Minbyul commited on
Commit
8af1b21
1 Parent(s): 501e263

End of training

Browse files
Files changed (4) hide show
  1. README.md +7 -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: llama2
3
  base_model: epfl-llm/meditron-7b
4
  tags:
 
5
  - trl
6
  - dpo
7
  - generated_from_trainer
 
 
 
 
 
8
  model-index:
9
  - name: meditron-7b-dpo-full-wo-kqa_golden-ep3
10
  results: []
@@ -15,7 +21,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # meditron-7b-dpo-full-wo-kqa_golden-ep3
17
 
18
- This model is a fine-tuned version of [epfl-llm/meditron-7b](https://huggingface.co/epfl-llm/meditron-7b) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
  - Loss: 0.4459
21
  - Rewards/chosen: -0.4566
 
2
  license: llama2
3
  base_model: epfl-llm/meditron-7b
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: meditron-7b-dpo-full-wo-kqa_golden-ep3
16
  results: []
 
21
 
22
  # meditron-7b-dpo-full-wo-kqa_golden-ep3
23
 
24
+ This model is a fine-tuned version of [epfl-llm/meditron-7b](https://huggingface.co/epfl-llm/meditron-7b) on the HuggingFaceH4/ultrafeedback_binarized dataset.
25
  It achieves the following results on the evaluation set:
26
  - Loss: 0.4459
27
  - Rewards/chosen: -0.4566
all_results.json CHANGED
@@ -1,5 +1,18 @@
1
  {
2
  "epoch": 1.0,
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  "train_loss": 0.5449820566177368,
4
  "train_runtime": 3295.9797,
5
  "train_samples": 12791,
 
1
  {
2
  "epoch": 1.0,
3
+ "eval_logits/chosen": -0.8309964537620544,
4
+ "eval_logits/rejected": -0.9202789068222046,
5
+ "eval_logps/chosen": -859.0581665039062,
6
+ "eval_logps/rejected": -1444.6895751953125,
7
+ "eval_loss": 0.4459438920021057,
8
+ "eval_rewards/accuracies": 0.8068181872367859,
9
+ "eval_rewards/chosen": -0.4565673768520355,
10
+ "eval_rewards/margins": 0.9446592926979065,
11
+ "eval_rewards/rejected": -1.4012266397476196,
12
+ "eval_runtime": 80.5227,
13
+ "eval_samples": 676,
14
+ "eval_samples_per_second": 8.395,
15
+ "eval_steps_per_second": 0.273,
16
  "train_loss": 0.5449820566177368,
17
  "train_runtime": 3295.9797,
18
  "train_samples": 12791,
config.json CHANGED
@@ -23,6 +23,6 @@
23
  "tie_word_embeddings": false,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.39.0.dev0",
26
- "use_cache": false,
27
  "vocab_size": 32017
28
  }
 
23
  "tie_word_embeddings": false,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.39.0.dev0",
26
+ "use_cache": true,
27
  "vocab_size": 32017
28
  }
eval_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_logits/chosen": -0.8309964537620544,
4
+ "eval_logits/rejected": -0.9202789068222046,
5
+ "eval_logps/chosen": -859.0581665039062,
6
+ "eval_logps/rejected": -1444.6895751953125,
7
+ "eval_loss": 0.4459438920021057,
8
+ "eval_rewards/accuracies": 0.8068181872367859,
9
+ "eval_rewards/chosen": -0.4565673768520355,
10
+ "eval_rewards/margins": 0.9446592926979065,
11
+ "eval_rewards/rejected": -1.4012266397476196,
12
+ "eval_runtime": 80.5227,
13
+ "eval_samples": 676,
14
+ "eval_samples_per_second": 8.395,
15
+ "eval_steps_per_second": 0.273
16
+ }