Hugging Face's logo

Gunslinger3D
/
fine-tuning-Phi2-with-webglm-qa-with-lora_4

PEFT
Safetensors
Generated from Trainer
Model card Files Community
Gunslinger3D commited on
Commit
009440b
·
verified ·
1 Parent(s): 709be23

fine-tuning-Phi2-with-webglm-qa-with-lora_4

Browse files
Files changed (4) hide show
  1. README.md +112 -0
  2. adapter_config.json +30 -0
  3. adapter_model.safetensors +3 -0
  4. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,112 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: peft
4
+ tags:
5
+ - generated_from_trainer
6
+ base_model: microsoft/phi-2
7
+ model-index:
8
+ - name: fine-tuning-Phi2-with-webglm-qa-with-lora_4
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # fine-tuning-Phi2-with-webglm-qa-with-lora_4
16
+
17
+ This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on an unknown dataset.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 0.1147
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 5e-05
39
+ - train_batch_size: 2
40
+ - eval_batch_size: 2
41
+ - seed: 42
42
+ - gradient_accumulation_steps: 5
43
+ - total_train_batch_size: 10
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: linear
46
+ - lr_scheduler_warmup_steps: 50
47
+ - training_steps: 500
48
+ - mixed_precision_training: Native AMP
49
+
50
+ ### Training results
51
+
52
+ | Training Loss | Epoch | Step | Validation Loss |
53
+ |:-------------:|:-----:|:----:|:---------------:|
54
+ | 8.2444 | 0.2 | 10 | 7.8267 |
55
+ | 7.4754 | 0.4 | 20 | 6.3605 |
56
+ | 4.8314 | 0.6 | 30 | 3.1457 |
57
+ | 1.7327 | 0.8 | 40 | 0.6363 |
58
+ | 0.5438 | 1.0 | 50 | 0.5673 |
59
+ | 0.4569 | 1.2 | 60 | 0.4906 |
60
+ | 0.4491 | 1.39 | 70 | 0.4269 |
61
+ | 0.367 | 1.59 | 80 | 0.3729 |
62
+ | 0.2821 | 1.79 | 90 | 0.3323 |
63
+ | 0.2414 | 1.99 | 100 | 0.3013 |
64
+ | 0.2521 | 2.19 | 110 | 0.2772 |
65
+ | 0.2135 | 2.39 | 120 | 0.2603 |
66
+ | 0.1982 | 2.59 | 130 | 0.2446 |
67
+ | 0.2186 | 2.79 | 140 | 0.2278 |
68
+ | 0.1741 | 2.99 | 150 | 0.2144 |
69
+ | 0.1781 | 3.19 | 160 | 0.2062 |
70
+ | 0.1702 | 3.39 | 170 | 0.1928 |
71
+ | 0.157 | 3.59 | 180 | 0.1846 |
72
+ | 0.1469 | 3.78 | 190 | 0.1770 |
73
+ | 0.1644 | 3.98 | 200 | 0.1705 |
74
+ | 0.1458 | 4.18 | 210 | 0.1654 |
75
+ | 0.1282 | 4.38 | 220 | 0.1623 |
76
+ | 0.1537 | 4.58 | 230 | 0.1568 |
77
+ | 0.1197 | 4.78 | 240 | 0.1509 |
78
+ | 0.1327 | 4.98 | 250 | 0.1464 |
79
+ | 0.1349 | 5.18 | 260 | 0.1436 |
80
+ | 0.1052 | 5.38 | 270 | 0.1409 |
81
+ | 0.127 | 5.58 | 280 | 0.1381 |
82
+ | 0.1303 | 5.78 | 290 | 0.1365 |
83
+ | 0.1063 | 5.98 | 300 | 0.1338 |
84
+ | 0.1145 | 6.18 | 310 | 0.1300 |
85
+ | 0.1101 | 6.37 | 320 | 0.1287 |
86
+ | 0.1088 | 6.57 | 330 | 0.1280 |
87
+ | 0.1062 | 6.77 | 340 | 0.1254 |
88
+ | 0.1016 | 6.97 | 350 | 0.1238 |
89
+ | 0.1005 | 7.17 | 360 | 0.1232 |
90
+ | 0.1084 | 7.37 | 370 | 0.1220 |
91
+ | 0.101 | 7.57 | 380 | 0.1204 |
92
+ | 0.1065 | 7.77 | 390 | 0.1200 |
93
+ | 0.0943 | 7.97 | 400 | 0.1191 |
94
+ | 0.0848 | 8.17 | 410 | 0.1184 |
95
+ | 0.0913 | 8.37 | 420 | 0.1175 |
96
+ | 0.1115 | 8.57 | 430 | 0.1169 |
97
+ | 0.091 | 8.76 | 440 | 0.1161 |
98
+ | 0.1009 | 8.96 | 450 | 0.1154 |
99
+ | 0.0966 | 9.16 | 460 | 0.1150 |
100
+ | 0.0931 | 9.36 | 470 | 0.1147 |
101
+ | 0.0922 | 9.56 | 480 | 0.1150 |
102
+ | 0.0912 | 9.76 | 490 | 0.1148 |
103
+ | 0.0915 | 9.96 | 500 | 0.1147 |
104
+
105
+
106
+ ### Framework versions
107
+
108
+ - PEFT 0.7.1
109
+ - Transformers 4.36.2
110
+ - Pytorch 2.0.0
111
+ - Datasets 2.15.0
112
+ - Tokenizers 0.15.0
adapter_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": null,
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layers_pattern": null,
10
+ "layers_to_transform": null,
11
+ "loftq_config": {},
12
+ "lora_alpha": 32,
13
+ "lora_dropout": 0.05,
14
+ "megatron_config": null,
15
+ "megatron_core": "megatron.core",
16
+ "modules_to_save": null,
17
+ "peft_type": "LORA",
18
+ "r": 16,
19
+ "rank_pattern": {},
20
+ "revision": null,
21
+ "target_modules": [
22
+ "q_proj",
23
+ "k_proj",
24
+ "v_proj",
25
+ "fc1",
26
+ "dense",
27
+ "fc2"
28
+ ],
29
+ "task_type": "CAUSAL_LM"
30
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:103eca31b19d0b8519424a6ce1f462ba1dc0d548234ecc1bf25988187a19c7c0
3
+ size 94428896
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46a91e763bd1d75bbe275e179c3a598225e0e03d68ed9c502f0b4b6287810304
3
+ size 4283