Hugging Face's logo

just1nseo
/
zephyr-dpo-qlora-uf-ours-uffull-5e-7

PEFT
TensorBoard
Safetensors
mistral
alignment-handbook
trl
dpo
Generated from Trainer
4-bit precision
bitsandbytes
Model card Files Metrics Community
just1nseo commited on
Commit
45f5c7c
•
1 Parent(s): 49045bf

Model save

Browse files
Files changed (6) hide show
  1. README.md +118 -0
  2. adapter_model.safetensors +1 -1
  3. all_results.json +8 -0
  4. runs/Jul16_03-11-51_notebook-deployment-48-7d9b6c99-p5kv4/events.out.tfevents.1721100016.notebook-deployment-48-7d9b6c99-p5kv4.69914.0 +2 -2
  5. train_results.json +8 -0
  6. trainer_state.json +0 -0
README.md ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: alignment-handbook/zephyr-7b-sft-full
3
+ library_name: peft
4
+ license: apache-2.0
5
+ tags:
6
+ - trl
7
+ - dpo
8
+ - generated_from_trainer
9
+ model-index:
10
+ - name: zephyr-dpo-qlora-uf-ours-uffull-5e-7
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # zephyr-dpo-qlora-uf-ours-uffull-5e-7
18
+
19
+ This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.5925
22
+ - Rewards/chosen: -0.2025
23
+ - Rewards/rejected: -0.5175
24
+ - Rewards/accuracies: 0.7060
25
+ - Rewards/margins: 0.3149
26
+ - Rewards/margins Max: 1.1833
27
+ - Rewards/margins Min: -0.5415
28
+ - Rewards/margins Std: 0.5824
29
+ - Logps/rejected: -317.5993
30
+ - Logps/chosen: -304.7070
31
+ - Logits/rejected: -2.5752
32
+ - Logits/chosen: -2.6056
33
+
34
+ ## Model description
35
+
36
+ More information needed
37
+
38
+ ## Intended uses & limitations
39
+
40
+ More information needed
41
+
42
+ ## Training and evaluation data
43
+
44
+ More information needed
45
+
46
+ ## Training procedure
47
+
48
+ ### Training hyperparameters
49
+
50
+ The following hyperparameters were used during training:
51
+ - learning_rate: 5e-07
52
+ - train_batch_size: 4
53
+ - eval_batch_size: 8
54
+ - seed: 42
55
+ - distributed_type: multi-GPU
56
+ - num_devices: 2
57
+ - gradient_accumulation_steps: 2
58
+ - total_train_batch_size: 16
59
+ - total_eval_batch_size: 16
60
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
61
+ - lr_scheduler_type: cosine
62
+ - lr_scheduler_warmup_ratio: 0.1
63
+ - num_epochs: 1
64
+
65
+ ### Training results
66
+
67
+ | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Rewards/margins Max | Rewards/margins Min | Rewards/margins Std | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
68
+ |:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:-------------------:|:-------------------:|:-------------------:|:--------------:|:------------:|:---------------:|:-------------:|
69
+ | 0.6931 | 0.02 | 100 | 0.6930 | 0.0002 | -0.0000 | 0.5130 | 0.0002 | 0.0047 | -0.0043 | 0.0030 | -265.8543 | -284.4388 | -2.7682 | -2.8031 |
70
+ | 0.692 | 0.05 | 200 | 0.6923 | 0.0017 | 0.0000 | 0.6220 | 0.0017 | 0.0099 | -0.0057 | 0.0051 | -265.8525 | -284.2892 | -2.7668 | -2.8017 |
71
+ | 0.6903 | 0.07 | 300 | 0.6908 | 0.0067 | 0.0019 | 0.6520 | 0.0048 | 0.0253 | -0.0125 | 0.0125 | -265.6623 | -283.7856 | -2.7627 | -2.7978 |
72
+ | 0.6888 | 0.1 | 400 | 0.6880 | 0.0104 | -0.0004 | 0.6645 | 0.0108 | 0.0545 | -0.0264 | 0.0268 | -265.8943 | -283.4167 | -2.7573 | -2.7924 |
73
+ | 0.6827 | 0.12 | 500 | 0.6834 | 0.0345 | 0.0138 | 0.6820 | 0.0207 | 0.0989 | -0.0454 | 0.0479 | -264.4715 | -281.0052 | -2.7529 | -2.7877 |
74
+ | 0.6831 | 0.14 | 600 | 0.6776 | 0.0296 | -0.0039 | 0.6910 | 0.0335 | 0.1552 | -0.0696 | 0.0745 | -266.2422 | -281.4937 | -2.7479 | -2.7827 |
75
+ | 0.6652 | 0.17 | 700 | 0.6700 | 0.0086 | -0.0427 | 0.6820 | 0.0513 | 0.2350 | -0.1057 | 0.1128 | -270.1202 | -283.5948 | -2.7382 | -2.7726 |
76
+ | 0.6486 | 0.19 | 800 | 0.6615 | -0.0198 | -0.0921 | 0.6805 | 0.0723 | 0.3237 | -0.1470 | 0.1565 | -275.0622 | -286.4378 | -2.7367 | -2.7702 |
77
+ | 0.6457 | 0.22 | 900 | 0.6531 | -0.0599 | -0.1549 | 0.6755 | 0.0950 | 0.4216 | -0.1947 | 0.2059 | -281.3418 | -290.4436 | -2.7168 | -2.7500 |
78
+ | 0.6356 | 0.24 | 1000 | 0.6449 | -0.0625 | -0.1814 | 0.6785 | 0.1188 | 0.5225 | -0.2486 | 0.2583 | -283.9890 | -290.7086 | -2.7042 | -2.7362 |
79
+ | 0.6465 | 0.26 | 1100 | 0.6378 | -0.0291 | -0.1702 | 0.6775 | 0.1411 | 0.6108 | -0.2946 | 0.3031 | -282.8690 | -287.3659 | -2.6982 | -2.7301 |
80
+ | 0.6121 | 0.29 | 1200 | 0.6317 | -0.0658 | -0.2261 | 0.6780 | 0.1603 | 0.6847 | -0.3354 | 0.3418 | -288.4626 | -291.0350 | -2.6893 | -2.7208 |
81
+ | 0.6113 | 0.31 | 1300 | 0.6287 | -0.1819 | -0.3556 | 0.6820 | 0.1737 | 0.7287 | -0.3416 | 0.3621 | -301.4144 | -302.6470 | -2.6941 | -2.7251 |
82
+ | 0.6058 | 0.34 | 1400 | 0.6234 | -0.1290 | -0.3204 | 0.6775 | 0.1914 | 0.7908 | -0.3943 | 0.3995 | -297.8902 | -297.3538 | -2.6823 | -2.7135 |
83
+ | 0.6169 | 0.36 | 1500 | 0.6194 | -0.1244 | -0.3286 | 0.6790 | 0.2042 | 0.8341 | -0.4094 | 0.4197 | -298.7180 | -296.9003 | -2.6648 | -2.6957 |
84
+ | 0.5809 | 0.38 | 1600 | 0.6163 | -0.1125 | -0.3291 | 0.6800 | 0.2167 | 0.8823 | -0.4243 | 0.4399 | -298.7659 | -295.7021 | -2.6547 | -2.6853 |
85
+ | 0.5979 | 0.41 | 1700 | 0.6161 | -0.2126 | -0.4403 | 0.6805 | 0.2276 | 0.9153 | -0.4469 | 0.4624 | -309.8821 | -305.7201 | -2.6466 | -2.6773 |
86
+ | 0.6034 | 0.43 | 1800 | 0.6124 | -0.1652 | -0.4014 | 0.6805 | 0.2362 | 0.9410 | -0.4507 | 0.4726 | -305.9889 | -300.9712 | -2.6365 | -2.6672 |
87
+ | 0.5983 | 0.45 | 1900 | 0.6144 | -0.0531 | -0.2743 | 0.6900 | 0.2212 | 0.8923 | -0.3931 | 0.4327 | -293.2797 | -289.7628 | -2.6389 | -2.6689 |
88
+ | 0.5822 | 0.48 | 2000 | 0.6049 | -0.1502 | -0.4096 | 0.6885 | 0.2593 | 1.0070 | -0.4697 | 0.4998 | -306.8109 | -299.4801 | -2.6378 | -2.6679 |
89
+ | 0.6013 | 0.5 | 2100 | 0.6034 | -0.1787 | -0.4453 | 0.6870 | 0.2666 | 1.0331 | -0.4819 | 0.5137 | -310.3860 | -302.3300 | -2.6289 | -2.6593 |
90
+ | 0.6018 | 0.53 | 2200 | 0.6019 | -0.1572 | -0.4295 | 0.6925 | 0.2723 | 1.0473 | -0.4896 | 0.5205 | -308.8055 | -300.1773 | -2.6287 | -2.6585 |
91
+ | 0.6121 | 0.55 | 2300 | 0.6010 | -0.2434 | -0.5217 | 0.6905 | 0.2783 | 1.0633 | -0.4893 | 0.5289 | -318.0273 | -308.7991 | -2.6178 | -2.6476 |
92
+ | 0.5698 | 0.57 | 2400 | 0.5979 | -0.1902 | -0.4780 | 0.6920 | 0.2878 | 1.0879 | -0.4939 | 0.5369 | -313.6557 | -303.4752 | -2.6092 | -2.6389 |
93
+ | 0.5656 | 0.6 | 2500 | 0.5992 | -0.2708 | -0.5597 | 0.6985 | 0.2889 | 1.0980 | -0.5097 | 0.5454 | -321.8217 | -311.5382 | -2.5991 | -2.6291 |
94
+ | 0.5795 | 0.62 | 2600 | 0.5950 | -0.2109 | -0.5113 | 0.6950 | 0.3003 | 1.1206 | -0.5079 | 0.5533 | -316.9805 | -305.5476 | -2.5944 | -2.6244 |
95
+ | 0.5909 | 0.65 | 2700 | 0.5945 | -0.2006 | -0.5044 | 0.6950 | 0.3038 | 1.1335 | -0.5150 | 0.5598 | -316.2979 | -304.5152 | -2.5934 | -2.6235 |
96
+ | 0.6097 | 0.67 | 2800 | 0.5938 | -0.2035 | -0.5091 | 0.6975 | 0.3055 | 1.1391 | -0.5171 | 0.5610 | -316.7604 | -304.8101 | -2.5909 | -2.6210 |
97
+ | 0.5776 | 0.69 | 2900 | 0.5929 | -0.2142 | -0.5232 | 0.7040 | 0.3091 | 1.1530 | -0.5251 | 0.5673 | -318.1778 | -305.8716 | -2.5874 | -2.6177 |
98
+ | 0.575 | 0.72 | 3000 | 0.5948 | -0.1848 | -0.4886 | 0.6980 | 0.3039 | 1.1465 | -0.5243 | 0.5647 | -314.7165 | -302.9333 | -2.5861 | -2.6165 |
99
+ | 0.5767 | 0.74 | 3100 | 0.5936 | -0.1972 | -0.5061 | 0.7010 | 0.3089 | 1.1551 | -0.5276 | 0.5690 | -316.4648 | -304.1734 | -2.5862 | -2.6166 |
100
+ | 0.5642 | 0.77 | 3200 | 0.5937 | -0.1943 | -0.5034 | 0.7010 | 0.3091 | 1.1615 | -0.5332 | 0.5726 | -316.1906 | -303.8846 | -2.5867 | -2.6170 |
101
+ | 0.5767 | 0.79 | 3300 | 0.5914 | -0.2376 | -0.5569 | 0.7050 | 0.3193 | 1.1828 | -0.5330 | 0.5823 | -321.5458 | -308.2144 | -2.5828 | -2.6131 |
102
+ | 0.5685 | 0.81 | 3400 | 0.5914 | -0.2246 | -0.5434 | 0.7045 | 0.3188 | 1.1858 | -0.5380 | 0.5834 | -320.1958 | -306.9150 | -2.5800 | -2.6103 |
103
+ | 0.5687 | 0.84 | 3500 | 0.5909 | -0.2343 | -0.5556 | 0.7045 | 0.3214 | 1.1905 | -0.5370 | 0.5855 | -321.4169 | -307.8832 | -2.5779 | -2.6082 |
104
+ | 0.5598 | 0.86 | 3600 | 0.5924 | -0.2063 | -0.5212 | 0.7060 | 0.3150 | 1.1819 | -0.5400 | 0.5817 | -317.9754 | -305.0805 | -2.5781 | -2.6084 |
105
+ | 0.5639 | 0.89 | 3700 | 0.5921 | -0.2090 | -0.5258 | 0.7055 | 0.3168 | 1.1849 | -0.5399 | 0.5831 | -318.4354 | -305.3578 | -2.5751 | -2.6056 |
106
+ | 0.5931 | 0.91 | 3800 | 0.5930 | -0.1985 | -0.5119 | 0.7060 | 0.3134 | 1.1790 | -0.5399 | 0.5802 | -317.0424 | -304.3084 | -2.5778 | -2.6081 |
107
+ | 0.5542 | 0.93 | 3900 | 0.5929 | -0.1989 | -0.5128 | 0.7060 | 0.3139 | 1.1807 | -0.5398 | 0.5808 | -317.1321 | -304.3491 | -2.5760 | -2.6064 |
108
+ | 0.5713 | 0.96 | 4000 | 0.5926 | -0.2022 | -0.5175 | 0.7050 | 0.3153 | 1.1831 | -0.5407 | 0.5823 | -317.6028 | -304.6741 | -2.5743 | -2.6048 |
109
+ | 0.5725 | 0.98 | 4100 | 0.5925 | -0.2025 | -0.5175 | 0.7060 | 0.3149 | 1.1833 | -0.5415 | 0.5824 | -317.5993 | -304.7070 | -2.5752 | -2.6056 |
110
+
111
+
112
+ ### Framework versions
113
+
114
+ - PEFT 0.7.1
115
+ - Transformers 4.39.0.dev0
116
+ - Pytorch 2.1.2+cu121
117
+ - Datasets 2.14.6
118
+ - Tokenizers 0.15.2
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1577fedaea18596a5891baa023ddb2443d80c81b6aeea99139340cbf717ca4f4
3
  size 671150064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f166177a6f7019483154049fe98e36b1ebf2ef9046f1ceec90f963b907b13c5e
3
  size 671150064
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.6106676421631342,
4
+ "train_runtime": 67977.0297,
5
+ "train_samples": 66812,
6
+ "train_samples_per_second": 0.983,
7
+ "train_steps_per_second": 0.061
8
+ }
runs/Jul16_03-11-51_notebook-deployment-48-7d9b6c99-p5kv4/events.out.tfevents.1721100016.notebook-deployment-48-7d9b6c99-p5kv4.69914.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:60a96f76d1d5aafc8512c7bc37d1cdc245ccc6db9af8bc6b79341b73bda7459f
3
- size 404929
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbd428a1f0fbb200e7b7c976f68ab8488c77f6c2610b25c00c4c92a344a32d74
3
+ size 411443
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.6106676421631342,
4
+ "train_runtime": 67977.0297,
5
+ "train_samples": 66812,
6
+ "train_samples_per_second": 0.983,
7
+ "train_steps_per_second": 0.061
8
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff