Feature Extraction
Transformers
Safetensors
English
Chinese
emova
Omni-modal-LLM
Multi-modal-LLM
Emotional-spoken-dialogue
custom_code
Eval Results
KaiChen1998 commited on
Commit
ceaa304
·
verified ·
1 Parent(s): a228a4e

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +51 -52
config.json CHANGED
@@ -1,52 +1,51 @@
1
- {
2
- "_name_or_path": "D:\\temp\\emova_data\\checkpoints\\emova_qwen2_5_3b",
3
- "architectures": [
4
- "EMOVAForConditionalGeneration"
5
- ],
6
- "auto_map": {
7
- "AutoConfig": "configuration_emova.EMOVAConfig",
8
- "AutoModel": "modeling_emova.EMOVAForConditionalGeneration",
9
- "AutoModelForCausalLM": "modeling_emova.EMOVAForConditionalGeneration"
10
- },
11
- "default_system_prompt": "You are a helpful assistant.",
12
- "ignore_index": -100,
13
- "image_token_index": 155761,
14
- "mm_projector_config": {
15
- "mlp_depth": 2,
16
- "trainable": true,
17
- "type": "MLPProjector"
18
- },
19
- "model_type": "emova",
20
- "text_config": {
21
- "architectures": [
22
- "Qwen2ForCausalLM"
23
- ],
24
- "bos_token_id": 151643,
25
- "eos_token_id": 151645,
26
- "hidden_size": 2048,
27
- "intermediate_size": 11008,
28
- "max_position_embeddings": 32768,
29
- "max_window_layers": 70,
30
- "model_type": "qwen2",
31
- "num_attention_heads": 16,
32
- "num_hidden_layers": 36,
33
- "num_key_value_heads": 2,
34
- "rope_theta": 1000000.0,
35
- "sliding_window": null,
36
- "tie_word_embeddings": true,
37
- "torch_dtype": "float32",
38
- "use_sliding_window": false,
39
- "vocab_size": 155776
40
- },
41
- "tie_word_embeddings": false,
42
- "torch_dtype": "bfloat16",
43
- "transformers_version": "4.44.2",
44
- "vision_config": {
45
- "max_pixels": 3211264,
46
- "model_type": "qwen2_vl",
47
- "pretrained_model_name_or_path": "Emova-ollm/qwen2vit600m",
48
- "trainable": true,
49
- "type": "Qwen2VisionTower",
50
- "unfreeze_mm_vision_tower": true
51
- }
52
- }
 
1
+ {
2
+ "architectures": [
3
+ "EMOVAForConditionalGeneration"
4
+ ],
5
+ "auto_map": {
6
+ "AutoConfig": "configuration_emova.EMOVAConfig",
7
+ "AutoModel": "modeling_emova.EMOVAForConditionalGeneration",
8
+ "AutoModelForCausalLM": "modeling_emova.EMOVAForConditionalGeneration"
9
+ },
10
+ "default_system_prompt": "You are a helpful assistant.",
11
+ "ignore_index": -100,
12
+ "image_token_index": 155761,
13
+ "mm_projector_config": {
14
+ "mlp_depth": 2,
15
+ "trainable": true,
16
+ "type": "MLPProjector"
17
+ },
18
+ "model_type": "emova",
19
+ "text_config": {
20
+ "architectures": [
21
+ "Qwen2ForCausalLM"
22
+ ],
23
+ "bos_token_id": 151643,
24
+ "eos_token_id": 151645,
25
+ "hidden_size": 2048,
26
+ "intermediate_size": 11008,
27
+ "max_position_embeddings": 32768,
28
+ "max_window_layers": 70,
29
+ "model_type": "qwen2",
30
+ "num_attention_heads": 16,
31
+ "num_hidden_layers": 36,
32
+ "num_key_value_heads": 2,
33
+ "rope_theta": 1000000.0,
34
+ "sliding_window": null,
35
+ "tie_word_embeddings": true,
36
+ "torch_dtype": "float32",
37
+ "use_sliding_window": false,
38
+ "vocab_size": 155776
39
+ },
40
+ "tie_word_embeddings": false,
41
+ "torch_dtype": "bfloat16",
42
+ "transformers_version": "4.44.2",
43
+ "vision_config": {
44
+ "max_pixels": 3211264,
45
+ "model_type": "qwen2_vl",
46
+ "pretrained_model_name_or_path": "Emova-ollm/qwen2vit600m",
47
+ "trainable": true,
48
+ "type": "Qwen2VisionTower",
49
+ "unfreeze_mm_vision_tower": true
50
+ }
51
+ }