HuggingFaceTB/SmolVLM2-2.2B-Instruct

1{2  "_name_or_path": "/fsx/miquel/smolvlmvideo/checkpoints/final-visionUnfrozen-balanced/checkpoint-6550",3  "architectures": [4    "SmolVLMForConditionalGeneration"5  ],6  "image_token_id": 49190,7  "model_type": "smolvlm",8  "pad_token_id": 128002,9  "scale_factor": 3,10  "text_config": {11    "_flash_attn_2_enabled": true,12    "_name_or_path": "None",13    "architectures": [14      "VLlama3ForCausalLM"15    ],16    "head_dim": 64,17    "hidden_size": 2048,18    "intermediate_size": 8192,19    "max_position_embeddings": 8192,20    "model_type": "llama",21    "neftune_noise_alpha": 0.0,22    "num_hidden_layers": 24,23    "pad_token_id": 2,24    "perceiver_config": {25      "_attn_implementation_autoset": false,26      "_name_or_path": "",27      "add_cross_attention": false,28      "architectures": null,29      "attention_dropout": 0.0,30      "bad_words_ids": null,31      "begin_suppress_tokens": null,32      "bos_token_id": null,33      "chunk_size_feed_forward": 0,34      "cross_attention_hidden_size": null,35      "decoder_start_token_id": null,36      "diversity_penalty": 0.0,37      "do_sample": false,38      "early_stopping": false,39      "encoder_no_repeat_ngram_size": 0,40      "eos_token_id": null,41      "exponential_decay_length_penalty": null,42      "finetuning_task": null,43      "forced_bos_token_id": null,44      "forced_eos_token_id": null,45      "hidden_act": "silu",46      "id2label": {47        "0": "LABEL_0",48        "1": "LABEL_1"49      },50      "is_decoder": false,51      "is_encoder_decoder": false,52      "label2id": {53        "LABEL_0": 0,54        "LABEL_1": 155      },56      "length_penalty": 1.0,57      "max_length": 20,58      "min_length": 0,59      "model_type": "vllama3",60      "no_repeat_ngram_size": 0,61      "num_beam_groups": 1,62      "num_beams": 1,63      "num_key_value_heads": 1,64      "num_return_sequences": 1,65      "output_attentions": false,66      "output_hidden_states": false,67      "output_scores": false,68      "pad_token_id": null,69      "prefix": null,70      "problem_type": null,71      "pruned_heads": {},72      "qk_layer_norms_perceiver": false,73      "remove_invalid_values": false,74      "repetition_penalty": 1.0,75      "resampler_depth": 6,76      "resampler_head_dim": 96,77      "resampler_n_heads": 16,78      "resampler_n_latents": 64,79      "return_dict": true,80      "return_dict_in_generate": false,81      "sep_token_id": null,82      "suppress_tokens": null,83      "task_specific_params": null,84      "temperature": 1.0,85      "tf_legacy_loss": false,86      "tie_encoder_decoder": false,87      "tie_word_embeddings": true,88      "tokenizer_class": null,89      "top_k": 50,90      "top_p": 1.0,91      "torch_dtype": null,92      "torchscript": false,93      "transformers_version": "4.46.0",94      "typical_p": 1.0,95      "use_bfloat16": false96    },97    "pixel_shuffle_factor": 3,98    "qk_layer_norms": false,99    "rms_norm_eps": 1e-05,100    "rope_theta": 130000,101    "torch_dtype": "bfloat16",102    "transformers.js_config": {103      "kv_cache_dtype": {104        "fp16": "float16",105        "q4f16": "float16"106      }107    },108    "use_resampler": false,109    "vocab_size": 49280110  },111  "tie_word_embeddings": false,112  "torch_dtype": "float32",113  "transformers_version": "4.47.1",114  "use_cache": false,115  "use_reentrant_checkpointing": false,116  "vision_config": {117    "image_size": 384,118    "intermediate_size": 4304,119    "max_image_size": {120      "longest_edge": 384121    },122    "model_type": "smolvlm_vision",123    "num_hidden_layers": 27,124    "patch_size": 14,125    "size": {126      "longest_edge": 1920127    },128    "tie_word_embeddings": false,129    "use_base_siglip": false130  },131  "vocab_size": 49280132}133