zhangbofei commited on
Commit
7514121
1 Parent(s): c364545

feat: fix .json

Browse files
.gitignore CHANGED
@@ -1,7 +1,5 @@
1
  dependency
2
- *.json
3
  *.log
4
  *.pyc
5
  *__pycache__*
6
- *.json
7
  serve_images
 
1
  dependency
 
2
  *.log
3
  *.pyc
4
  *__pycache__*
 
5
  serve_images
checkpoints/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "Lin-Chen/open-llava-next-llama3-8b",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 256,
14
+ "lora_dropout": 0.05,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 64,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "k_proj",
24
+ "q_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoints/config.json ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Lin-Chen/open-llava-next-llama3-8b",
3
+ "architectures": [
4
+ "LlavaLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 128000,
9
+ "eos_token_id": 128001,
10
+ "freeze_mm_mlp_adapter": false,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 4096,
13
+ "image_aspect_ratio": "pad",
14
+ "image_grid_pinpoints": [
15
+ [
16
+ 336,
17
+ 672
18
+ ],
19
+ [
20
+ 672,
21
+ 336
22
+ ],
23
+ [
24
+ 672,
25
+ 672
26
+ ],
27
+ [
28
+ 1008,
29
+ 336
30
+ ],
31
+ [
32
+ 336,
33
+ 1008
34
+ ]
35
+ ],
36
+ "initializer_range": 0.02,
37
+ "intermediate_size": 14336,
38
+ "max_position_embeddings": 8192,
39
+ "mm_hidden_size": 1024,
40
+ "mm_patch_merge_type": "flat",
41
+ "mm_projector_lr": null,
42
+ "mm_projector_type": "mlp2x_gelu",
43
+ "mm_use_im_patch_token": false,
44
+ "mm_use_im_start_end": false,
45
+ "mm_vision_select_feature": "patch",
46
+ "mm_vision_select_layer": -2,
47
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
48
+ "mm_vision_tower_lr": 2e-06,
49
+ "model_type": "llava_llama",
50
+ "num_attention_heads": 32,
51
+ "num_hidden_layers": 32,
52
+ "num_key_value_heads": 8,
53
+ "pad_token_id": 128256,
54
+ "pretraining_tp": 1,
55
+ "rms_norm_eps": 1e-05,
56
+ "rope_scaling": null,
57
+ "rope_theta": 500000.0,
58
+ "tie_word_embeddings": false,
59
+ "tokenizer_model_max_length": 2048,
60
+ "tokenizer_padding_side": "right",
61
+ "torch_dtype": "bfloat16",
62
+ "transformers_version": "4.37.2",
63
+ "tune_mm_mlp_adapter": false,
64
+ "unfreeze_mm_vision_tower": true,
65
+ "use_cache": true,
66
+ "use_mm_proj": true,
67
+ "vocab_size": 128257
68
+ }
checkpoints/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff