Hugging Face's logo

wzhouad
/
zephyr-7b-dpo-full

Text Generation
Transformers
Safetensors
llama
trl
dpo
Generated from Trainer
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
wzhouad commited on
Commit
a6f0bf6
•
1 Parent(s): 8e106d1

Model save

Browse files
Files changed (8) hide show
  1. README.md +28 -32
  2. all_results.json +4 -4
  3. model-00001-of-00003.safetensors +1 -1
  4. model-00002-of-00003.safetensors +1 -1
  5. model-00003-of-00003.safetensors +1 -1
  6. train_results.json +4 -4
  7. trainer_state.json +0 -0
  8. training_args.bin +1 -1
README.md CHANGED
@@ -2,34 +2,30 @@
2
  license: mit
3
  base_model: HuggingFaceH4/mistral-7b-sft-beta
4
  tags:
5
- - alignment-handbook
 
6
  - generated_from_trainer
7
- datasets:
8
- - HuggingFaceH4/hh-rlhf-h4
9
  model-index:
10
- - name: baseline2
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
- # baseline2
18
 
19
- This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the HuggingFaceH4/hh-rlhf-h4 dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.0260
22
- - Rewards/chosen: -2.1934
23
- - Rewards/rejected: -2.7139
24
- - Rewards/accuracies: 0.7034
25
- - Rewards/margins: 0.5205
26
- - Logps/rejected: -422.5251
27
- - Logps/chosen: -363.9185
28
- - Logits/rejected: -1.3643
29
- - Logits/chosen: -1.4823
30
- - Debug/policy Weights: 0.0429
31
- - Debug/losses: 0.0248
32
- - Debug/raw Losses: 0.5706
33
 
34
  ## Model description
35
 
@@ -64,20 +60,20 @@ The following hyperparameters were used during training:
64
 
65
  ### Training results
66
 
67
- | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen | Debug/policy Weights | Debug/losses | Debug/raw Losses |
68
- |:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:--------------------:|:------------:|:----------------:|
69
- | 0.1732 | 0.08 | 100 | 0.1665 | -0.1359 | -0.1700 | 0.6082 | 0.0341 | -168.1360 | -158.1616 | -2.7242 | -2.7324 | 0.2430 | 0.1649 | 0.6787 |
70
- | 0.0483 | 0.16 | 200 | 0.0413 | -1.3001 | -1.4580 | 0.6054 | 0.1579 | -296.9352 | -274.5804 | -2.5074 | -2.5178 | 0.0625 | 0.0392 | 0.6508 |
71
- | 0.036 | 0.24 | 300 | 0.0367 | -1.5860 | -1.8750 | 0.6465 | 0.2890 | -338.6323 | -303.1715 | -2.3859 | -2.3966 | 0.0580 | 0.0355 | 0.6226 |
72
- | 0.0407 | 0.32 | 400 | 0.0425 | -1.5769 | -1.9682 | 0.6586 | 0.3914 | -347.9584 | -302.2592 | -2.3752 | -2.3958 | 0.0683 | 0.0413 | 0.6047 |
73
- | 0.033 | 0.4 | 500 | 0.0298 | -1.9640 | -2.3686 | 0.6632 | 0.4046 | -387.9910 | -340.9718 | -2.3550 | -2.3756 | 0.0479 | 0.0283 | 0.6049 |
74
- | 0.0236 | 0.48 | 600 | 0.0252 | -2.1962 | -2.6613 | 0.6716 | 0.4651 | -417.2640 | -364.1955 | -2.0262 | -2.0674 | 0.0404 | 0.0239 | 0.5903 |
75
- | 0.0278 | 0.56 | 700 | 0.0294 | -2.0777 | -2.5844 | 0.6828 | 0.5067 | -409.5729 | -352.3430 | -1.7539 | -1.8296 | 0.0478 | 0.0280 | 0.5799 |
76
- | 0.0196 | 0.64 | 800 | 0.0217 | -2.4830 | -2.9611 | 0.6791 | 0.4781 | -447.2491 | -392.8779 | -1.3359 | -1.4513 | 0.0350 | 0.0207 | 0.5850 |
77
- | 0.0277 | 0.72 | 900 | 0.0257 | -2.2966 | -2.8007 | 0.6884 | 0.5041 | -431.2014 | -374.2322 | -1.3720 | -1.4823 | 0.0419 | 0.0244 | 0.5759 |
78
- | 0.0252 | 0.8 | 1000 | 0.0267 | -2.1757 | -2.6982 | 0.6996 | 0.5225 | -420.9574 | -362.1451 | -1.4056 | -1.5164 | 0.0440 | 0.0255 | 0.5724 |
79
- | 0.0285 | 0.88 | 1100 | 0.0267 | -2.1709 | -2.6957 | 0.6996 | 0.5248 | -420.7030 | -361.6649 | -1.3619 | -1.4794 | 0.0440 | 0.0254 | 0.5707 |
80
- | 0.0253 | 0.96 | 1200 | 0.0260 | -2.1934 | -2.7139 | 0.7034 | 0.5205 | -422.5251 | -363.9185 | -1.3643 | -1.4823 | 0.0429 | 0.0248 | 0.5706 |
81
 
82
 
83
  ### Framework versions
 
2
  license: mit
3
  base_model: HuggingFaceH4/mistral-7b-sft-beta
4
  tags:
5
+ - trl
6
+ - dpo
7
  - generated_from_trainer
 
 
8
  model-index:
9
+ - name: zephyr-7b-dpo-full
10
  results: []
11
  ---
12
 
13
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
  should probably proofread and complete it, then remove this comment. -->
15
 
16
+ # zephyr-7b-dpo-full
17
 
18
+ This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.5440
21
+ - Rewards/chosen: -2.2940
22
+ - Rewards/rejected: -3.0054
23
+ - Rewards/accuracies: 0.7090
24
+ - Rewards/margins: 0.7114
25
+ - Logps/rejected: -451.6765
26
+ - Logps/chosen: -373.9785
27
+ - Logits/rejected: 0.3244
28
+ - Logits/chosen: 0.0742
 
 
 
29
 
30
  ## Model description
31
 
 
60
 
61
  ### Training results
62
 
63
+ | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
64
+ |:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
65
+ | 0.6789 | 0.08 | 100 | 0.6770 | -0.1062 | -0.1422 | 0.5914 | 0.0360 | -165.3552 | -155.1927 | -2.7255 | -2.7337 |
66
+ | 0.6062 | 0.16 | 200 | 0.6079 | -1.0212 | -1.3873 | 0.6670 | 0.3660 | -289.8622 | -246.6971 | -2.3696 | -2.3856 |
67
+ | 0.5965 | 0.24 | 300 | 0.5907 | -1.3779 | -1.8008 | 0.6623 | 0.4229 | -331.2100 | -282.3621 | -2.2450 | -2.2656 |
68
+ | 0.5729 | 0.32 | 400 | 0.5711 | -1.6763 | -2.2404 | 0.6828 | 0.5640 | -375.1720 | -312.2064 | -1.2920 | -1.3760 |
69
+ | 0.5645 | 0.4 | 500 | 0.5639 | -2.0721 | -2.6869 | 0.6987 | 0.6147 | -419.8194 | -351.7883 | -0.6091 | -0.7860 |
70
+ | 0.5513 | 0.48 | 600 | 0.5582 | -2.9237 | -3.5389 | 0.7108 | 0.6152 | -505.0223 | -436.9386 | 0.1224 | -0.1054 |
71
+ | 0.5571 | 0.56 | 700 | 0.5559 | -2.7971 | -3.5456 | 0.7043 | 0.7485 | -505.6961 | -424.2823 | 0.2980 | 0.0356 |
72
+ | 0.5609 | 0.64 | 800 | 0.5469 | -2.4314 | -3.0831 | 0.7108 | 0.6517 | -459.4439 | -387.7092 | 0.1922 | -0.0312 |
73
+ | 0.5514 | 0.72 | 900 | 0.5474 | -2.4774 | -3.2082 | 0.6996 | 0.7308 | -471.9533 | -392.3096 | 0.5382 | 0.2860 |
74
+ | 0.527 | 0.8 | 1000 | 0.5454 | -2.5040 | -3.2071 | 0.7080 | 0.7031 | -471.8454 | -394.9711 | 0.6372 | 0.3871 |
75
+ | 0.5487 | 0.88 | 1100 | 0.5444 | -2.2851 | -2.9963 | 0.7090 | 0.7112 | -450.7599 | -373.0831 | 0.4336 | 0.1858 |
76
+ | 0.5483 | 0.96 | 1200 | 0.5440 | -2.2940 | -3.0054 | 0.7090 | 0.7114 | -451.6765 | -373.9785 | 0.3244 | 0.0742 |
77
 
78
 
79
  ### Framework versions
all_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 1.0,
3
- "train_loss": 0.049385896711877195,
4
- "train_runtime": 11611.5481,
5
  "train_samples": 160800,
6
- "train_samples_per_second": 13.848,
7
- "train_steps_per_second": 0.108
8
  }
 
1
  {
2
  "epoch": 1.0,
3
+ "train_loss": 0.5712926928784438,
4
+ "train_runtime": 11451.0976,
5
  "train_samples": 160800,
6
+ "train_samples_per_second": 14.042,
7
+ "train_steps_per_second": 0.11
8
  }
model-00001-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d91c6951684710ee13cfc6493657d73e5e07ec420fb9e3f03057df8a2410eee4
3
  size 4943162336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18e1cec63bd40f863dc594533ae9ac02d7bcdd4f57a17c1ef5d63193122a0814
3
  size 4943162336
model-00002-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:45a598088007ec1a9a7caaca9462759689bb90fd469cdf968e25a0b50190ed6f
3
  size 4999819336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:565d4244afeda54e7f62be9e162a16c6892085c081422f02c7a001ecce587eb6
3
  size 4999819336
model-00003-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b16b3c8e88783d9d90ab30f2ce3f323612375d8cd3c2436ffa74a856527e56c5
3
  size 4540516344
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0debf1533b3a9f2ffea91ddec7f947ba3d1c43476aedcef3273235a227bb4ce5
3
  size 4540516344
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 1.0,
3
- "train_loss": 0.049385896711877195,
4
- "train_runtime": 11611.5481,
5
  "train_samples": 160800,
6
- "train_samples_per_second": 13.848,
7
- "train_steps_per_second": 0.108
8
  }
 
1
  {
2
  "epoch": 1.0,
3
+ "train_loss": 0.5712926928784438,
4
+ "train_runtime": 11451.0976,
5
  "train_samples": 160800,
6
+ "train_samples_per_second": 14.042,
7
+ "train_steps_per_second": 0.11
8
  }
trainer_state.json CHANGED
The diff for this file is too large to render. See raw diff
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7b563d8ba7853a3ff70d20b889cac2dc1b6f510a9e1fcf017bbae2c1ee40b07d
3
  size 5944
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef3f3bcb1d637ffd73632ad00af47d3006ac1e6c1f0c109c90bd802bdaba6dcd
3
  size 5944