llava_dit_stage1 / config.json
shijiayang
upload
1fbe192
{
"_flash_attn_2_enabled": true,
"_name_or_path": "/mnt/bn/bohanzhainas1/Public_Models/vicuna-7b-v1.5",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"ensemble_size": 1,
"eos_token_id": 2,
"freeze_mm_mlp_adapter": false,
"hidden_act": "silu",
"hidden_size": 4096,
"image_aspect_ratio": "square",
"image_grid_pinpoints": null,
"img_size": 512,
"initializer_range": 0.02,
"intermediate_size": 11008,
"max_position_embeddings": 4096,
"mm_hidden_size": 4608,
"mm_projector_type": "mlp2x_gelu",
"mm_use_im_patch_token": false,
"mm_use_im_start_end": false,
"mm_vision_select_feature": "patch",
"mm_vision_select_layer": -2,
"mm_vision_tower": "facebook/DiT-XL-2-512",
"model_type": "llava_llama",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 32,
"pad_token_id": 0,
"pretraining_tp": 1,
"prompt": "",
"rms_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 10000.0,
"t": 1,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.38.2",
"tune_mm_mlp_adapter": true,
"up_ft_index": -1,
"use_cache": true,
"use_mm_proj": true,
"vision_tower": "facebook/DiT-XL-2-512",
"vocab_size": 32000
}