Visual Question Answering
Transformers
English
qwen2
text-generation
multimodal large language model
large video-language model
Inference Endpoints
lixin4ever commited on
Commit
a7126d9
·
verified ·
1 Parent(s): 2d704ac

Upload 2 files

Browse files
Files changed (2) hide show
  1. config.json +43 -0
  2. trainer_state.json +0 -0
config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Qwen/Qwen2-72B-Instruct",
3
+ "architectures": [
4
+ "Qwen2ForCausalLM"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 151643,
8
+ "eos_token_id": 151645,
9
+ "freeze_mm_mlp_adapter": false,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 8192,
12
+ "image_aspect_ratio": "square",
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 29568,
15
+ "max_position_embeddings": 32768,
16
+ "max_window_layers": 80,
17
+ "mm_hidden_size": 1024,
18
+ "mm_projector_lr": null,
19
+ "mm_projector_type": "stc_connector",
20
+ "mm_use_im_patch_token": false,
21
+ "mm_use_im_start_end": false,
22
+ "mm_vision_select_feature": "patch",
23
+ "mm_vision_select_layer": -2,
24
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
25
+ "model_type": "qwen2",
26
+ "num_attention_heads": 64,
27
+ "num_frames": 8,
28
+ "num_hidden_layers": 80,
29
+ "num_key_value_heads": 8,
30
+ "rms_norm_eps": 1e-06,
31
+ "rope_theta": 1000000.0,
32
+ "sliding_window": 131072,
33
+ "tie_word_embeddings": false,
34
+ "tokenizer_model_max_length": 2048,
35
+ "tokenizer_padding_side": "right",
36
+ "torch_dtype": "bfloat16",
37
+ "transformers_version": "4.41.2",
38
+ "tune_mm_mlp_adapter": true,
39
+ "use_cache": true,
40
+ "use_mm_proj": true,
41
+ "use_sliding_window": false,
42
+ "vocab_size": 152064
43
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff