Upload 6 files
Browse files- README.md +9 -0
- adapter_config.json +27 -0
- adapter_model.bin +3 -0
- config.json +39 -0
- non_lora_trainables.bin +3 -0
- trainer_state.json +79 -0
README.md
CHANGED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: peft
|
3 |
+
---
|
4 |
+
## Training procedure
|
5 |
+
|
6 |
+
### Framework versions
|
7 |
+
|
8 |
+
|
9 |
+
- PEFT 0.4.0
|
adapter_config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"auto_mapping": null,
|
3 |
+
"base_model_name_or_path": "liuhaotian/llava-v1-0719-336px-lora-merge-vicuna-13b-v1.3",
|
4 |
+
"bias": "none",
|
5 |
+
"fan_in_fan_out": false,
|
6 |
+
"inference_mode": true,
|
7 |
+
"init_lora_weights": true,
|
8 |
+
"layers_pattern": null,
|
9 |
+
"layers_to_transform": null,
|
10 |
+
"lora_alpha": 16,
|
11 |
+
"lora_dropout": 0.05,
|
12 |
+
"modules_to_save": null,
|
13 |
+
"peft_type": "LORA",
|
14 |
+
"r": 64,
|
15 |
+
"revision": null,
|
16 |
+
"target_modules": [
|
17 |
+
"up_proj",
|
18 |
+
"v_proj",
|
19 |
+
"k_proj",
|
20 |
+
"mm_projector",
|
21 |
+
"gate_proj",
|
22 |
+
"q_proj",
|
23 |
+
"down_proj",
|
24 |
+
"o_proj"
|
25 |
+
],
|
26 |
+
"task_type": "CAUSAL_LM"
|
27 |
+
}
|
adapter_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c0f528aa4754c5d1a84d815ff787bda6f2bb285401e1c368c20e1707eb5e7de7
|
3 |
+
size 501676373
|
config.json
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "liuhaotian/llava-v1-0719-336px-lora-merge-vicuna-13b-v1.3",
|
3 |
+
"architectures": [
|
4 |
+
"LlavaLlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"bos_token_id": 1,
|
7 |
+
"eos_token_id": 2,
|
8 |
+
"freeze_mm_mlp_adapter": false,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 5120,
|
11 |
+
"image_aspect_ratio": "square",
|
12 |
+
"image_grid_pinpoints": null,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 13824,
|
15 |
+
"max_position_embeddings": 2048,
|
16 |
+
"mm_hidden_size": 1024,
|
17 |
+
"mm_resampler_type": null,
|
18 |
+
"mm_use_im_patch_token": false,
|
19 |
+
"mm_use_im_start_end": false,
|
20 |
+
"mm_vision_select_feature": "patch",
|
21 |
+
"mm_vision_select_layer": -2,
|
22 |
+
"mm_vision_tower": "openai/clip-vit-large-patch14-336",
|
23 |
+
"model_type": "llava",
|
24 |
+
"num_attention_heads": 40,
|
25 |
+
"num_hidden_layers": 40,
|
26 |
+
"num_key_value_heads": 40,
|
27 |
+
"pad_token_id": 0,
|
28 |
+
"pretraining_tp": 1,
|
29 |
+
"rms_norm_eps": 1e-06,
|
30 |
+
"rope_scaling": null,
|
31 |
+
"tie_word_embeddings": false,
|
32 |
+
"torch_dtype": "float16",
|
33 |
+
"transformers_version": "4.31.0",
|
34 |
+
"tune_mm_mlp_adapter": false,
|
35 |
+
"tune_mm_vision_resampler": false,
|
36 |
+
"use_cache": true,
|
37 |
+
"use_mm_proj": true,
|
38 |
+
"vocab_size": 32000
|
39 |
+
}
|
non_lora_trainables.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cc6b9cc23771073378c3887f3524986a3b61cac85cbc085f322b318241bc6845
|
3 |
+
size 455
|
trainer_state.json
ADDED
@@ -0,0 +1,79 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.0,
|
5 |
+
"global_step": 9,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.33,
|
12 |
+
"learning_rate": 2e-05,
|
13 |
+
"loss": 2.0078,
|
14 |
+
"step": 1
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.67,
|
18 |
+
"learning_rate": 1.9238795325112867e-05,
|
19 |
+
"loss": 1.8555,
|
20 |
+
"step": 2
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 1.0,
|
24 |
+
"learning_rate": 1.7071067811865477e-05,
|
25 |
+
"loss": 1.9688,
|
26 |
+
"step": 3
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 1.33,
|
30 |
+
"learning_rate": 1.3826834323650899e-05,
|
31 |
+
"loss": 1.9648,
|
32 |
+
"step": 4
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 1.67,
|
36 |
+
"learning_rate": 1e-05,
|
37 |
+
"loss": 1.9531,
|
38 |
+
"step": 5
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 2.0,
|
42 |
+
"learning_rate": 6.173165676349103e-06,
|
43 |
+
"loss": 1.8594,
|
44 |
+
"step": 6
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 2.33,
|
48 |
+
"learning_rate": 2.9289321881345257e-06,
|
49 |
+
"loss": 1.8164,
|
50 |
+
"step": 7
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 2.67,
|
54 |
+
"learning_rate": 7.612046748871327e-07,
|
55 |
+
"loss": 2.0547,
|
56 |
+
"step": 8
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 3.0,
|
60 |
+
"learning_rate": 0.0,
|
61 |
+
"loss": 1.8125,
|
62 |
+
"step": 9
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 3.0,
|
66 |
+
"step": 9,
|
67 |
+
"total_flos": 3615157490024448.0,
|
68 |
+
"train_loss": 1.9214409722222223,
|
69 |
+
"train_runtime": 539.9369,
|
70 |
+
"train_samples_per_second": 0.533,
|
71 |
+
"train_steps_per_second": 0.017
|
72 |
+
}
|
73 |
+
],
|
74 |
+
"max_steps": 9,
|
75 |
+
"num_train_epochs": 3,
|
76 |
+
"total_flos": 3615157490024448.0,
|
77 |
+
"trial_name": null,
|
78 |
+
"trial_params": null
|
79 |
+
}
|