Minbyul commited on
Commit
8a9e02c
1 Parent(s): a6310e0

End of training

Browse files
Files changed (4) hide show
  1. README.md +17 -1
  2. all_results.json +13 -0
  3. config.json +1 -1
  4. eval_results.json +16 -0
README.md CHANGED
@@ -2,9 +2,15 @@
2
  license: apache-2.0
3
  base_model: Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1
4
  tags:
 
5
  - trl
6
  - dpo
7
  - generated_from_trainer
 
 
 
 
 
8
  model-index:
9
  - name: mistral-7b-wo-kqa_golden-iter-dpo-step2
10
  results: []
@@ -15,7 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # mistral-7b-wo-kqa_golden-iter-dpo-step2
17
 
18
- This model is a fine-tuned version of [Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1](https://huggingface.co/Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1) on an unknown dataset.
 
 
 
 
 
 
 
 
 
 
19
 
20
  ## Model description
21
 
 
2
  license: apache-2.0
3
  base_model: Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1
4
  tags:
5
+ - alignment-handbook
6
  - trl
7
  - dpo
8
  - generated_from_trainer
9
+ - trl
10
+ - dpo
11
+ - generated_from_trainer
12
+ datasets:
13
+ - HuggingFaceH4/ultrafeedback_binarized
14
  model-index:
15
  - name: mistral-7b-wo-kqa_golden-iter-dpo-step2
16
  results: []
 
21
 
22
  # mistral-7b-wo-kqa_golden-iter-dpo-step2
23
 
24
+ This model is a fine-tuned version of [Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1](https://huggingface.co/Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1) on the HuggingFaceH4/ultrafeedback_binarized dataset.
25
+ It achieves the following results on the evaluation set:
26
+ - Loss: 0.5463
27
+ - Rewards/chosen: -0.5234
28
+ - Rewards/rejected: -1.3980
29
+ - Rewards/accuracies: 0.8125
30
+ - Rewards/margins: 0.8746
31
+ - Logps/rejected: -518.7740
32
+ - Logps/chosen: -233.3091
33
+ - Logits/rejected: -3.7466
34
+ - Logits/chosen: -3.9181
35
 
36
  ## Model description
37
 
all_results.json CHANGED
@@ -1,5 +1,18 @@
1
  {
2
  "epoch": 0.99,
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  "train_loss": 0.49557680382833375,
4
  "train_runtime": 1348.5444,
5
  "train_samples": 5827,
 
1
  {
2
  "epoch": 0.99,
3
+ "eval_logits/chosen": -3.918135404586792,
4
+ "eval_logits/rejected": -3.7465851306915283,
5
+ "eval_logps/chosen": -233.3091278076172,
6
+ "eval_logps/rejected": -518.7740478515625,
7
+ "eval_loss": 0.5462985038757324,
8
+ "eval_rewards/accuracies": 0.8125,
9
+ "eval_rewards/chosen": -0.5233834981918335,
10
+ "eval_rewards/margins": 0.8745772242546082,
11
+ "eval_rewards/rejected": -1.3979607820510864,
12
+ "eval_runtime": 37.2528,
13
+ "eval_samples": 290,
14
+ "eval_samples_per_second": 7.785,
15
+ "eval_steps_per_second": 0.268,
16
  "train_loss": 0.49557680382833375,
17
  "train_runtime": 1348.5444,
18
  "train_samples": 5827,
config.json CHANGED
@@ -21,6 +21,6 @@
21
  "tie_word_embeddings": false,
22
  "torch_dtype": "bfloat16",
23
  "transformers_version": "4.39.0.dev0",
24
- "use_cache": false,
25
  "vocab_size": 32000
26
  }
 
21
  "tie_word_embeddings": false,
22
  "torch_dtype": "bfloat16",
23
  "transformers_version": "4.39.0.dev0",
24
+ "use_cache": true,
25
  "vocab_size": 32000
26
  }
eval_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.99,
3
+ "eval_logits/chosen": -3.918135404586792,
4
+ "eval_logits/rejected": -3.7465851306915283,
5
+ "eval_logps/chosen": -233.3091278076172,
6
+ "eval_logps/rejected": -518.7740478515625,
7
+ "eval_loss": 0.5462985038757324,
8
+ "eval_rewards/accuracies": 0.8125,
9
+ "eval_rewards/chosen": -0.5233834981918335,
10
+ "eval_rewards/margins": 0.8745772242546082,
11
+ "eval_rewards/rejected": -1.3979607820510864,
12
+ "eval_runtime": 37.2528,
13
+ "eval_samples": 290,
14
+ "eval_samples_per_second": 7.785,
15
+ "eval_steps_per_second": 0.268
16
+ }