Upload config.json with huggingface_hub
Browse files- config.json +38 -0
config.json
ADDED
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "llava-hf/llava-1.5-7b-hf",
|
3 |
+
"architectures": [
|
4 |
+
"LlavaForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"ignore_index": -100,
|
7 |
+
"image_token_index": 32000,
|
8 |
+
"model_type": "llava",
|
9 |
+
"pad_token_id": 32001,
|
10 |
+
"projector_hidden_act": "gelu",
|
11 |
+
"text_config": {
|
12 |
+
"_name_or_path": "lmsys/vicuna-7b-v1.5",
|
13 |
+
"architectures": [
|
14 |
+
"LlamaForCausalLM"
|
15 |
+
],
|
16 |
+
"max_position_embeddings": 4096,
|
17 |
+
"model_type": "llama",
|
18 |
+
"rms_norm_eps": 1e-05,
|
19 |
+
"torch_dtype": "float16",
|
20 |
+
"vocab_size": 32064
|
21 |
+
},
|
22 |
+
"tie_word_embeddings": false,
|
23 |
+
"transformers_version": "4.39.3",
|
24 |
+
"use_cache": true,
|
25 |
+
"vision_config": {
|
26 |
+
"hidden_size": 1024,
|
27 |
+
"image_size": 336,
|
28 |
+
"intermediate_size": 4096,
|
29 |
+
"model_type": "clip_vision_model",
|
30 |
+
"num_attention_heads": 16,
|
31 |
+
"num_hidden_layers": 24,
|
32 |
+
"patch_size": 14,
|
33 |
+
"projection_dim": 768,
|
34 |
+
"vocab_size": 32000
|
35 |
+
},
|
36 |
+
"vision_feature_layer": -2,
|
37 |
+
"vision_feature_select_strategy": "default"
|
38 |
+
}
|