config.json

{
  "_name_or_path": "",
  "architectures": [
    "OryxLlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "freeze_mm_mlp_adapter": false,
  "freeze_mm_vision_resampler": false,
  "hidden_act": "silu",
  "hidden_size": 7168,
  "initializer_range": 0.02,
  "intermediate_size": 20480,
  "max_position_embeddings": 16384,
  "mlp_bias": false,
  "mm_hidden_size": 1152,
  "mm_projector_lr": null,
  "mm_projector_type": "simple_mlp_twoview",
  "mm_resampler_type": "dynamic_compressor",
  "mm_use_im_patch_token": false,
  "mm_use_im_start_end": false,
  "mm_vision_select_feature": "patch",
  "mm_vision_select_layer": -1,
  "mm_vision_tower": "oryx_vit:/path/to/oryx_vit",
  "mm_vision_tower_lr": null,
  "model_type": "oryx_llama",
  "num_attention_heads": 56,
  "num_hidden_layers": 60,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 5000000,
  "tie_word_embeddings": false,
  "tokenizer_model_max_length": 16384,
  "tokenizer_padding_side": "right",
  "torch_dtype": "bfloat16",
  "transformers_version": "4.39.2",
  "tune_mm_mlp_adapter": false,
  "tune_mm_vision_resampler": false,
  "unfreeze_mm_vision_tower": false,
  "use_cache": true,
  "use_mm_proj": true,
  "vocab_size": 64000
}