NVILA-15B / vision_tower /config.json
klldmofashi's picture
Upload files with `vila-upload`.
eb96957 verified
raw
history blame contribute delete
635 Bytes
{
"_name_or_path": "runs/train/qwen25-14b-dynamic-s2-stage3-20241129213531/model/vision_tower",
"architectures": [
"SiglipVisionModel"
],
"attention_dropout": 0.0,
"hidden_act": "gelu_pytorch_tanh",
"hidden_size": 1152,
"image_size": 448,
"intermediate_size": 4304,
"layer_norm_eps": 1e-06,
"model_type": "siglip_vision_model",
"num_attention_heads": 16,
"num_channels": 3,
"num_hidden_layers": 27,
"num_image_tokens": 256,
"patch_size": 14,
"projection_dim": 2048,
"projector_hidden_act": "gelu_fast",
"torch_dtype": "bfloat16",
"transformers_version": "4.46.0",
"vision_use_head": false
}