config.json

{
    "vocab_size": 32000,
    "hidden_size": 4096,
    "tie_word_embeddings": false,
    "max_position_embeddings": 2048,
    "num_hidden_layers": 32,
    "rms_norm_eps": 1e-05,
    "num_attention_heads": 32,
    "intermediate_size": 11008,
    "hidden_act": "silu",
    "mm_projector_type": "mlp2x_gelu",
    "tune_mm_mlp_adapter": false,
    "mm_vision_select_layer": -2,
    "mm_vision_select_feature": "patch",
    "mm_hidden_size": 1024,
    "rope_scaling": null,
    "model_type": "llava",
    "mm_use_im_patch_token": false,
    "mm_use_im_start_end": false,
    "bos_token_id": 1,
    "eos_token_id": 2,
    "image_aspect_ratio": "pad",
    "freeze_mm_mlp_adapter": false,
    "freeze_mm_vision_resampler": false,
    "architectures": [
        "LlavaLlamaForCausalLM"
    ],
    "pad_token_id": 0,
    "tune_mm_vision_resampler": false,
    "unfreeze_mm_vision_tower": false,
    "use_cache": true,
    "mm_vision_tower": "cerebras/Cerebras-ViT-L-336-patch14-llava7b-ShareGPT4V"
}