VLM_WebSight_finetuned / config.json
VictorSanh's picture
perhaps it's the model type
b7b417b
raw
history blame
1.63 kB
{
"_commit_hash": null,
"_name_or_path": "None",
"additional_vocab_size": 2,
"alpha_initializer": "zeros",
"alpha_type": "float",
"alphas_initializer_range": 0.0,
"architectures": [
"Img2HTMLForVisionText2Text"
],
"attention_dropout": 0.0,
"auto_map": {
"AutoConfig": "configuration_img2html.Img2HTMLConfig",
"AutoModelForCausalLM": "modeling_img2html.Img2HTMLForVisionText2Text"
},
"bos_token_id": 1,
"cross_layer_interval": 1,
"eos_token_id": 2,
"freeze_lm_head": false,
"freeze_text_layers": false,
"freeze_text_module_exceptions": [],
"freeze_vision_layers": false,
"freeze_vision_module_exceptions": [],
"hidden_act": "silu",
"hidden_size": 4096,
"image_token_id": 32001,
"initializer_range": 0.02,
"intermediate_size": 14336,
"max_position_embeddings": 32768,
"model_type": "img2html",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"pad_token_id": 0,
"perceiver_config": {
"resampler_depth": 3,
"resampler_head_dim": 96,
"resampler_n_heads": 16,
"resampler_n_latents": 64,
"qk_layer_norms_perceiver": true
},
"qk_layer_norms": true,
"rms_norm_eps": 1e-05,
"rope_theta": 10000.0,
"sliding_window": 4096,
"tie_word_embeddings": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.34.0.dev0",
"use_cache": true,
"use_resampler": true,
"vision_config": {
"hidden_size": 1152,
"image_size": 960,
"intermediate_size": 4304,
"model_type": "img2html",
"num_attention_heads": 16,
"num_hidden_layers": 27,
"patch_size": 14
},
"vocab_size": 32000
}