hugodk-sch commited on
Commit
3231889
1 Parent(s): 7594385

End of training

Browse files
Files changed (3) hide show
  1. README.md +13 -10
  2. all_results.json +13 -0
  3. eval_results.json +10 -10
README.md CHANGED
@@ -1,10 +1,13 @@
1
  ---
2
  library_name: peft
3
  tags:
 
4
  - trl
5
  - dpo
6
  - generated_from_trainer
7
  base_model: NbAiLab/nb-gpt-j-6B-v2
 
 
8
  model-index:
9
  - name: aftonposten-6b-align-scan
10
  results: []
@@ -15,17 +18,17 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # aftonposten-6b-align-scan
17
 
18
- This model is a fine-tuned version of [NbAiLab/nb-gpt-j-6B-v2](https://huggingface.co/NbAiLab/nb-gpt-j-6B-v2) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.9998
21
- - Rewards/chosen: 0.0135
22
- - Rewards/rejected: 0.0132
23
- - Rewards/accuracies: 0.4904
24
- - Rewards/margins: 0.0004
25
- - Logps/rejected: -37.5002
26
- - Logps/chosen: -34.0177
27
- - Logits/rejected: -2.2386
28
- - Logits/chosen: -2.2435
29
 
30
  ## Model description
31
 
 
1
  ---
2
  library_name: peft
3
  tags:
4
+ - alignment-handbook
5
  - trl
6
  - dpo
7
  - generated_from_trainer
8
  base_model: NbAiLab/nb-gpt-j-6B-v2
9
+ datasets:
10
+ - hugodk-sch/aftonposten_title_prefs
11
  model-index:
12
  - name: aftonposten-6b-align-scan
13
  results: []
 
18
 
19
  # aftonposten-6b-align-scan
20
 
21
+ This model is a fine-tuned version of [data/ap-gpt-j-6b-sft-qlora-04-08](https://huggingface.co/data/ap-gpt-j-6b-sft-qlora-04-08) on the hugodk-sch/aftonposten_title_prefs dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.9960
24
+ - Rewards/chosen: 0.0097
25
+ - Rewards/rejected: 0.0053
26
+ - Rewards/accuracies: 0.5141
27
+ - Rewards/margins: 0.0044
28
+ - Logps/rejected: -37.5100
29
+ - Logps/chosen: -34.0224
30
+ - Logits/rejected: -2.2385
31
+ - Logits/chosen: -2.2434
32
 
33
  ## Model description
34
 
all_results.json CHANGED
@@ -1,5 +1,18 @@
1
  {
2
  "epoch": 1.0,
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  "train_loss": 0.9929689927534624,
4
  "train_runtime": 3255.1859,
5
  "train_samples": 3079,
 
1
  {
2
  "epoch": 1.0,
3
+ "eval_logits/chosen": -2.2434139251708984,
4
+ "eval_logits/rejected": -2.2385354042053223,
5
+ "eval_logps/chosen": -34.0224494934082,
6
+ "eval_logps/rejected": -37.51002883911133,
7
+ "eval_loss": 0.9960440397262573,
8
+ "eval_rewards/accuracies": 0.5141196250915527,
9
+ "eval_rewards/chosen": 0.0096820630133152,
10
+ "eval_rewards/margins": 0.004409347660839558,
11
+ "eval_rewards/rejected": 0.005272718146443367,
12
+ "eval_runtime": 145.6562,
13
+ "eval_samples": 343,
14
+ "eval_samples_per_second": 2.355,
15
+ "eval_steps_per_second": 0.295,
16
  "train_loss": 0.9929689927534624,
17
  "train_runtime": 3255.1859,
18
  "train_samples": 3079,
eval_results.json CHANGED
@@ -1,15 +1,15 @@
1
  {
2
  "epoch": 1.0,
3
- "eval_logits/chosen": -2.243932008743286,
4
- "eval_logits/rejected": -2.239053249359131,
5
- "eval_logps/chosen": -34.03658676147461,
6
- "eval_logps/rejected": -37.49867630004883,
7
- "eval_loss": 1.0139278173446655,
8
- "eval_rewards/accuracies": 0.4663621187210083,
9
- "eval_rewards/chosen": -0.0014253630070015788,
10
- "eval_rewards/margins": -0.013985722325742245,
11
- "eval_rewards/rejected": 0.01256035827100277,
12
- "eval_runtime": 145.6648,
13
  "eval_samples": 343,
14
  "eval_samples_per_second": 2.355,
15
  "eval_steps_per_second": 0.295
 
1
  {
2
  "epoch": 1.0,
3
+ "eval_logits/chosen": -2.2434139251708984,
4
+ "eval_logits/rejected": -2.2385354042053223,
5
+ "eval_logps/chosen": -34.0224494934082,
6
+ "eval_logps/rejected": -37.51002883911133,
7
+ "eval_loss": 0.9960440397262573,
8
+ "eval_rewards/accuracies": 0.5141196250915527,
9
+ "eval_rewards/chosen": 0.0096820630133152,
10
+ "eval_rewards/margins": 0.004409347660839558,
11
+ "eval_rewards/rejected": 0.005272718146443367,
12
+ "eval_runtime": 145.6562,
13
  "eval_samples": 343,
14
  "eval_samples_per_second": 2.355,
15
  "eval_steps_per_second": 0.295