Image-Text-to-Text
Transformers
TensorBoard
Safetensors
feature-extraction
conversational
custom_code
Yin-Xie's picture
Upload config.json with huggingface_hub
ecabe9b verified
raw
history blame
2.51 kB
{
"architectures": [
"LLaVAOneVision1_5_ForConditionalGeneration"
],
"image_token_id": 151655,
"auto_map": {
"AutoConfig": "configuration_llavaonevision1_5.Llavaonevision1_5Config",
"AutoModel": "modeling_llavaonevision1_5.LLaVAOneVision1_5_ForConditionalGeneration",
"AutoModelForCausalLM": "modeling_llavaonevision1_5.LLaVAOneVision1_5_ForConditionalGeneration"
},
"text_config": {
"attention_bias": false,
"attention_dropout": 0.0,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 4096,
"image_token_id": null,
"initializer_range": 0.02,
"intermediate_size": 12288,
"tie_word_embeddings": false,
"layer_types": [
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention"
],
"max_position_embeddings": 32768,
"max_window_layers": 36,
"model_type": "LLaVAOneVision1_5_text",
"num_attention_heads": 32,
"num_hidden_layers": 36,
"num_key_value_heads": 8,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 1000000.0,
"sliding_window": null,
"use_cache": true,
"use_sliding_window": false,
"video_token_id": null,
"vocab_size": 151936
},
"torch_dtype": "bfloat16",
"transformers_version": "4.53.0",
"video_token_id": 151656,
"vision_config": {
"depth": 24,
"embed_dim": 1024,
"hidden_act": "gelu",
"hidden_size": 1024,
"in_channels": 3,
"initializer_range": 0.02,
"intermediate_size": 4096,
"layer_norm_eps": 1e-05,
"model_type": "rice_vit",
"num_heads": 16,
"patch_size": 14,
"spatial_merge_size": 2,
"temporal_patch_size": 1,
"text_hidden_size": 4096
},
"vocab_size": 151936
}