config.json

{
 "_name_or_path": "work_dirs/llava_phi3_mini_4k_instruct_full_clip_vit_large_p14_336_full_e2_gpu8_internvl_finetune/iter_19810_hf/llama_llm",
 "architectures": [
 "LlavaLlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 32000,
 "hidden_act": "silu",
 "hidden_size": 3072,
 "image_aspect_ratio": "pad",
 "initializer_range": 0.02,
 "intermediate_size": 8192,
 "max_position_embeddings": 4096,
 "mm_hidden_size": 1024,
 "mm_projector_type": "mlp2x_gelu",
 "mm_use_im_patch_token": false,
 "mm_use_im_start_end": false,
 "mm_vision_select_feature": "patch",
 "mm_vision_select_layer": -2,
 "mm_vision_tower": "work_dirs/llava_phi3_mini_4k_instruct_full_clip_vit_large_p14_336_full_e2_gpu8_internvl_finetune/iter_19810_hf/visual_encoder",
 "model_type": "llava_llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 32,
 "num_key_value_heads": 32,
 "original_max_position_embeddings": 4096,
 "pad_token_id": 32000,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "sliding_window": 2048,
 "tie_word_embeddings": false,
 "torch_dtype": "float16",
 "transformers_version": "4.40.1",
 "unfreeze_mm_vision_tower": true,
 "use_cache": true,
 "use_mm_proj": true,
 "vocab_size": 32064
}