config.json

{
  "architectures": [
    "MineralForConditionalGeneration"
  ],
  "model_type": "mineral",
  "vocab_size": 32000,
  "hidden_size": 2048,
  "intermediate_size": 5632,
  "num_hidden_layers": 22,
  "num_attention_heads": 16,
  "num_key_value_heads": 16,
  "hidden_act": "silu",
  "max_position_embeddings": 8192,
  "initializer_range": 0.02,
  "rms_norm_eps": 1e-06,
  "use_cache": true,
  "pad_token_id": 0,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "tie_word_embeddings": false,
  "rope_theta": 10000.0,
  "rope_scaling": null,
  "attention_bias": false,
  "attention_dropout": 0.0,
  "mlp_bias": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.40.0",
  "vision_config": {
    "hidden_size": 1024,
    "image_size": 384,
    "intermediate_size": 4096,
    "num_attention_heads": 16,
    "num_hidden_layers": 24,
    "patch_size": 14,
    "projection_dim": 2048,
    "vocab_size": 32000,
    "num_channels": 3,
    "layer_norm_eps": 1e-06,
    "hidden_act": "quick_gelu",
    "attention_dropout": 0.0
  },
  "vision_feature_layer": -2,
  "vision_feature_select_strategy": "default",
  "image_token_index": 32000,
  "projector_hidden_act": "gelu",
  "vision_start_token": "<|image_start|>",
  "vision_end_token": "<|image_end|>",
  "image_seq_length": 729,
  "ignore_index": -100
}