config.json

{
  "return_dict": true,
  "output_hidden_states": false,
  "torchscript": false,
  "dtype": null,
  "pruned_heads": {},
  "tie_word_embeddings": true,
  "chunk_size_feed_forward": 0,
  "is_encoder_decoder": false,
  "is_decoder": false,
  "cross_attention_hidden_size": null,
  "add_cross_attention": false,
  "tie_encoder_decoder": false,
  "architectures": [
    "VillanovaVLMForConditionalGeneration"
  ],
  "finetuning_task": null,
  "id2label": {
    "0": "LABEL_0",
    "1": "LABEL_1"
  },
  "label2id": {
    "LABEL_0": 0,
    "LABEL_1": 1
  },
  "task_specific_params": null,
  "problem_type": null,
  "tokenizer_class": null,
  "prefix": null,
  "bos_token_id": null,
  "pad_token_id": null,
  "eos_token_id": null,
  "sep_token_id": null,
  "decoder_start_token_id": null,
  "max_length": 20,
  "min_length": 0,
  "do_sample": false,
  "early_stopping": false,
  "num_beams": 1,
  "temperature": 1.0,
  "top_k": 50,
  "top_p": 1.0,
  "typical_p": 1.0,
  "repetition_penalty": 1.0,
  "length_penalty": 1.0,
  "no_repeat_ngram_size": 0,
  "encoder_no_repeat_ngram_size": 0,
  "bad_words_ids": null,
  "num_return_sequences": 1,
  "output_scores": false,
  "return_dict_in_generate": false,
  "forced_bos_token_id": null,
  "forced_eos_token_id": null,
  "remove_invalid_values": false,
  "exponential_decay_length_penalty": null,
  "suppress_tokens": null,
  "begin_suppress_tokens": null,
  "num_beam_groups": 1,
  "diversity_penalty": 0.0,
  "_name_or_path": "villanova-vlm",
  "transformers_version": "4.57.3",
  "tf_legacy_loss": false,
  "use_bfloat16": false,
  "vision_config": {
    "hidden_size": 1024,
    "image_size": 384,
    "patch_size": 16,
    "num_patches": 576,
    "num_hidden_layers": 24,
    "num_attention_heads": 16,
    "intermediate_size": 4096,
    "model_name": "ViT-L-16-SigLIP-384",
    "pretrained": "webli"
  },
  "projector_config": {
    "num_layers": 2,
    "input_size": 1024,
    "output_size": 2560,
    "hidden_size": 2560,
    "activation": "gelu",
    "use_layer_norm": false,
    "bias": true,
    "output_scale": 1.0
  },
  "text_config": {
    "architectures": [
      "VillanovaVLM"
    ],
    "auto_map": {
      "AutoConfig": "villanova_config.VillanovaVLMConfig",
      "AutoModelForCausalLM": "villanova_vlm.VillanovaVLM"
    },
    "bos_token_id": 1,
    "dtype": "float32",
    "eos_token_id": 2,
    "freeze_vision_encoder": true,
    "image_seq_length": 576,
    "image_token_index": 256000,
    "model_type": "villanova_vlm",
    "pad_token_id": 2,
    "projector_hidden_act": "gelu",
    "projector_hidden_size": 2560,
    "projector_num_layers": 2,
    "projector_output_scale": 1.0,
    "projector_use_output_norm": false,
    "text_config": {
      "_name_or_path": "/media/storage/store1/gabburo/models/villanova-sal-2b-w_const_pretrain_dcos1000k-1100k_to3e-6_v2-step=1099999",
      "architectures": [
        "LlamaForCausalLM"
      ],
      "attention_bias": false,
      "attention_dropout": 0.0,
      "dtype": "bfloat16",
      "head_dim": 128,
      "hidden_act": "silu",
      "hidden_size": 2560,
      "initializer_range": 0.014,
      "intermediate_size": 10240,
      "max_position_embeddings": 4096,
      "mlp_bias": false,
      "model_type": "llama",
      "num_attention_heads": 20,
      "num_hidden_layers": 18,
      "num_key_value_heads": 4,
      "pretraining_tp": 1,
      "rms_norm_eps": 1e-05,
      "rope_scaling": null,
      "rope_theta": 10000,
      "tie_word_embeddings": true,
      "use_cache": true,
      "vocab_size": 256001
    },
    "transformers_version": "4.57.3",
    "vision_config": {
      "backend": "openclip",
      "encoder_name": "ViT-L-16-SigLIP-384",
      "hidden_size": 1024,
      "image_size": 384,
      "num_patches": 576,
      "patch_size": 16,
      "pretrained": "webli"
    },
    "vision_feature_layer": -1,
    "vision_feature_select_strategy": "default"
  },
  "image_token_index": 256000,
  "vocab_size": 256001,
  "hidden_size": 2560,
  "model_type": "villanova",
  "output_attentions": false,
  "auto_map": {
    "AutoConfig": "configuration_villanova.VillanovaConfig",
    "AutoModelForImageTextToText": "modeling_villanova.VillanovaVLMForConditionalGeneration",
    "AutoProcessor": "processing_villanova.VillanovaProcessor"
  }
}