config.json

{
  "_name_or_path": "MoVA-8B",
  "architectures": [
    "MovaLlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128001,
  "expert_channels": [1536, 512, 256, 1536, 768, 768, 256],
  "freeze_mm_mlp_adapter": false,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "image_aspect_ratio": "pad",
  "image_feat_size": 48,
  "image_grid_pinpoints": null,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 8192,
  "mm_hidden_size": 1024,
  "mm_projector_freeze_text_encoder": false,
  "mm_projector_text_encoder": "bert-large-uncased",
  "mm_projector_type": "mov_adapter",
  "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": [
    "openai/clip-vit-large-patch14-336",
    "facebook/dinov2-large",
    "experts/vary-vit-base-patch16-1024",
    "experts/codetr-vit-large/pytorch_model.bin",
    "google/pix2struct-large",
    "google/deplot",
    "experts/BiomedCLIP-PubMedBERT_256-vit_base_patch16_224",
    "experts/sam-vit-huge/pytorch_model.bin"
  ],
  "model_type": "mova",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "num_projector_layers": 3,
  "num_routing_tokens": 144,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "topk_experts": 3,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.37.2",
  "tune_mm_mlp_adapter": false,
  "use_cache": true,
  "use_mm_proj": true,
  "vocab_size": 128256
}