_attn.k_norm": false, "language_model.model.layers.23.self_attn.rope": false, "language_model.model.layers.23.mlp.gate_proj": true, "language_model.model.layers.23.mlp.down_proj": true, "language_model.model.layers.23.mlp.up_proj": true, "language_model.model.layers.23.input_layernorm": false, "language_model.model.layers.23.post_attention_layernorm": false, "language_model.model.layers.23.pre_feedforward_layernorm": false, "language_model.model.layers.23.post_feedforward_layernorm": false, "language_model.model.layers.24.self_attn.q_proj": true, "language_model.model.layers.24.self_attn.k_proj": true, "language_model.model.layers.24.self_attn.v_proj": true, "language_model.model.layers.24.self_attn.o_proj": true, "language_model.model.layers.24.self_attn.q_norm": false, "language_model.model.layers.24.self_attn.k_norm": false, "language_model.model.layers.24.self_attn.rope": false, "language_model.model.layers.24.mlp.gate_proj": true, "language_model.model.layers.24.mlp.down_proj": true, "language_model.model.layers.24.mlp.up_proj": true, "language_model.model.layers.24.input_layernorm": false, "language_model.model.layers.24.post_attention_layernorm": false, "language_model.model.layers.24.pre_feedforward_layernorm": false, "language_model.model.layers.24.post_feedforward_layernorm": false, "language_model.model.layers.25.self_attn.q_proj": true, "language_model.model.layers.25.self_attn.k_proj": true, "language_model.model.layers.25.self_attn.v_proj": true, "language_model.model.layers.25.self_attn.o_proj": true, "language_model.model.layers.25.self_attn.q_norm": false, "language_model.model.layers.25.self_attn.k_norm": false, "language_model.model.layers.25.self_attn.rope": false, "language_model.model.layers.25.mlp.gate_proj": true, "language_model.model.layers.25.mlp.down_proj": true, "language_model.model.layers.25.mlp.up_proj": true, "language_model.model.layers.25.input_layernorm": false, "language_model.model.layers.25.post_attention_layernorm": false, "language_model.model.layers.25.pre_feedforward_layernorm": false, "language_model.model.layers.25.post_feedforward_layernorm": false, "language_model.model.layers.26.self_attn.q_proj": true, "language_model.model.layers.26.self_attn.k_proj": true, "language_model.model.layers.26.self_attn.v_proj": true, "language_model.model.layers.26.self_attn.o_proj": true, "language_model.model.layers.26.self_attn.q_norm": false, "language_model.model.layers.26.self_attn.k_norm": false, "language_model.model.layers.26.self_attn.rope": false, "language_model.model.layers.26.mlp.gate_proj": true, "language_model.model.layers.26.mlp.down_proj": true, "language_model.model.layers.26.mlp.up_proj": true, "language_model.model.layers.26.input_layernorm": false, "language_model.model.layers.26.post_attention_layernorm": false, "language_model.model.layers.26.pre_feedforward_layernorm": false, "language_model.model.layers.26.post_feedforward_layernorm": false, "language_model.model.layers.27.self_attn.q_proj": true, "language_model.model.layers.27.self_attn.k_proj": true, "language_model.model.layers.27.self_attn.v_proj": true, "language_model.model.layers.27.self_attn.o_proj": true, "language_model.model.layers.27.self_attn.q_norm": false, "language_model.model.layers.27.self_attn.k_norm": false, "language_model.model.layers.27.self_attn.rope": false, "language_model.model.layers.27.mlp.gate_proj": true, "language_model.model.layers.27.mlp.down_proj": true, "language_model.model.layers.27.mlp.up_proj": true, "language_model.model.layers.27.input_layernorm": false, "language_model.model.layers.27.post_attention_layernorm": false, "language_model.model.layers.27.pre_feedforward_layernorm": false, "language_model.model.layers.27.post_feedforward_layernorm": false, "language_model.model.layers.28.self_attn.q_proj": true, "language_model.model.layers.28.self_attn.k_proj": true, "language_model.model.layers.28.self_attn.v_proj": true, "language_model.model.layers.28.self_attn.o_proj": true, "language_model.model.layers.28.self_attn.q_norm": false, "language_model.model.layers.28.self_attn.k_norm": false, "language_model.model.layers.28.self_attn.rope": false, "language_model.model.layers.28.mlp.gate_proj": true, "language_model.model.layers.28.mlp.down_proj": true, "language_model.model.layers.28.mlp.up_proj": true, "language_model.model.layers.28.input_layernorm": false, "language_model.model.layers.28.post_attention_layernorm": false, "language_model.model.layers.28.pre_feedforward_layernorm": false, "language_model.model.layers.28.post_feedforward_layernorm": false, "language_model.model.layers.29.self_attn.q_proj": true, "language_model.model.layers.29.self_attn.k_proj": true, "language_model.model.layers.29.self_attn.v_proj": true, "language_model.model.layers.29.self_attn.o_proj": true, "language_model.model.layers.29.self_attn.q_norm": false, "language_model.model.layers.29.self_attn.k_norm": false, "language_model.model.layers.29.self_attn.rope": false, "language_model.model.layers.29.mlp.gate_proj": true, "language_model.model.layers.29.mlp.down_proj": true, "language_model.model.layers.29.mlp.up_proj": true, "language_model.model.layers.29.input_layernorm": false, "language_model.model.layers.29.post_attention_layernorm": false, "language_model.model.layers.29.pre_feedforward_layernorm": false, "language_model.model.layers.29.post_feedforward_layernorm": false, "language_model.model.layers.30.self_attn.q_proj": true, "language_model.model.layers.30.self_attn.k_proj": true, "language_model.model.layers.30.self_attn.v_proj": true, "language_model.model.layers.30.self_attn.o_proj": true, "language_model.model.layers.30.self_attn.q_norm": false, "language_model.model.layers.30.self_attn.k_norm": false, "language_model.model.layers.30.self_attn.rope": false, "language_model.model.layers.30.mlp.gate_proj": true, "language_model.model.layers.30.mlp.down_proj": true, "language_model.model.layers.30.mlp.up_proj": true, "language_model.model.layers.30.input_layernorm": false, "language_model.model.layers.30.post_attention_layernorm": false, "language_model.model.layers.30.pre_feedforward_layernorm": false, "language_model.model.layers.30.post_feedforward_layernorm": false, "language_model.model.layers.31.self_attn.q_proj": true, "language_model.model.layers.31.self_attn.k_proj": true, "language_model.model.layers.31.self_attn.v_proj": true, "language_model.model.layers.31.self_attn.o_proj": true, "language_model.model.layers.31.self_attn.q_norm": false, "language_model.model.layers.31.self_attn.k_norm": false, "language_model.model.layers.31.self_attn.rope": false, "language_model.model.layers.31.mlp.gate_proj": true, "language_model.model.layers.31.mlp.down_proj": true, "language_model.model.layers.31.mlp.up_proj": true, "language_model.model.layers.31.input_layernorm": false, "language_model.model.layers.31.post_attention_layernorm": false, "language_model.model.layers.31.pre_feedforward_layernorm": false, "language_model.model.layers.31.post_feedforward_layernorm": false, "language_model.model.layers.32.self_attn.q_proj": true, "language_model.model.layers.32.self_attn.k_proj": true, "language_model.model.layers.32.self_attn.v_proj": true, "language_model.model.layers.32.self_attn.o_proj": true, "language_model.model.layers.32.self_attn.q_norm": false, "language_model.model.layers.32.self_attn.k_norm": false, "language_model.model.layers.32.self_attn.rope": false, "language_model.model.layers.32.mlp.gate_proj": true, "language_model.model.layers.32.mlp.down_proj": true, "language_model.model.layers.32.mlp.up_proj": true, "language_model.model.layers.32.input_layernorm": false, "language_model.model.layers.32.post_attention_layernorm": false, "language_model.model.layers.32.pre_feedforward_layernorm": false, "language_model.model.layers.32.post_feedforward_layernorm": false, "language_model.model.layers.33.self_attn.q_proj": true, "language_model.model.layers.33.self_attn.k_proj": true, "language_model.model.layers.33.self_attn.v_proj": true, "language_model.model.layers.33.self_attn.o_proj": true, "language_model.model.layers.33.self_attn.q_norm": false, "language_model.model.layers.33.self_attn.k_norm": false, "language_model.model.layers.33.self_attn.rope": false, "language_model.model.layers.33.mlp.gate_proj": true, "language_model.model.layers.33.mlp.down_proj": true, "language_model.model.layers.33.mlp.up_proj": true, "language_model.model.layers.33.input_layernorm": false, "language_model.model.layers.33.post_attention_layernorm": false, "language_model.model.layers.33.pre_feedforward_layernorm": false, "language_model.model.layers.33.post_feedforward_layernorm": false, "language_model.model.layers.34.self_attn.q_proj": true, "language_model.model.layers.34.self_attn.k_proj": true, "language_model.model.layers.34.self_attn.v_proj": true, "language_model.model.layers.34.self_attn.o_proj": true, "language_model.model.layers.34.self_attn.q_norm": false, "language_model.model.layers.34.self_attn.k_norm": false, "language_model.model.layers.34.self_attn.rope": false, "language_model.model.layers.34.mlp.gate_proj": true, "language_model.model.layers.34.mlp.down_proj": true, "language_model.model.layers.34.mlp.up_proj": true, "language_model.model.layers.34.input_layernorm": false, "language_model.model.layers.34.post_attention_layernorm": false, "language_model.model.layers.34.pre_feedforward_layernorm": false, "language_model.model.layers.34.post_feedforward_layernorm": false, "language_model.model.layers.35.self_attn.q_proj": true, "language_model.model.layers.35.self_attn.k_proj": true, "language_model.model.layers.35.self_attn.v_proj": true, "language_model.model.layers.35.self_attn.o_proj": true, "language_model.model.layers.35.self_attn.q_norm": false, "language_model.model.layers.35.self_attn.k_norm": false, "language_model.model.layers.35.self_attn.rope": false, "language_model.model.layers.35.mlp.gate_proj": true, "language_model.model.layers.35.mlp.down_proj": true, "language_model.model.layers.35.mlp.up_proj": true, "language_model.model.layers.35.input_layernorm": false, "language_model.model.layers.35.post_attention_layernorm": false, "language_model.model.layers.35.pre_feedforward_layernorm": false, "language_model.model.layers.35.post_feedforward_layernorm": false, "language_model.model.layers.36.self_attn.q_proj": true, "language_model.model.layers.36.self_attn.k_proj": true, "language_model.model.layers.36.self_attn.v_proj": true, "language_model.model.layers.36.self_attn.o_proj": true, "language_model.model.layers.36.self_attn.q_norm": false, "language_model.model.layers.36.self_attn.k_norm": false, "language_model.model.layers.36.self_attn.rope": false, "language_model.model.layers.36.mlp.gate_proj": true, "language_model.model.layers.36.mlp.down_proj": true, "language_model.model.layers.36.mlp.up_proj": true, "language_model.model.layers.36.input_layernorm": false, "language_model.model.layers.36.post_attention_layernorm": false, "language_model.model.layers.36.pre_feedforward_layernorm": false, "language_model.model.layers.36.post_feedforward_layernorm": false, "language_model.model.layers.37.self_attn.q_proj": true, "language_model.model.layers.37.self_attn.k_proj": true, "language_model.model.layers.37.self_attn.v_proj": true, "language_model.model.layers.37.self_attn.o_proj": true, "language_model.model.layers.37.self_attn.q_norm": false, "language_model.model.layers.37.self_attn.k_norm": false, "language_model.model.layers.37.self_attn.rope": false, "language_model.model.layers.37.mlp.gate_proj": true, "language_model.model.layers.37.mlp.down_proj": true, "language_model.model.layers.37.mlp.up_proj": true, "language_model.model.layers.37.input_layernorm": false, "language_model.model.layers.37.post_attention_layernorm": false, "language_model.model.layers.37.pre_feedforward_layernorm": false, "language_model.model.layers.37.post_feedforward_layernorm": false, "language_model.model.layers.38.self_attn.q_proj": true, "language_model.model.layers.38.self_attn.k_proj": true, "language_model.model.layers.38.self_attn.v_proj": true, "language_model.model.layers.38.self_attn.o_proj": true, "language_model.model.layers.38.self_attn.q_norm": false, "language_model.model.layers.38.self_attn.k_norm": false, "language_model.model.layers.38.self_attn.rope": false, "language_model.model.layers.38.mlp.gate_proj": true, "language_model.model.layers.38.mlp.down_proj": true, "language_model.model.layers.38.mlp.up_proj": true, "language_model.model.layers.38.input_layernorm": false, "language_model.model.layers.38.post_attention_layernorm": false, "language_model.model.layers.38.pre_feedforward_layernorm": false, "language_model.model.layers.38.post_feedforward_layernorm": false, "language_model.model.layers.39.self_attn.q_proj": true, "language_model.model.layers.39.self_attn.k_proj": true, "language_model.model.layers.39.self_attn.v_proj": true, "language_model.model.layers.39.self_attn.o_proj": true, "language_model.model.layers.39.self_attn.q_norm": false, "language_model.model.layers.39.self_attn.k_norm": false, "language_model.model.layers.39.self_attn.rope": false, "language_model.model.layers.39.mlp.gate_proj": true, "language_model.model.layers.39.mlp.down_proj": true, "language_model.model.layers.39.mlp.up_proj": true, "language_model.model.layers.39.input_layernorm": false, "language_model.model.layers.39.post_attention_layernorm": false, "language_model.model.layers.39.pre_feedforward_layernorm": false, "language_model.model.layers.39.post_feedforward_layernorm": false, "language_model.model.layers.40.self_attn.q_proj": true, "language_model.model.layers.40.self_attn.k_proj": true, "language_model.model.layers.40.self_attn.v_proj": true, "language_model.model.layers.40.self_attn.o_proj": true, "language_model.model.layers.40.self_attn.q_norm": false, "language_model.model.layers.40.self_attn.k_norm": false, "language_model.model.layers.40.self_attn.rope": false, "language_model.model.layers.40.mlp.gate_proj": true, "language_model.model.layers.40.mlp.down_proj": true, "language_model.model.layers.40.mlp.up_proj": true, "language_model.model.layers.40.input_layernorm": false, "language_model.model.layers.40.post_attention_layernorm": false, "language_model.model.layers.40.pre_feedforward_layernorm": false, "language_model.model.layers.40.post_feedforward_layernorm": false, "language_model.model.layers.41.self_attn.q_proj": true, "language_model.model.layers.41.self_attn.k_proj": true, "language_model.model.layers.41.self_attn.v_proj": true, "language_model.model.layers.41.self_attn.o_proj": true, "language_model.model.layers.41.self_attn.q_norm": false, "language_model.model.layers.41.self_attn.k_norm": false, "language_model.model.layers.41.self_attn.rope": false, "language_model.model.layers.41.mlp.gate_proj": true, "language_model.model.layers.41.mlp.down_proj": true, "language_model.model.layers.41.mlp.up_proj": true, "language_model.model.layers.41.input_layernorm": false, "language_model.model.layers.41.post_attention_layernorm": false, "language_model.model.layers.41.pre_feedforward_layernorm": false, "language_model.model.layers.41.post_feedforward_layernorm": false, "language_model.model.layers.42.self_attn.q_proj": true, "language_model.model.layers.42.self_attn.k_proj": true, "language_model.model.layers.42.self_attn.v_proj": true, "language_model.model.layers.42.self_attn.o_proj": true, "language_model.model.layers.42.self_attn.q_norm": false, "language_model.model.layers.42.self_attn.k_norm": false, "language_model.model.layers.42.self_attn.rope": false, "language_model.model.layers.42.mlp.gate_proj": true, "language_model.model.layers.42.mlp.down_proj": true, "language_model.model.layers.42.mlp.up_proj": true, "language_model.model.layers.42.input_layernorm": false, "language_model.model.layers.42.post_attention_layernorm": false, "language_model.model.layers.42.pre_feedforward_layernorm": false, "language_model.model.layers.42.post_feedforward_layernorm": false, "language_model.model.layers.43.self_attn.q_proj": true, "language_model.model.layers.43.self_attn.k_proj": true, "language_model.model.layers.43.self_attn.v_proj": true, "language_model.model.layers.43.self_attn.o_proj": true, "language_model.model.layers.43.self_attn.q_norm": false, "language_model.model.layers.43.self_attn.k_norm": false, "language_model.model.layers.43.self_attn.rope": false, "language_model.model.layers.43.mlp.gate_proj": true, "language_model.model.layers.43.mlp.down_proj": true, "language_model.model.layers.43.mlp.up_proj": true, "language_model.model.layers.43.input_layernorm": false, "language_model.model.layers.43.post_attention_layernorm": false, "language_model.model.layers.43.pre_feedforward_layernorm": false, "language_model.model.layers.43.post_feedforward_layernorm": false, "language_model.model.layers.44.self_attn.q_proj": true, "language_model.model.layers.44.self_attn.k_proj": true, "language_model.model.layers.44.self_attn.v_proj": true, "language_model.model.layers.44.self_attn.o_proj": true, "language_model.model.layers.44.self_attn.q_norm": false, "language_model.model.layers.44.self_attn.k_norm": false, "language_model.model.layers.44.self_attn.rope": false, "language_model.model.layers.44.mlp.gate_proj": true, "language_model.model.layers.44.mlp.down_proj": true, "language_model.model.layers.44.mlp.up_proj": true, "language_model.model.layers.44.input_layernorm": false, "language_model.model.layers.44.post_attention_layernorm": false, "language_model.model.layers.44.pre_feedforward_layernorm": false, "language_model.model.layers.44.post_feedforward_layernorm": false, "language_model.model.layers.45.self_attn.q_proj": true, "language_model.model.layers.45.self_attn.k_proj": true, "language_model.model.layers.45.self_attn.v_proj": true, "language_model.model.layers.45.self_attn.o_proj": true, "language_model.model.layers.45.self_attn.q_norm": false, "language_model.model.layers.45.self_attn.k_norm": false, "language_model.model.layers.45.self_attn.rope": false, "language_model.model.layers.45.mlp.gate_proj": true, "language_model.model.layers.45.mlp.down_proj": true, "language_model.model.layers.45.mlp.up_proj": true, "language_model.model.layers.45.input_layernorm": false, "language_model.model.layers.45.post_attention_layernorm": false, "language_model.model.layers.45.pre_feedforward_layernorm": false, "language_model.model.layers.45.post_feedforward_layernorm": false, "language_model.model.layers.46.self_attn.q_proj": true, "language_model.model.layers.46.self_attn.k_proj": true, "language_model.model.layers.46.self_attn.v_proj": true, "language_model.model.layers.46.self_attn.o_proj": true, "language_model.model.layers.46.self_attn.q_norm": false, "language_model.model.layers.46.self_attn.k_norm": false, "language_model.model.layers.46.self_attn.rope": false, "language_model.model.layers.46.mlp.gate_proj": true, "language_model.model.layers.46.mlp.down_proj": true, "language_model.model.layers.46.mlp.up_proj": true, "language_model.model.layers.46.input_layernorm": false, "language_model.model.layers.46.post_attention_layernorm": false, "language_model.model.layers.46.pre_feedforward_layernorm": false, "language_model.model.layers.46.post_feedforward_layernorm": false, "language_model.model.layers.47.self_attn.q_proj": true, "language_model.model.layers.47.self_attn.k_proj": true, "language_model.model.layers.47.self_attn.v_proj": true, "language_model.model.layers.47.self_attn.o_proj": true, "language_model.model.layers.47.self_attn.q_norm": false, "language_model.model.layers.47.self_attn.k_norm": false, "language_model.model.layers.47.self_attn.rope": false, "language_model.model.layers.47.mlp.gate_proj": true, "language_model.model.layers.47.mlp.down_proj": true, "language_model.model.layers.47.mlp.up_proj": true, "language_model.model.layers.47.input_layernorm": false, "language_model.model.layers.47.post_attention_layernorm": false, "language_model.model.layers.47.pre_feedforward_layernorm": false, "language_model.model.layers.47.post_feedforward_layernorm": false, "language_model.model.layers.48.self_attn.q_proj": true, "language_model.model.layers.48.self_attn.k_proj": true, "language_model.model.layers.48.self_attn.v_proj": true, "language_model.model.layers.48.self_attn.o_proj": true, "language_model.model.layers.48.self_attn.q_norm": false, "language_model.model.layers.48.self_attn.k_norm": false, "language_model.model.layers.48.self_attn.rope": false, "language_model.model.layers.48.mlp.gate_proj": true, "language_model.model.layers.48.mlp.down_proj": true, "language_model.model.layers.48.mlp.up_proj": true, "language_model.model.layers.48.input_layernorm": false, "language_model.model.layers.48.post_attention_layernorm": false, "language_model.model.layers.48.pre_feedforward_layernorm": false, "language_model.model.layers.48.post_feedforward_layernorm": false, "language_model.model.layers.49.self_attn.q_proj": true, "language_model.model.layers.49.self_attn.k_proj": true, "language_model.model.layers.49.self_attn.v_proj": true, "language_model.model.layers.49.self_attn.o_proj": true, "language_model.model.layers.49.self_attn.q_norm": false, "language_model.model.layers.49.self_attn.k_norm": false, "language_model.model.layers.49.self_attn.rope": false, "language_model.model.layers.49.mlp.gate_proj": true, "language_model.model.layers.49.mlp.down_proj": true, "language_model.model.layers.49.mlp.up_proj": true, "language_model.model.layers.49.input_layernorm": false, "language_model.model.layers.49.post_attention_layernorm": false, "language_model.model.layers.49.pre_feedforward_layernorm": false, "language_model.model.layers.49.post_feedforward_layernorm": false, "language_model.model.layers.50.self_attn.q_proj": true, "language_model.model.layers.50.self_attn.k_proj": true, "language_model.model.layers.50.self_attn.v_proj": true, "language_model.model.layers.50.self_attn.o_proj": true, "language_model.model.layers.50.self_attn.q_norm": false, "language_model.model.layers.50.self_attn.k_norm": false, "language_model.model.layers.50.self_attn.rope": false, "language_model.model.layers.50.mlp.gate_proj": true, "language_model.model.layers.50.mlp.down_proj": true, "language_model.model.layers.50.mlp.up_proj": true, "language_model.model.layers.50.input_layernorm": false, "language_model.model.layers.50.post_attention_layernorm": false, "language_model.model.layers.50.pre_feedforward_layernorm": false, "language_model.model.layers.50.post_feedforward_layernorm": false, "language_model.model.layers.51.self_attn.q_proj": true, "language_model.model.layers.51.self_attn.k_proj": true, "language_model.model.layers.51.self_attn.v_proj": true, "language_model.model.layers.51.self_attn.o_proj": true, "language_model.model.layers.51.self_attn.q_norm": false, "language_model.model.layers.51.self_attn.k_norm": false, "language_model.model.layers.51.self_attn.rope": false, "language_model.model.layers.51.mlp.gate_proj": true, "language_model.model.layers.51.mlp.down_proj": true, "language_model.model.layers.51.mlp.up_proj": true, "language_model.model.layers.51.input_layernorm": false, "language_model.model.layers.51.post_attention_layernorm": false, "language_model.model.layers.51.pre_feedforward_layernorm": false, "language_model.model.layers.51.post_feedforward_layernorm": false, "language_model.model.layers.52.self_attn.q_proj": true, "language_model.model.layers.52.self_attn.k_proj": true, "language_model.model.layers.52.self_attn.v_proj": true, "language_model.model.layers.52.self_attn.o_proj": true, "language_model.model.layers.52.self_attn.q_norm": false, "language_model.model.layers.52.self_attn.k_norm": false, "language_model.model.layers.52.self_attn.rope": false, "language_model.model.layers.52.mlp.gate_proj": true, "language_model.model.layers.52.mlp.down_proj": true, "language_model.model.layers.52.mlp.up_proj": true, "language_model.model.layers.52.input_layernorm": false, "language_model.model.layers.52.post_attention_layernorm": false, "language_model.model.layers.52.pre_feedforward_layernorm": false, "language_model.model.layers.52.post_feedforward_layernorm": false, "language_model.model.layers.53.self_attn.q_proj": true, "language_model.model.layers.53.self_attn.k_proj": true, "language_model.model.layers.53.self_attn.v_proj": true, "language_model.model.layers.53.self_attn.o_proj": true, "language_model.model.layers.53.self_attn.q_norm": false, "language_model.model.layers.53.self_attn.k_norm": false, "language_model.model.layers.53.self_attn.rope": false, "language_model.model.layers.53.mlp.gate_proj": true, "language_model.model.layers.53.mlp.down_proj": true, "language_model.model.layers.53.mlp.up_proj": true, "language_model.model.layers.53.input_layernorm": false, "language_model.model.layers.53.post_attention_layernorm": false, "language_model.model.layers.53.pre_feedforward_layernorm": false, "language_model.model.layers.53.post_feedforward_layernorm": false, "language_model.model.layers.54.self_attn.q_proj": true, "language_model.model.layers.54.self_attn.k_proj": true, "language_model.model.layers.54.self_attn.v_proj": true, "language_model.model.layers.54.self_attn.o_proj": true, "language_model.model.layers.54.self_attn.q_norm": false, "language_model.model.layers.54.self_attn.k_norm": false, "language_model.model.layers.54.self_attn.rope": false, "language_model.model.layers.54.mlp.gate_proj": true, "language_model.model.layers.54.mlp.down_proj": true, "language_model.model.layers.54.mlp.up_proj": true, "language_model.model.layers.54.input_layernorm": false, "language_model.model.layers.54.post_attention_layernorm": false, "language_model.model.layers.54.pre_feedforward_layernorm": false, "language_model.model.layers.54.post_feedforward_layernorm": false, "language_model.model.layers.55.self_attn.q_proj": true, "language_model.model.layers.55.self_attn.k_proj": true, "language_model.model.layers.55.self_attn.v_proj": true, "language_model.model.layers.55.self_attn.o_proj": true, "language_model.model.layers.55.self_attn.q_norm": false, "language_model.model.layers.55.self_attn.k_norm": false, "language_model.model.layers.55.self_attn.rope": false, "language_model.model.layers.55.mlp.gate_proj": true, "language_model.model.layers.55.mlp.down_proj": true, "language_model.model.layers.55.mlp.up_proj": true, "language_model.model.layers.55.input_layernorm": false, "language_model.model.layers.55.post_attention_layernorm": false, "language_model.model.layers.55.pre_feedforward_layernorm": false, "language_model.model.layers.55.post_feedforward_layernorm": false, "language_model.model.layers.56.self_attn.q_proj": true, "language_model.model.layers.56.self_attn.k_proj": true, "language_model.model.layers.56.self_attn.v_proj": true, "language_model.model.layers.56.self_attn.o_proj": true, "language_model.model.layers.56.self_attn.q_norm": false, "language_model.model.layers.56.self_attn.k_norm": false, "language_model.model.layers.56.self_attn.rope": false, "language_model.model.layers.56.mlp.gate_proj": true, "language_model.model.layers.56.mlp.down_proj": true, "language_model.model.layers.56.mlp.up_proj": true, "language_model.model.layers.56.input_layernorm": false, "language_model.model.layers.56.post_attention_layernorm": false, "language_model.model.layers.56.pre_feedforward_layernorm": false, "language_model.model.layers.56.post_feedforward_layernorm": false, "language_model.model.layers.57.self_attn.q_proj": true, "language_model.model.layers.57.self_attn.k_proj": true, "language_model.model.layers.57.self_attn.v_proj": true, "language_model.model.layers.57.self_attn.o_proj": true, "language_model.model.layers.57.self_attn.q_norm": false, "language_model.model.layers.57.self_attn.k_norm": false, "language_model.model.layers.57.self_attn.rope": false, "language_model.model.layers.57.mlp.gate_proj": true, "language_model.model.layers.57.mlp.down_proj": true, "language_model.model.layers.57.mlp.up_proj": true, "language_model.model.layers.57.input_layernorm": false, "language_model.model.layers.57.post_attention_layernorm": false, "language_model.model.layers.57.pre_feedforward_layernorm": false, "language_model.model.layers.57.post_feedforward_layernorm": false, "language_model.model.layers.58.self_attn.q_proj": true, "language_model.model.layers.58.self_attn.k_proj": true, "language_model.model.layers.58.self_attn.v_proj": true, "language_model.model.layers.58.self_attn.o_proj": true, "language_model.model.layers.58.self_attn.q_norm": false, "language_model.model.layers.58.self_attn.k_norm": false, "language_model.model.layers.58.self_attn.rope": false, "language_model.model.layers.58.mlp.gate_proj": true, "language_model.model.layers.58.mlp.down_proj": true, "language_model.model.layers.58.mlp.up_proj": true, "language_model.model.layers.58.input_layernorm": false, "language_model.model.layers.58.post_attention_layernorm": false, "language_model.model.layers.58.pre_feedforward_layernorm": false, "language_model.model.layers.58.post_feedforward_layernorm": false, "language_model.model.layers.59.self_attn.q_proj": true, "language_model.model.layers.59.self_attn.k_proj": true, "language_model.model.layers.59.self_attn.v_proj": true, "language_model.model.layers.59.self_attn.o_proj": true, "language_model.model.layers.59.self_attn.q_norm": false, "language_model.model.layers.59.self_attn.k_norm": false, "language_model.model.layers.59.self_attn.rope": false, "language_model.model.layers.59.mlp.gate_proj": true, "language_model.model.layers.59.mlp.down_proj": true, "language_model.model.layers.59.mlp.up_proj": true, "language_model.model.layers.59.input_layernorm": false, "language_model.model.layers.59.post_attention_layernorm": false, "language_model.model.layers.59.pre_feedforward_layernorm": false, "language_model.model.layers.59.post_feedforward_layernorm": false, "language_model.model.layers.60.self_attn.q_proj": true, "language_model.model.layers.60.self_attn.k_proj": true, "language_model.model.layers.60.self_attn.v_proj": true, "language_model.model.layers.60.self_attn.o_proj": true, "language_model.model.layers.60.self_attn.q_norm": false, "language_model.model.layers.60.self_attn.k_norm": false, "language_model.model.layers.60.self_attn.rope": false, "language_model.model.layers.60.mlp.gate_proj": true, "language_model.model.layers.60.mlp.down_proj": true, "language_model.model.layers.60.mlp.up_proj": true, "language_model.model.layers.60.input_layernorm": false, "language_model.model.layers.60.post_attention_layernorm": false, "language_model.model.layers.60.pre_feedforward_layernorm": false, "language_model.model.layers.60.post_feedforward_layernorm": false, "language_model.model.layers.61.self_attn.q_proj": true, "language_model.model.layers.61.self_attn.k_proj": true, "language_model.model.layers.61.self_attn.v_proj": true, "language_model.model.layers.61.self_attn.o_proj": true, "language_model.model.layers.61.self_attn.q_norm": false, "language_model.model.layers.61.self_attn.k_norm": false, "language_model.model.layers.61.self_attn.rope": false, "language_model.model.layers.61.mlp.gate_proj": true, "language_model.model.layers.61.mlp.down_proj": true, "language_model.model.layers.61.mlp.up_proj": true, "language_model.model.layers.61.input_layernorm": false, "language_model.model.layers.61.post_attention_layernorm": false, "language_model.model.layers.61.pre_feedforward_layernorm": false, "language_model.model.layers.61.post_feedforward_layernorm": false, "language_model.model.norm": false, "language_model.lm_head": true, "multi_modal_projector.mm_soft_emb_norm": false, "multi_modal_projector.avg_pool": false }, "quantization_config": { "group_size": 64, "bits": 6, "vision_tower.vision_model.embeddings.patch_embedding": false, "vision_tower.vision_model.embeddings.position_embedding": false, "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.0.layer_norm1": false, "vision_tower.vision_model.encoder.layers.0.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.0.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.0.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.0.layer_norm2": false, "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.1.layer_norm1": false, "vision_tower.vision_model.encoder.layers.1.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.1.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.1.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.1.layer_norm2": false, "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.2.layer_norm1": false, "vision_tower.vision_model.encoder.layers.2.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.2.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.2.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.2.layer_norm2": false, "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.3.layer_norm1": false, "vision_tower.vision_model.encoder.layers.3.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.3.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.3.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.3.layer_norm2": false, "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.4.layer_norm1": false, "vision_tower.vision_model.encoder.layers.4.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.4.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.4.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.4.layer_norm2": false, "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.5.layer_norm1": false, "vision_tower.vision_model.encoder.layers.5.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.5.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.5.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.5.layer_norm2": false, "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.6.layer_norm1": false, "vision_tower.vision_model.encoder.layers.6.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.6.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.6.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.6.layer_norm2": false, "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.7.layer_norm1": false, "vision_tower.vision_model.encoder.layers.7.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.7.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.7.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.7.layer_norm2": false, "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.8.layer_norm1": false, "vision_tower.vision_model.encoder.layers.8.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.8.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.8.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.8.layer_norm2": false, "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.9.layer_norm1": false, "vision_tower.vision_model.encoder.layers.9.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.9.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.9.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.9.layer_norm2": false, "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.10.layer_norm1": false, "vision_tower.vision_model.encoder.layers.10.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.10.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.10.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.10.layer_norm2": false, "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.11.layer_norm1": false, "vision_tower.vision_model.encoder.layers.11.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.11.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.11.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.11.layer_norm2": false, "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.12.layer_norm1": false, "vision_tower.vision_model.encoder.layers.12.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.12.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.12.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.12.layer_norm2": false, "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.13.layer_norm1": false, "vision_tower.vision_model.encoder.layers.13.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.13.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.13.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.13.layer_norm2": false, "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.14.layer_norm1": false, "vision_tower.vision_model.encoder.layers.14.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.14.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.14.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.14.layer_norm2": false, "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.15.layer_norm1": false, "vision_tower.vision_model.encoder.layers.15.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.15.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.15.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.15.layer_norm2": false, "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.16.layer_norm1": false, "vision_tower.vision_model.encoder.layers.16.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.16.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.16.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.16.layer_norm2": false, "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.17.layer_norm1": false, "vision_tower.vision_model.encoder.layers.17.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.17.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.17.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.17.layer_norm2": false, "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.18.layer_norm1": false, "vision_tower.vision_model.encoder.layers.18.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.18.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.18.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.18.layer_norm2": false, "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.19.layer_norm1": false, "vision_tower.vision_model.encoder.layers.19.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.19.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.19.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.19.layer_norm2": false, "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.20.layer_norm1": false, "vision_tower.vision_model.encoder.layers.20.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.20.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.20.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.20.layer_norm2": false, "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.21.layer_norm1": false, "vision_tower.vision_model.encoder.layers.21.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.21.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.21.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.21.layer_norm2": false, "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.22.layer_norm1": false, "vision_tower.vision_model.encoder.layers.22.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.22.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.22.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.22.layer_norm2": false, "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.23.layer_norm1": false, "vision_tower.vision_model.encoder.layers.23.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.23.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.23.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.23.layer_norm2": false, "vision_tower.vision_model.encoder.layers.24.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.24.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.24.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.24.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.24.layer_norm1": false, "vision_tower.vision_model.encoder.layers.24.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.24.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.24.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.24.layer_norm2": false, "vision_tower.vision_model.encoder.layers.25.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.25.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.25.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.25.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.25.layer_norm1": false, "vision_tower.vision_model.encoder.layers.25.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.25.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.25.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.25.layer_norm2": false, "vision_tower.vision_model.encoder.layers.26.self_attn.q_proj": false, "vision_tower.vision_model.encoder.layers.26.self_attn.k_proj": false, "vision_tower.vision_model.encoder.layers.26.self_attn.v_proj": false, "vision_tower.vision_model.encoder.layers.26.self_attn.out_proj": false, "vision_tower.vision_model.encoder.layers.26.layer_norm1": false, "vision_tower.vision_model.encoder.layers.26.mlp.activation_fn": false, "vision_tower.vision_model.encoder.layers.26.mlp.fc1": false, "vision_tower.vision_model.encoder.layers.26.mlp.fc2": false, "vision_tower.vision_model.encoder.layers.26.layer_norm2": false, "vision_tower.vision_model.post_layernorm": false, "language_model.model.embed_tokens": true, "language_model.model.layers.0.self_attn.q_proj": true, "language_model.model.layers.0.self_attn.k_proj": true, "language_model.model.layers.0.self_attn.v_proj": true, "language_model.model.layers.0.self_attn.o_proj": true, "language_model.model.layers.0.self_attn.q_norm": false, "language_model.model.layers.0.self_attn.k_norm": false, "language_model.model.layers.0.self_attn.rope": false, "language_model.model.layers.0.mlp.gate_proj": true, "language_model.model.layers.0.mlp.down_proj": true, "language_model.model.layers.0.mlp.up_proj": true, "language_model.model.layers.0.input_layernorm": false, "language_model.model.layers.0.post_attention_layernorm": false, "language_model.model.layers.0.pre_feedforward_layernorm": false, "language_model.model.layers.0.post_feedforward_layernorm": false, "language_model.model.layers.1.self_attn.q_proj": true, "language_model.model.layers.1.self_attn.k_proj": true, "language_model.model.layers.1.self_attn.v_proj": true, "language_model.model.layers.1.self_attn.o_proj": true, "language_model.model.layers.1.self_attn.q_norm": false, "language_model.model.layers.1.self_attn.k_norm": false, "language_model.model.layers.1.self_attn.rope": false, "language_model.model.layers.1.mlp.gate_proj": true, "language_model.model.layers.1.mlp.down_proj": true, "language_model.model.layers.1.mlp.up_proj": true, "language_model.model.layers.1.input_layernorm": false, "language_model.model.layers.1.post_attention_layernorm": false, "language_model.model.layers.1.pre_feedforward_layernorm": false, "language_model.model.layers.1.post_feedforward_layernorm": false, "language_model.model.layers.2.self_attn.q_proj": true, "language_model.model.layers.2.self_attn.k_proj": true, "language_model.model.layers.2.self_attn.v_proj": true, "language_model.model.layers.2.self_attn.o_proj": true, "language_model.model.layers.2.self_attn.q_norm": false, "language_model.model.layers.2.self_attn.k_norm": false, "language_model.model.layers.2.self_attn.rope": false, "language_model.model.layers.2.mlp.gate_proj": true, "language_model.model.layers.2.mlp.down_proj": true, "language_model.model.layers.2.mlp.up_proj": true, "language_model.model.layers.2.input_layernorm": false, "language_model.model.layers.2.post_attention_layernorm": false, "language_model.model.layers.2.pre_feedforward_layernorm": false, "language_model.model.layers.2.post_feedforward_layernorm": false, "language_model.model.layers.3.self_attn.q_proj": true, "language_model.model.layers.3.self_attn.k_proj": true, "language_model.model.layers.3.self_attn.v_proj": true, "language_model.model.layers.3.self_attn.o_proj": true, "language_model.model.layers.3.self_attn.q_norm": false, "language_model.model.layers.3.self_attn.k_norm": false, "language_model.model.layers.3.self_attn.rope": false, "language_model.model.layers.3.mlp.gate_proj": true, "language_model.model.layers.3.mlp.down_proj": true, "language_model.model.layers.3.mlp.up_proj": true, "language_model.model.layers.3.input_layernorm": false, "language_model.model.layers.3.post_attention_layernorm": false, "language_model.model.layers.3.pre_feedforward_layernorm": false, "language_model.model.layers.3.post_feedforward_layernorm": false, "language_model.model.layers.4.self_attn.q_proj": true, "language_model.model.layers.4.self_attn.k_proj": true, "language_model.model.layers.4.self_attn.v_proj": true, "language_model.model.layers.4.self_attn.o_proj": true, "language_model.model.layers.4.self_attn.q_norm": false, "language_model.model.layers.4.self_attn.k_norm": false, "language_model.model.layers.4.self_attn.rope": false, "language_model.model.layers.4.mlp.gate_proj": true, "language_model.model.layers.4.mlp.down_proj": true, "language_model.model.layers.4.mlp.up_proj": true, "language_model.model.layers.4.input_layernorm": false, "language_model.model.layers.4.post_attention_layernorm": false, "language_model.model.layers.4.pre_feedforward_layernorm": false, "language_model.model.layers.4.post_feedforward_layernorm": false, "language_model.model.layers.5.self_attn.q_proj": true, "language_model.model.layers.5.self_attn.k_proj": true, "language_model.model.layers.5.self_attn.v_proj": true, "language_model.model.layers.5.self_attn.o_proj": true, "language_model.model.layers.5.self_attn.q_norm": false, "language_model.model.layers.5.self_attn.k_norm": false, "language_model.model.layers.5.self_attn.rope": false, "language_model.model.layers.5.mlp.gate_proj": true, "language_model.model.layers.5.mlp.down_proj": true, "language_model.model.layers.5.mlp.up_proj": true, "language_model.model.layers.5.input_layernorm": false, "language_model.model.layers.5.post_attention_layernorm": false, "language_model.model.layers.5.pre_feedforward_layernorm": false, "language_model.model.layers.5.post_feedforward_layernorm": false, "language_model.model.layers.6.self_attn.q_proj": true, "language_model.model.layers.6.self_attn.k_proj": true, "language_model.model.layers.6.self_attn.v_proj": true, "language_model.model.layers.6.self_attn.o_proj": true, "language_model.model.layers.6.self_attn.q_norm": false, "language_model.model.layers.6.self_attn.k_norm": false, "language_model.model.layers.6.self_attn.rope": false, "language_model.model.layers.6.mlp.gate_proj": true, "language_model.model.layers.6.mlp.down_proj": true, "language_model.model.layers.6.mlp.up_proj": true, "language_model.model.layers.6.input_layernorm": false, "language_model.model.layers.6.post_attention_layernorm": false, "language_model.model.layers.6.pre_feedforward_layernorm": false, "language_model.model.layers.6.post_feedforward_layernorm": false, "language_model.model.layers.7.self_attn.q_proj": true, "language_model.model.layers.7.self_attn.k_proj": true, "language_model.model.layers.7.self_attn.v_proj": true, "language_model.model.layers.7.self_attn.o_proj": true, "language_model.model.layers.7.self_attn.q_norm": false, "language_model.model.layers.7.self_attn.k_norm": false, "language_model.model.layers.7.self_attn.rope": false, "language_model.model.layers.7.mlp.gate_proj": true, "language_model.model.layers.7.mlp.down_proj": true, "language_model.model.layers.7.mlp.up_proj": true, "language_model.model.layers.7.input_layernorm": false, "language_model.model.layers.7.post_attention_layernorm": false, "language_model.model.layers.7.pre_feedforward_layernorm": false, "language_model.model.layers.7.post_feedforward_layernorm": false, "language_model.model.layers.8.self_attn.q_proj": true, "language_model.model.layers.8.self_attn.k_proj": true, "language_model.model.layers.8.self_attn.v_proj": true, "language_model.model.layers.8.self_attn.o_proj": true, "language_model.model.layers.8.self_attn.q_norm": false, "language_model.model.layers.8.self_attn.k_norm": false, "language_model.model.layers.8.self_attn.rope": false, "language_model.model.layers.8.mlp.gate_proj": true, "language_model.model.layers.8.mlp.down_proj": true, "language_model.model.layers.8.mlp.up_proj": true, "language_model.model.layers.8.input_layernorm": false, "language_model.model.layers.8.post_attention_layernorm": false, "language_model.model.layers.8.pre_feedforward_layernorm": false, "language_model.model.layers.8.post_feedforward_layernorm": false, "language_model.model.layers.9.self_attn.q_proj": true, "language_model.model.layers.9.self_attn.k_proj": true, "language_model.model.layers.9.self_attn.v_proj": true, "language_model.model.layers.9.self_attn.o_proj": true, "language_model.model.layers.9.self_attn.q_norm": false, "language_model.model.layers.9.self_attn.k_norm": false, "language_model.model.layers.9.self_attn.rope": false, "language_model.model.layers.9.mlp.gate_proj": true, "language_model.model.layers.9.mlp.down_proj": true, "language_model.model.layers.9.mlp.up_proj": true, "language_model.model.layers.9.input_layernorm": false, "language_model.model.layers.9.post_attention_layernorm": false, "language_model.model.layers.9.pre_feedforward_layernorm": false, "language_model.model.layers.9.post_feedforward_layernorm": false, "language_model.model.layers.10.self_attn.q_proj": true, "language_model.model.layers.10.self_attn.k_proj": true, "language_model.model.layers.10.self_attn.v_proj": true, "language_model.model.layers.10.self_attn.o_proj": true, "language_model.model.layers.10.self_attn.q_norm": false, "language_model.model.layers.10.self_attn.k_norm": false, "language_model.model.layers.10.self_attn.rope": false, "language_model.model.layers.10.mlp.gate_proj": true, "language_model.model.layers.10.mlp.down_proj": true, "language_model.model.layers.10.mlp.up_proj": true, "language_model.model.layers.10.input_layernorm": false, "language_model.model.layers.10.post_attention_layernorm": false, "language_model.model.layers.10.pre_feedforward_layernorm": false, "language_model.model.layers.10.post_feedforward_layernorm": false, "language_model.model.layers.11.self_attn.q_proj": true, "language_model.model.layers.11.self_attn.k_proj": true, "language_model.model.layers.11.self_attn.v_proj": true, "language_model.model.layers.11.self_attn.o_proj": true, "language_model.model.layers.11.self_attn.q_norm": false, "language_model.model.layers.11.self_attn.k_norm": false, "language_model.model.layers.11.self_attn.rope": false, "language_model.model.layers.11.mlp.gate_proj": true, "language_model.model.layers.11.mlp.down_proj": true, "language_model.model.layers.11.mlp.up_proj": true, "language_model.model.layers.11.input_layernorm": false, "language_model.model.layers.11.post_attention_layernorm": false, "language_model.model.layers.11.pre_feedforward_layernorm": false, "language_model.model.layers.11.post_feedforward_layernorm": false, "language_model.model.layers.12.self_attn.q_proj": true, "language_model.model.layers.12.self_attn.k_proj": true, "language_model.model.layers.12.self_attn.v_proj": true, "language_model.model.layers.12.self_attn.o_proj": true, "language_model.model.layers.12.self_attn.q_norm": false, "language_model.model.layers.12.self_attn.k_norm": false, "language_model.model.layers.12.self_attn.rope": false, "language_model.model.layers.12.mlp.gate_proj": true, "language_model.model.layers.12.mlp.down_proj": true, "language_model.model.layers.12.mlp.up_proj": true, "language_model.model.layers.12.input_layernorm": false, "language_model.model.layers.12.post_attention_layernorm": false, "language_model.model.layers.12.pre_feedforward_layernorm": false, "language_model.model.layers.12.post_feedforward_layernorm": false, "language_model.model.layers.13.self_attn.q_proj": true, "language_model.model.layers.13.self_attn.k_proj": true, "language_model.model.layers.13.self_attn.v_proj": true, "language_model.model.layers.13.self_attn.o_proj": true, "language_model.model.layers.13.self_attn.q_norm": false, "language_model.model.layers.13.self_attn.k_norm": false, "language_model.model.layers.13.self_attn.rope": false, "language_model.model.layers.13.mlp.gate_proj": true, "language_model.model.layers.13.mlp.down_proj": true, "language_model.model.layers.13.mlp.up_proj": true, "language_model.model.layers.13.input_layernorm": false, "language_model.model.layers.13.post_attention_layernorm": false, "language_model.model.layers.13.pre_feedforward_layernorm": false, "language_model.model.layers.13.post_feedforward_layernorm": false, "language_model.model.layers.14.self_attn.q_proj": true, "language_model.model.layers.14.self_attn.k_proj": true, "language_model.model.layers.14.self_attn.v_proj": true, "language_model.model.layers.14.self_attn.o_proj": true, "language_model.model.layers.14.self_attn.q_norm": false, "language_model.model.layers.14.self_attn.k_norm": false, "language_model.model.layers.14.self_attn.rope": false, "language_model.model.layers.14.mlp.gate_proj": true, "language_model.model.layers.14.mlp.down_proj": true, "language_model.model.layers.14.mlp.up_proj": true, "language_model.model.layers.14.input_layernorm": false, "language_model.model.layers.14.post_attention_layernorm": false, "language_model.model.layers.14.pre_feedforward_layernorm": false, "language_model.model.layers.14.post_feedforward_layernorm": false, "language_model.model.layers.15.self_attn.q_proj": true, "language_model.model.layers.15.self_attn.k_proj": true, "language_model.model.layers.15.self_attn.v_proj": true, "language_model.model.layers.15.self_attn.o_proj": true, "language_model.model.layers.15.self_attn.q_norm": false, "language_model.model.layers.15.self_attn.k_norm": false, "language_model.model.layers.15.self_attn.rope": false, "language_model.model.layers.15.mlp.gate_proj": true, "language_model.model.layers.15.mlp.down_proj": true, "language_model.model.layers.15.mlp.up_proj": true, "language_model.model.layers.15.input_layernorm": false, "language_model.model.layers.15.post_attention_layernorm": false, "language_model.model.layers.15.pre_feedforward_layernorm": false, "language_model.model.layers.15.post_feedforward_layernorm": false, "language_model.model.layers.16.self_attn.q_proj": true, "language_model.model.layers.16.self_attn.k_proj": true, "language_model.model.layers.16.self_attn.v_proj": true, "language_model.model.layers.16.self_attn.o_proj": true, "language_model.model.layers.16.self_attn.q_norm": false, "language_model.model.layers.16.self_attn.k_norm": false, "language_model.model.layers.16.self_attn.rope": false, "language_model.model.layers.16.mlp.gate_proj": true, "language_model.model.layers.16.mlp.down_proj": true, "language_model.model.layers.16.mlp.up_proj": true, "language_model.model.layers.16.input_layernorm": false, "language_model.model.layers.16.post_attention_layernorm": false, "language_model.model.layers.16.pre_feedforward_layernorm": false, "language_model.model.layers.16.post_feedforward_layernorm": false, "language_model.model.layers.17.self_attn.q_proj": true, "language_model.model.layers.17.self_attn.k_proj": true, "language_model.model.layers.17.self_attn.v_proj": true, "language_model.model.layers.17.self_attn.o_proj": true, "language_model.model.layers.17.self_attn.q_norm": false, "language_model.model.layers.17.self_attn.k_norm": false, "language_model.model.layers.17.self_attn.rope": false, "language_model.model.layers.17.mlp.gate_proj": true, "language_model.model.layers.17.mlp.down_proj": true, "language_model.model.layers.17.mlp.up_proj": true, "language_model.model.layers.17.input_layernorm": false, "language_model.model.layers.17.post_attention_layernorm": false, "language_model.model.layers.17.pre_feedforward_layernorm": false, "language_model.model.layers.17.post_feedforward_layernorm": false, "language_model.model.layers.18.self_attn.q_proj": true, "language_model.model.layers.18.self_attn.k_proj": true, "language_model.model.layers.18.self_attn.v_proj": true, "language_model.model.layers.18.self_attn.o_proj": true, "language_model.model.layers.18.self_attn.q_norm": false, "language_model.model.layers.18.self_attn.k_norm": false, "language_model.model.layers.18.self_attn.rope": false, "language_model.model.layers.18.mlp.gate_proj": true, "language_model.model.layers.18.mlp.down_proj": true, "language_model.model.layers.18.mlp.up_proj": true, "language_model.model.layers.18.input_layernorm": false, "language_model.model.layers.18.post_attention_layernorm": false, "language_model.model.layers.18.pre_feedforward_layernorm": false, "language_model.model.layers.18.post_feedforward_layernorm": false, "language_model.model.layers.19.self_attn.q_proj": true, "language_model.model.layers.19.self_attn.k_proj": true, "language_model.model.layers.19.self_attn.v_proj": true, "language_model.model.layers.19.self_attn.o_proj": true, "language_model.model.layers.19.self_attn.q_norm": false, "language_model.model.layers.19.self_attn.k_norm": false, "language_model.model.layers.19.self_attn.rope": false, "language_model.model.layers.19.mlp.gate_proj": true, "language_model.model.layers.19.mlp.down_proj": true, "language_model.model.layers.19.mlp.up_proj": true, "language_model.model.layers.19.input_layernorm": false, "language_model.model.layers.19.post_attention_layernorm": false, "language_model.model.layers.19.pre_feedforward_layernorm": false, "language_model.model.layers.19.post_feedforward_layernorm": false, "language_model.model.layers.20.self_attn.q_proj": true, "language_model.model.layers.20.self_attn.k_proj": true, "language_model.model.layers.20.self_attn.v_proj": true, "language_model.model.layers.20.self_attn.o_proj": true, "language_model.model.layers.20.self_attn.q_norm": false, "language_model.model.layers.20.self_attn.k_norm": false, "language_model.model.layers.20.self_attn.rope": false, "language_model.model.layers.20.mlp.gate_proj": true, "language_model.model.layers.20.mlp.down_proj": true, "language_model.model.layers.20.mlp.up_proj": true, "language_model.model.layers.20.input_layernorm": false, "language_model.model.layers.20.post_attention_layernorm": false, "language_model.model.layers.20.pre_feedforward_layernorm": false, "language_model.model.layers.20.post_feedforward_layernorm": false, "language_model.model.layers.21.self_attn.q_proj": true, "language_model.model.layers.21.self_attn.k_proj": true, "language_model.model.layers.21.self_attn.v_proj": true, "language_model.model.layers.21.self_attn.o_proj": true, "language_model.model.layers.21.self_attn.q_norm": false, "language_model.model.layers.21.self_attn.k_norm": false, "language_model.model.layers.21.self_attn.rope": false, "language_model.model.layers.21.mlp.gate_proj": true, "language_model.model.layers.21.mlp.down_proj": true, "language_model.model.layers.21.mlp.up_proj": true, "language_model.model.layers.21.input_layernorm": false, "language_model.model.layers.21.post_attention_layernorm": false, "language_model.model.layers.21.pre_feedforward_layernorm": false, "language_model.model.layers.21.post_feedforward_layernorm": false, "language_model.model.layers.22.self_attn.q_proj": true, "language_model.model.layers.22.self_attn.k_proj": true, "language_model.model.layers.22.self_attn.v_proj": true, "language_model.model.layers.22.self_attn.o_proj": true, "language_model.model.layers.22.self_attn.q_norm": false, "language_model.model.layers.22.self_attn.k_norm": false, "language_model.model.layers.22.self_attn.rope": false, "language_model.model.layers.22.mlp.gate_proj": true, "language_model.model.layers.22.mlp.down_proj": true, "language_model.model.layers.22.mlp.up_proj": true, "language_model.model.layers.22.input_layernorm": false, "language_model.model.layers.22.post_attention_layernorm": false, "language_model.model.layers.22.pre_feedforward_layernorm": false, "language_model.model.layers.22.post_feedforward_layernorm": false, "language_model.model.layers.23.self_attn.q_proj": true, "language_model.model.layers.23.self_attn.k_proj": true, "language_model.model.layers.23.self_attn.v_proj": true, "language_model.model.layers.23.self_attn.o_proj": true, "language_model.model.layers.23.self_attn.q_norm": false, "language_model.model.layers.23.self_attn.k_norm": false, "language_model.model.layers.23.self_attn.rope": false, "language_model.model.layers.23.mlp.gate_proj": true, "language_model.model.layers.23.mlp.down_proj": true, "language_model.model.layers.23.mlp.up_proj": true, "language_model.model.layers.23.input_layernorm": false, "language_model.model.layers.23.post_attention_layernorm": false, "language_model.model.layers.23.pre_feedforward_layernorm": false, "language_model.model.layers.23.post_feedforward_layernorm": false, "language_model.model.layers.24.self_attn.q_proj": true, "language_model.model.layers.24.self_attn.k_proj": true, "language_model.model.layers.24.self_attn.v_proj": true, "language_model.model.layers.24.self_attn.o_proj": true, "language_model.model.layers.24.self_attn.q_norm": false, "language_model.model.layers.24.self_attn.k_norm": false, "language_model.model.layers.24.self_attn.rope": false, "language_model.model.layers.24.mlp.gate_proj": true, "language_model.model.layers.24.mlp.down_proj": true, "language_model.model.layers.24.mlp.up_proj": true, "language_model.model.layers.24.input_layernorm": false, "language_model.model.layers.24.post_attention_layernorm": false, "language_model.model.layers.24.pre_feedforward_layernorm": false, "language_model.model.layers.24.post_feedforward_layernorm": false, "language_model.model.layers.25.self_attn.q_proj": true, "language_model.model.layers.25.self_attn.k_proj": true, "language_model.model.layers.25.self_attn.v_proj": true, "language_model.model.layers.25.self_attn.o_proj": true, "language_model.model.layers.25.self_attn.q_norm": false, "language_model.model.layers.25.self_attn.k_norm": false, "language_model.model.layers.25.self_attn.rope": false, "language_model.model.layers.25.mlp.gate_proj": true, "language_model.model.layers.25.mlp.down_proj": true, "language_model.model.layers.25.mlp.up_proj": true, "language_model.model.layers.25.input_layernorm": false, "language_model.model.layers.25.post_attention_layernorm": false, "language_model.model.layers.25.pre_feedforward_layernorm": false, "language_model.model.layers.25.post_feedforward_layernorm": false, "language_model.model.layers.26.self_attn.q_proj": true, "language_model.model.layers.26.self_attn.k_proj": true, "language_model.model.layers.26.self_attn.v_proj": true, "language_model.model.layers.26.self_attn.o_proj": true, "language_model.model.layers.26.self_attn.q_norm": false, "language_model.model.layers.26.self_attn.k_norm": false, "language_model.model.layers.26.self_attn.rope": false, "language_model.model.layers.26.mlp.gate_proj": true, "language_model.model.layers.26.mlp.down_proj": true, "language_model.model.layers.26.mlp.up_proj": true, "language_model.model.layers.26.input_layernorm": false, "language_model.model.layers.26.post_attention_layernorm": false, "language_model.model.layers.26.pre_feedforward_layernorm": false, "language_model.model.layers.26.post_feedforward_layernorm": false, "language_model.model.layers.27.self_attn.q_proj": true, "language_model.model.layers.27.self_attn.k_proj": true, "language_model.model.layers.27.self_attn.v_proj": true, "language_model.model.layers.27.self_attn.o_proj": true, "language_model.model.layers.27.self_attn.q_norm": false, "language_model.model.layers.27.self_attn.k_norm": false, "language_model.model.layers.27.self_attn.rope": false, "language_model.model.layers.27.mlp.gate_proj": true, "language_model.model.layers.27.mlp.down_proj": true, "language_model.model.layers.27.mlp.up_proj": true, "language_model.model.layers.27.input_layernorm": false, "language_model.model.layers.27.post_attention_layernorm": false, "language_model.model.layers.27.pre_feedforward_layernorm": false, "language_model.model.layers.27.post_feedforward_layernorm": false, "language_model.model.layers.28.self_attn.q_proj": true, "language_model.model.layers.28.self_attn.k_proj": true, "language_model.model.layers.28.self_attn.v_proj": true, "language_model.model.layers.28.self_attn.o_proj": true, "language_model.model.layers.28.self_attn.q_norm": false, "language_model.model.layers.28.self_attn.k_norm": false, "language_model.model.layers.28.self_attn.rope": false, "language_model.model.layers.28.mlp.gate_proj": true, "language_model.model.layers.28.mlp.down_proj": true, "language_model.model.layers.28.mlp.up_proj": true, "language_model.model.layers.28.input_layernorm": false, "language_model.model.layers.28.post_attention_layernorm": false, "language_model.model.layers.28.pre_feedforward_layernorm": false, "language_model.model.layers.28.post_feedforward_layernorm": false, "language_model.model.layers.29.self_attn.q_proj": true, "language_model.model.layers.29.self_attn.k_proj": true, "language_model.model.layers.29.self_attn.v_proj": true, "language_model.model.layers.29.self_attn.o_proj": true, "language_model.model.layers.29.self_attn.q_norm": false, "language_model.model.layers.29.self_attn.k_norm": false, "language_model.model.layers.29.self_attn.rope": false, "language_model.model.layers.29.mlp.gate_proj": true, "language_model.model.layers.29.mlp.down_proj": true, "language_model.model.layers.29.mlp.up_proj": true, "language_model.model.layers.29.input_layernorm": false, "language_model.model.layers.29.post_attention_layernorm": false, "language_model.model.layers.29.pre_feedforward_layernorm": false, "language_model.model.layers.29.post_feedforward_layernorm": false, "language_model.model.layers.30.self_attn.q_proj": true, "language_model.model.layers.30.self_attn.k_proj": true, "language_model.model.layers.30.self_attn.v_proj": true, "language_model.model.layers.30.self_attn.o_proj": true, "language_model.model.layers.30.self_attn.q_norm": false, "language_model.model.layers.30.self_attn.k_norm": false, "language_model.model.layers.30.self_attn.rope": false, "language_model.model.layers.30.mlp.gate_proj": true, "language_model.model.layers.30.mlp.down_proj": true, "language_model.model.layers.30.mlp.up_proj": true, "language_model.model.layers.30.input_layernorm": false, "language_model.model.layers.30.post_attention_layernorm": false, "language_model.model.layers.30.pre_feedforward_layernorm": false, "language_model.model.layers.30.post_feedforward_layernorm": false, "language_model.model.layers.31.self_attn.q_proj": true, "language_model.model.layers.31.self_attn.k_proj": true, "language_model.model.layers.31.self_attn.v_proj": true, "language_model.model.layers.31.self_attn.o_proj": true, "language_model.model.layers.31.self_attn.q_norm": false, "language_model.model.layers.31.self_attn.k_norm": false, "language_model.model.layers.31.self_attn.rope": false, "language_model.model.layers.31.mlp.gate_proj": true, "language_model.model.layers.31.mlp.down_proj": true, "language_model.model.layers.31.mlp.up_proj": true, "language_model.model.layers.31.input_layernorm": false, "language_model.model.layers.31.post_attention_layernorm": false, "language_model.model.layers.31.pre_feedforward_layernorm": false, "language_model.model.layers.31.post_feedforward_layernorm": false, "language_model.model.layers.32.self_attn.q_proj": true, "language_model.model.layers.32.self_attn.k_proj": true, "language_model.model.layers.32.self_attn.v_proj": true, "language_model.model.layers.32.self_attn.o_proj": true, "language_model.model.layers.32.self_attn.q_norm": false, "language_model.model.layers.32.self_attn.k_norm": false, "language_model.model.layers.32.self_attn.rope": false, "language_model.model.layers.32.mlp.gate_proj": true, "language_model.model.layers.32.mlp.down_proj": true, "language_model.model.layers.32.mlp.up_proj": true, "language_model.model.layers.32.input_layernorm": false, "language_model.model.layers.32.post_attention_layernorm": false, "language_model.model.layers.32.pre_feedforward_layernorm": false, "language_model.model.layers.32.post_feedforward_layernorm": false, "language_model.model.layers.33.self_attn.q_proj": true, "language_model.model.layers.33.self_attn.k_proj": true, "language_model.model.layers.33.self_attn.v_proj": true, "language_model.model.layers.33.self_attn.o_proj": true, "language_model.model.layers.33.self_attn.q_norm": false, "language_model.model.layers.33.self_attn.k_norm": false, "language_model.model.layers.33.self_attn.rope": false, "language_model.model.layers.33.mlp.gate_proj": true, "language_model.model.layers.33.mlp.down_proj": true, "language_model.model.layers.33.mlp.up_proj": true, "language_model.model.layers.33.input_layernorm": false, "language_model.model.layers.33.post_attention_layernorm": false, "language_model.model.layers.33.pre_feedforward_layernorm": false, "language_model.model.layers.33.post_feedforward_layernorm": false, "language_model.model.layers.34.self_attn.q_proj": true, "language_model.model.layers.34.self_attn.k_proj": true, "language_model.model.layers.34.self_attn.v_proj": true, "language_model.model.layers.34.self_attn.o_proj": true, "language_model.model.layers.34.self_attn.q_norm": false, "language_model.model.layers.34.self_attn.k_norm": false, "language_model.model.layers.34.self_attn.rope": false, "language_model.model.layers.34.mlp.gate_proj": true, "language_model.model.layers.34.mlp.down_proj": true, "language_model.model.layers.34.mlp.up_proj": true, "language_model.model.layers.34.input_layernorm": false, "language_model.model.layers.34.post_attention_layernorm": false, "language_model.model.layers.34.pre_feedforward_layernorm": false, "language_model.model.layers.34.post_feedforward_layernorm": false, "language_model.model.layers.35.self_attn.q_proj": true, "language_model.model.layers.35.self_attn.k_proj": true, "language_model.model.layers.35.self_attn.v_proj": true, "language_model.model.layers.35.self_attn.o_proj": true, "language_model.model.layers.35.self_attn.q_norm": false, "language_model.model.layers.35.self_attn.k_norm": false, "language_model.model.layers.35.self_attn.rope": false, "language_model.model.layers.35.mlp.gate_proj": true, "language_model.model.layers.35.mlp.down_proj": true, "language_model.model.layers.35.mlp.up_proj": true, "language_model.model.layers.35.input_layernorm": false, "language_model.model.layers.35.post_attention_layernorm": false, "language_model.model.layers.35.pre_feedforward_layernorm": false, "language_model.model.layers.35.post_feedforward_layernorm": false, "language_model.model.layers.36.self_attn.q_proj": true, "language_model.model.layers.36.self_attn.k_proj": true, "language_model.model.layers.36.self_attn.v_proj": true, "language_model.model.layers.36.self_attn.o_proj": true, "language_model.model.layers.36.self_attn.q_norm": false, "language_model.model.layers.36.self_attn.k_norm": false, "language_model.model.layers.36.self_attn.rope": false, "language_model.model.layers.36.mlp.gate_proj": true, "language_model.model.layers.36.mlp.down_proj": true, "language_model.model.layers.36.mlp.up_proj": true, "language_model.model.layers.36.input_layernorm": false, "language_model.model.layers.36.post_attention_layernorm": false, "language_model.model.layers.36.pre_feedforward_layernorm": false, "language_model.model.layers.36.post_feedforward_layernorm": false, "language_model.model.layers.37.self_attn.q_proj": true, "language_model.model.layers.37.self_attn.k_proj": true, "language_model.model.layers.37.self_attn.v_proj": true, "language_model.model.layers.37.self_attn.o_proj": true, "language_model.model.layers.37.self_attn.q_norm": false, "language_model.model.layers.37.self_attn.k_norm": false, "language_model.model.layers.37.self_attn.rope": false, "language_model.model.layers.37.mlp.gate_proj": true, "language_model.model.layers.37.mlp.down_proj": true, "language_model.model.layers.37.mlp.up_proj": true, "language_model.model.layers.37.input_layernorm": false, "language_model.model.layers.37.post_attention_layernorm": false, "language_model.model.layers.37.pre_feedforward_layernorm": false, "language_model.model.layers.37.post_feedforward_layernorm": false, "language_model.model.layers.38.self_attn.q_proj": true, "language_model.model.layers.38.self_attn.k_proj": true, "language_model.model.layers.38.self_attn.v_proj": true, "language_model.model.layers.38.self_attn.o_proj": true, "language_model.model.layers.38.self_attn.q_norm": false, "language_model.model.layers.38.self_attn.k_norm": false, "language_model.model.layers.38.self_attn.rope": false, "language_model.model.layers.38.mlp.gate_proj": true, "language_model.model.layers.38.mlp.down_proj": true, "language_model.model.layers.38.mlp.up_proj": true, "language_model.model.layers.38.input_layernorm": false, "language_model.model.layers.38.post_attention_layernorm": false, "language_model.model.layers.38.pre_feedforward_layernorm": false, "language_model.model.layers.38.post_feedforward_layernorm": false, "language_model.model.layers.39.self_attn.q_proj": true, "language_model.model.layers.39.self_attn.k_proj": true, "language_model.model.layers.39.self_attn.v_proj": true, "language_model.model.layers.39.self_attn.o_proj": true, "language_model.model.layers.39.self_attn.q_norm": false, "language_model.model.layers.39.self_attn.k_norm": false, "language_model.model.layers.39.self_attn.rope": false, "language_model.model.layers.39.mlp.gate_proj": true, "language_model.model.layers.39.mlp.down_proj": true, "language_model.model.layers.39.mlp.up_proj": true, "language_model.model.layers.39.input_layernorm": false, "language_model.model.layers.39.post_attention_layernorm": false, "language_model.model.layers.39.pre_feedforward_layernorm": false, "language_model.model.layers.39.post_feedforward_layernorm": false, "language_model.model.layers.40.self_attn.q_proj": true, "language_model.model.layers.40.self_attn.k_proj": true, "language_model.model.layers.40.self_attn.v_proj": true, "language_model.model.layers.40.self_attn.o_proj": true, "language_model.model.layers.40.self_attn.q_norm": false, "language_model.model.layers.40.self_attn.k_norm": false, "language_model.model.layers.40.self_attn.rope": false, "language_model.model.layers.40.mlp.gate_proj": true, "language_model.model.layers.40.mlp.down_proj": true, "language_model.model.layers.40.mlp.up_proj": true, "language_model.model.layers.40.input_layernorm": false, "language_model.model.layers.40.post_attention_layernorm": false, "language_model.model.layers.40.pre_feedforward_layernorm": false, "language_model.model.layers.40.post_feedforward_layernorm": false, "language_model.model.layers.41.self_attn.q_proj": true, "language_model.model.layers.41.self_attn.k_proj": true, "language_model.model.layers.41.self_attn.v_proj": true, "language_model.model.layers.41.self_attn.o_proj": true, "language_model.model.layers.41.self_attn.q_norm": false, "language_model.model.layers.41.self_attn.k_norm": false, "language_model.model.layers.41.self_attn.rope": false, "language_model.model.layers.41.mlp.gate_proj": true, "language_model.model.layers.41.mlp.down_proj": true, "language_model.model.layers.41.mlp.up_proj": true, "language_model.model.layers.41.input_layernorm": false, "language_model.model.layers.41.post_attention_layernorm": false, "language_model.model.layers.41.pre_feedforward_layernorm": false, "language_model.model.layers.41.post_feedforward_layernorm": false, "language_model.model.layers.42.self_attn.q_proj": true, "language_model.model.layers.42.self_attn.k_proj": true, "language_model.model.layers.42.self_attn.v_proj": true, "language_model.model.layers.42.self_attn.o_proj": true, "language_model.model.layers.42.self_attn.q_norm": false, "language_model.model.layers.42.self_attn.k_norm": false, "language_model.model.layers.42.self_attn.rope": false, "language_model.model.layers.42.mlp.gate_proj": true, "language_model.model.layers.42.mlp.down_proj": true, "language_model.model.layers.42.mlp.up_proj": true, "language_model.model.layers.42.input_layernorm": false, "language_model.model.layers.42.post_attention_layernorm": false, "language_model.model.layers.42.pre_feedforward_layernorm": false, "language_model.model.layers.42.post_feedforward_layernorm": false, "language_model.model.layers.43.self_attn.q_proj": true, "language_model.model.layers.43.self_attn.k_proj": true, "language_model.model.layers.43.self_attn.v_proj": true, "language_model.model.layers.43.self_attn.o_proj": true, "language_model.model.layers.43.self_attn.q_norm": false, "language_model.model.layers.43.self_attn.k_norm": false, "language_model.model.layers.43.self_attn.rope": false, "language_model.model.layers.43.mlp.gate_proj": true, "language_model.model.layers.43.mlp.down_proj": true, "language_model.model.layers.43.mlp.up_proj": true, "language_model.model.layers.43.input_layernorm": false, "language_model.model.layers.43.post_attention_layernorm": false, "language_model.model.layers.43.pre_feedforward_layernorm": false, "language_model.model.layers.43.post_feedforward_layernorm": false, "language_model.model.layers.44.self_attn.q_proj": true, "language_model.model.layers.44.self_attn.k_proj": true, "language_model.model.layers.44.self_attn.v_proj": true, "language_model.model.layers.44.self_attn.o_proj": true, "language_model.model.layers.44.self_attn.q_norm": false, "language_model.model.layers.44.self_attn.k_norm": false, "language_model.model.layers.44.self_attn.rope": false, "language_model.model.layers.44.mlp.gate_proj": true, "language_model.model.layers.44.mlp.down_proj": true, "language_model.model.layers.44.mlp.up_proj": true, "language_model.model.layers.44.input_layernorm": false, "language_model.model.layers.44.post_attention_layernorm": false, "language_model.model.layers.44.pre_feedforward_layernorm": false, "language_model.model.layers.44.post_feedforward_layernorm": false, "language_model.model.layers.45.self_attn.q_proj": true, "language_model.model.layers.45.self_attn.k_proj": true, "language_model.model.layers.45.self_attn.v_proj": true, "language_model.model.layers.45.self_attn.o_proj": true, "language_model.model.layers.45.self_attn.q_norm": false, "language_model.model.layers.45.self_attn.k_norm": false, "language_model.model.layers.45.self_attn.rope": false, "language_model.model.layers.45.mlp.gate_proj": true, "language_model.model.layers.45.mlp.down_proj": true, "language_model.model.layers.45.mlp.up_proj": true, "language_model.model.layers.45.input_layernorm": false, "language_model.model.layers.45.post_attention_layernorm": false, "language_model.model.layers.45.pre_feedforward_layernorm": false, "language_model.model.layers.45.post_feedforward_layernorm": false, "language_model.model.layers.46.self_attn.q_proj": true, "language_model.model.layers.46.self_attn.k_proj": true, "language_model.model.layers.46.self_attn.v_proj": true, "language_model.model.layers.46.self_attn.o_proj": true, "language_model.model.layers.46.self_attn.q_norm": false, "language_model.model.layers.46.self_attn.k_norm": false, "language_model.model.layers.46.self_attn.rope": false, "language_model.model.layers.46.mlp.gate_proj": true, "language_model.model.layers.46.mlp.down_proj": true, "language_model.model.layers.46.mlp.up_proj": true, "language_model.model.layers.46.input_layernorm": false, "language_model.model.layers.46.post_attention_layernorm": false, "language_model.model.layers.46.pre_feedforward_layernorm": false, "language_model.model.layers.46.post_feedforward_layernorm": false, "language_model.model.layers.47.self_attn.q_proj": true, "language_model.model.layers.47.self_attn.k_proj": true, "language_model.model.layers.47.self_attn.v_proj": true, "language_model.model.layers.47.self_attn.o_proj": true, "language_model.model.layers.47.self_attn.q_norm": false, "language_model.model.layers.47.self_attn.k_norm": false, "language_model.model.layers.47.self_attn.rope": false, "language_model.model.layers.47.mlp.gate_proj": true, "language_model.model.layers.47.mlp.down_proj": true, "language_model.model.layers.47.mlp.up_proj": true, "language_model.model.layers.47.input_layernorm": false, "language_model.model.layers.47.post_attention_layernorm": false, "language_model.model.layers.47.pre_feedforward_layernorm": false, "language_model.model.layers.47.post_feedforward_layernorm": false, "language_model.model.layers.48.self_attn.q_proj": true, "language_model.model.layers.48.self_attn.k_proj": true, "language_model.model.layers.48.self_attn.v_proj": true, "language_model.model.layers.48.self_attn.o_proj": true, "language_model.model.layers.48.self_attn.q_norm": false, "language_model.model.layers.48.self_attn.k_norm": false, "language_model.model.layers.48.self_attn.rope": false, "language_model.model.layers.48.mlp.gate_proj": true, "language_model.model.layers.48.mlp.down_proj": true, "language_model.model.layers.48.mlp.up_proj": true, "language_model.model.layers.48.input_layernorm": false, "language_model.model.layers.48.post_attention_layernorm": false, "language_model.model.layers.48.pre_feedforward_layernorm": false, "language_model.model.layers.48.post_feedforward_layernorm": false, "language_model.model.layers.49.self_attn.q_proj": true, "language_model.model.layers.49.self_attn.k_proj": true, "language_model.model.layers.49.self_attn.v_proj": true, "language_model.model.layers.49.self_attn.o_proj": true, "language_model.model.layers.49.self_attn.q_norm": false, "language_model.model.layers.49.self_attn.k_norm": false, "language_model.model.layers.49.self_attn.rope": false, "language_model.model.layers.49.mlp.gate_proj": true, "language_model.model.layers.49.mlp.down_proj": true, "language_model.model.layers.49.mlp.up_proj": true, "language_model.model.layers.49.input_layernorm": false, "language_model.model.layers.49.post_attention_layernorm": false, "language_model.model.layers.49.pre_feedforward_layernorm": false, "language_model.model.layers.49.post_feedforward_layernorm": false, "language_model.model.layers.50.self_attn.q_proj": true, "language_model.model.layers.50.self_attn.k_proj": true, "language_model.model.layers.50.self_attn.v_proj": true, "language_model.model.layers.50.self_attn.o_proj": true, "language_model.model.layers.50.self_attn.q_norm": false, "language_model.model.layers.50.self_attn.k_norm": false, "language_model.model.layers.50.self_attn.rope": false, "language_model.model.layers.50.mlp.gate_proj": true, "language_model.model.layers.50.mlp.down_proj": true, "language_model.model.layers.50.mlp.up_proj": true, "language_model.model.layers.50.input_layernorm": false, "language_model.model.layers.50.post_attention_layernorm": false, "language_model.model.layers.50.pre_feedforward_layernorm": false, "language_model.model.layers.50.post_feedforward_layernorm": false, "language_model.model.layers.51.self_attn.q_proj": true, "language_model.model.layers.51.self_attn.k_proj": true, "language_model.model.layers.51.self_attn.v_proj": true, "language_model.model.layers.51.self_attn.o_proj": true, "language_model.model.layers.51.self_attn.q_norm": false, "language_model.model.layers.51.self_attn.k_norm": false, "language_model.model.layers.51.self_attn.rope": false, "language_model.model.layers.51.mlp.gate_proj": true, "language_model.model.layers.51.mlp.down_proj": true, "language_model.model.layers.51.mlp.up_proj": true, "language_model.model.layers.51.input_layernorm": false, "language_model.model.layers.51.post_attention_layernorm": false, "language_model.model.layers.51.pre_feedforward_layernorm": false, "language_model.model.layers.51.post_feedforward_layernorm": false, "language_model.model.layers.52.self_attn.q_proj": true, "language_model.model.layers.52.self_attn.k_proj": true, "language_model.model.layers.52.self_attn.v_proj": true, "language_model.model.layers.52.self_attn.o_proj": true, "language_model.model.layers.52.self_attn.q_norm": false, "language_model.model.layers.52.self_attn.k_norm": false, "language_model.model.layers.52.self_attn.rope": false, "language_model.model.layers.52.mlp.gate_proj": true, "language_model.model.layers.52.mlp.down_proj": true, "language_model.model.layers.52.mlp.up_proj": true, "language_model.model.layers.52.input_layernorm": false, "language_model.model.layers.52.post_attention_layernorm": false, "language_model.model.layers.52.pre_feedforward_layernorm": false, "language_model.model.layers.52.post_feedforward_layernorm": false, "language_model.model.layers.53.self_attn.q_proj": true, "language_model.model.layers.53.self_attn.k_proj": true, "language_model.model.layers.53.self_attn.v_proj": true, "language_model.model.layers.53.self_attn.o_proj": true, "language_model.model.layers.53.self_attn.q_norm": false, "language_model.model.layers.53.self_attn.k_norm": false, "language_model.model.layers.53.self_attn.rope": false, "language_model.model.layers.53.mlp.gate_proj": true, "language_model.model.layers.53.mlp.down_proj": true, "language_model.model.layers.53.mlp.up_proj": true, "language_model.model.layers.53.input_layernorm": false, "language_model.model.layers.53.post_attention_layernorm": false, "language_model.model.layers.53.pre_feedforward_layernorm": false, "language_model.model.layers.53.post_feedforward_layernorm": false, "language_model.model.layers.54.self_attn.q_proj": true, "language_model.model.layers.54.self_attn.k_proj": true, "language_model.model.layers.54.self_attn.v_proj": true, "language_model.model.layers.54.self_attn.o_proj": true, "language_model.model.layers.54.self_attn.q_norm": false, "language_model.model.layers.54.self_attn.k_norm": false, "language_model.model.layers.54.self_attn.rope": false, "language_model.model.layers.54.mlp.gate_proj": true, "language_model.model.layers.54.mlp.down_proj": true, "language_model.model.layers.54.mlp.up_proj": true, "language_model.model.layers.54.input_layernorm": false, "language_model.model.layers.54.post_attention_layernorm": false, "language_model.model.layers.54.pre_feedforward_layernorm": false, "language_model.model.layers.54.post_feedforward_layernorm": false, "language_model.model.layers.55.self_attn.q_proj": true, "language_model.model.layers.55.self_attn.k_proj": true, "language_model.model.layers.55.self_attn.v_proj": true, "language_model.model.layers.55.self_attn.o_proj": true, "language_model.model.layers.55.self_attn.q_norm": false, "language_model.model.layers.55.self_attn.k_norm": false, "language_model.model.layers.55.self_attn.rope": false, "language_model.model.layers.55.mlp.gate_proj": true, "language_model.model.layers.55.mlp.down_proj": true, "language_model.model.layers.55.mlp.up_proj": true, "language_model.model.layers.55.input_layernorm": false, "language_model.model.layers.55.post_attention_layernorm": false, "language_model.model.layers.55.pre_feedforward_layernorm": false, "language_model.model.layers.55.post_feedforward_layernorm": false, "language_model.model.layers.56.self_attn.q_proj": true, "language_model.model.layers.56.self_attn.k_proj": true, "language_model.model.layers.56.self_attn.v_proj": true, "language_model.model.layers.56.self_attn.o_proj": true, "language_model.model.layers.56.self_attn.q_norm": false, "language_model.model.layers.56.self_attn.k_norm": false, "language_model.model.layers.56.self_attn.rope": false, "language_model.model.layers.56.mlp.gate_proj": true, "language_model.model.layers.56.mlp.down_proj": true, "language_model.model.layers.56.mlp.up_proj": true, "language_model.model.layers.56.input_layernorm": false, "language_model.model.layers.56.post_attention_layernorm": false, "language_model.model.layers.56.pre_feedforward_layernorm": false, "language_model.model.layers.56.post_feedforward_layernorm": false, "language_model.model.layers.57.self_attn.q_proj": true, "language_model.model.layers.57.self_attn.k_proj": true, "language_model.model.layers.57.self_attn.v_proj": true, "language_model.model.layers.57.self_attn.o_proj": true, "language_model.model.layers.57.self_attn.q_norm": false, "language_model.model.layers.57.self_attn.k_norm": false, "language_model.model.layers.57.self_attn.rope": false, "language_model.model.layers.57.mlp.gate_proj": true, "language_model.model.layers.57.mlp.down_proj": true, "language_model.model.layers.57.mlp.up_proj": true, "language_model.model.layers.57.input_layernorm": false, "language_model.model.layers.57.post_attention_layernorm": false, "language_model.model.layers.57.pre_feedforward_layernorm": false, "language_model.model.layers.57.post_feedforward_layernorm": false, "language_model.model.layers.58.self_attn.q_proj": true, "language_model.model.layers.58.self_attn.k_proj": true, "language_model.model.layers.58.self_attn.v_proj": true, "language_model.model.layers.58.self_attn.o_proj": true, "language_model.model.layers.58.self_attn.q_norm": false, "language_model.model.layers.58.self_attn.k_norm": false, "language_model.model.layers.58.self_attn.rope": false, "language_model.model.layers.58.mlp.gate_proj": true, "language_model.model.layers.58.mlp.down_proj": true, "language_model.model.layers.58.mlp.up_proj": true, "language_model.model.layers.58.input_layernorm": false, "language_model.model.layers.58.post_attention_layernorm": false, "language_model.model.layers.58.pre_feedforward_layernorm": false, "language_model.model.layers.58.post_feedforward_layernorm": false, "language_model.model.layers.59.self_attn.q_proj": true, "language_model.model.layers.59.self_attn.k_proj": true, "language_model.model.layers.59.self_attn.v_proj": true, "language_model.model.layers.59.self_attn.o_proj": true, "language_model.model.layers.59.self_attn.q_norm": false, "language_model.model.layers.59.self_attn.k_norm": false, "language_model.model.layers.59.self_attn.rope": false, "language_model.model.layers.59.mlp.gate_proj": true, "language_model.model.layers.59.mlp.down_proj": true, "language_model.model.layers.59.mlp.up_proj": true, "language_model.model.layers.59.input_layernorm": false, "language_model.model.layers.59.post_attention_layernorm": false, "language_model.model.layers.59.pre_feedforward_layernorm": false, "language_model.model.layers.59.post_feedforward_layernorm": false, "language_model.model.layers.60.self_attn.q_proj": true, "language_model.model.layers.60.self_attn.k_proj": true, "language_model.model.layers.60.self_attn.v_proj": true, "language_model.model.layers.60.self_attn.o_proj": true, "language_model.model.layers.60.self_attn.q_norm": false, "language_model.model.layers.60.self_attn.k_norm": false, "language_model.model.layers.60.self_attn.rope": false, "language_model.model.layers.60.mlp.gate_proj": true, "language_model.model.layers.60.mlp.down_proj": true, "language_model.model.layers.60.mlp.up_proj": true, "language_model.model.layers.60.input_layernorm": false, "language_model.model.layers.60.post_attention_layernorm": false, "language_model.model.layers.60.pre_feedforward_layernorm": false, "language_model.model.layers.60.post_feedforward_layernorm": false, "language_model.model.layers.61.self_attn.q_proj": true, "language_model.model.layers.61.self_attn.k_proj": true, "language_model.model.layers.61.self_attn.v_proj": true, "language_model.model.layers.61.self_attn.o_proj": true, "language_model.model.layers.61.self_attn.q_norm": false, "language_model.model.layers.61.self_attn.k_norm": false, "language_model.model.layers.61.self_attn.rope": false, "language_model.model.layers.61.mlp.gate_proj": true, "language_model.model.layers.61.mlp.down_proj": true, "language_model.model.layers.61.mlp.up_proj": true, "language_model.model.layers.61.input_layernorm": false, "language_model.model.layers.61.post_attention_layernorm": false, "language_model.model.layers.61.pre_feedforward_layernorm": false, "language_model.model.layers.61.post_feedforward_layernorm": false, "language_model.model.norm": false, "language_model.lm_head": true, "multi_modal_projector.mm_soft_emb_norm": false, "multi_modal_projector.avg_pool": false }, "remove_invalid_values": false, "repetition_penalty": 1.0, "return_dict": true, "return_dict_in_generate": false, "sep_token_id": null, "suppress_tokens": null, "task_specific_params": null, "temperature": 1.0, "text_config": { "return_dict": true, "output_hidden_states": false, "torchscript": false, "torch_dtype": "bfloat16", "use_bfloat16": false, "tf_legacy_loss": false, "pruned_heads": {}, "tie_word_embeddings": true, "chunk_size_feed_forward": 0, "is_encoder_decoder": false, "is_decoder": false, "cross_attention_hidden_size": null, "add_cross_attention": false, "tie_encoder_decoder": false, "max_length": 20, "min_length": 0, "do_sample": false, "early_stopping": false, "num_beams": 1, "num_beam_groups": 1, "diversity_penalty": 0.0, "temperature": 1.0, "top_k": 50, "top_p": 1.0, "typical_p": 1.0, "repetition_penalty": 1.0, "length_penalty": 1.0, "no_repeat_ngram_size": 0, "encoder_no_repeat_ngram_size": 0, "bad_words_ids": null, "num_return_sequences": 1, "output_scores": false, "return_dict_in_generate": false, "forced_bos_token_id": null, "forced_eos_token_id": null, "remove_invalid_values": false, "exponential_decay_length_penalty": null, "suppress_tokens": null, "begin_suppress_tokens": null, "architectures": null, "finetuning_task": null, "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "tokenizer_class": null, "prefix": null, "bos_token_id": 2, "pad_token_id": 0, "eos_token_id": 1, "sep_token_id": null, "decoder_start_token_id": null, "task_specific_params": null, "problem_type": null, "_name_or_path": "", "model_type": "gemma3_text", "vocab_size": 262208, "max_position_embeddings": 131072, "hidden_size": 5376, "intermediate_size": 21504, "num_hidden_layers": 62, "num_attention_heads": 32, "head_dim": 128, "num_key_value_heads": 16, "initializer_range": 0.02, "rms_norm_eps": 1e-06, "use_cache": true, "rope_theta": 1000000, "attention_bias": false, "attention_dropout": 0.0, "hidden_activation": "gelu_pytorch_tanh", "query_pre_attn_scalar": 168, "sliding_window": 1024, "final_logit_softcapping": null, "attn_logit_softcapping": null, "layer_types": [ "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention" ], "rope_local_base_freq": 10000, "rope_scaling": { "factor": 8.0, "rope_type": "linear" }, "output_attentions": false }, "tf_legacy_loss": false, "tie_encoder_decoder": false, "tie_word_embeddings": true, "tokenizer_class": null, "top_k": 50, "top_p": 1.0, "torchscript": false, "transformers_version": "4.53.1", "typical_p": 1.0, "use_bfloat16": false, "vision_config": { "return_dict": true, "output_hidden_states": false, "torchscript": false, "torch_dtype": "bfloat16", "use_bfloat16": false, "tf_legacy_loss": false, "pruned_heads": {}, "tie_word_embeddings": true, "chunk_size_feed_forward": 0, "is_encoder_decoder": false, "is_decoder": false, "cross_attention_hidden_size": null, "add_cross_attention": false, "tie_encoder_decoder": false, "max_length": 20, "min_length": 0, "do_sample": false, "early_stopping": false, "num_beams": 1, "num_beam_groups": 1, "diversity_penalty": 0.0, "temperature": 1.0, "top_k": 50, "top_p": 1.0, "typical_p": 1.0, "repetition_penalty": 1.0, "length_penalty": 1.0, "no_repeat_ngram_size": 0, "encoder_no_repeat_ngram_size": 0, "bad_words_ids": null, "num_return_sequences": 1, "output_scores": false, "return_dict_in_generate": false, "forced_bos_token_id": null, "forced_eos_token_id": null, "remove_invalid_values": false, "exponential_decay_length_penalty": null, "suppress_tokens": null, "begin_suppress_tokens": null, "architectures": null, "finetuning_task": null, "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "tokenizer_class": null, "prefix": null, "bos_token_id": null, "pad_token_id": null, "eos_token_id": null, "sep_token_id": null, "decoder_start_token_id": null, "task_specific_params": null, "problem_type": null, "_name_or_path": "", "model_type": "siglip_vision_model", "vision_use_head": false, "hidden_size": 1152, "intermediate_size": 4304, "num_hidden_layers": 27, "num_attention_heads": 16, "num_channels": 3, "patch_size": 14, "image_size": 896, "attention_dropout": 0.0, "layer_norm_eps": 1e-06, "hidden_act": "gelu_pytorch_tanh", "output_attentions": false } }