lewtun HF staff commited on
Commit
eb6baf0
1 Parent(s): 87a8563

End of training

Browse files
README.md CHANGED
@@ -2,9 +2,15 @@
2
  license: other
3
  base_model: lewtun/zephyr-7b-gemma-sft
4
  tags:
 
5
  - trl
6
  - dpo
7
  - generated_from_trainer
 
 
 
 
 
8
  model-index:
9
  - name: zephyr-7b-gemma-dpo
10
  results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # zephyr-7b-gemma-dpo
17
 
18
- This model is a fine-tuned version of [lewtun/zephyr-7b-gemma-sft](https://huggingface.co/lewtun/zephyr-7b-gemma-sft) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.4736
21
- - Rewards/chosen: -3.4575
22
- - Rewards/rejected: -4.9556
23
- - Rewards/accuracies: 0.75
24
- - Rewards/margins: 1.4980
25
- - Logps/rejected: -459.1662
26
- - Logps/chosen: -431.5707
27
- - Logits/rejected: 86.3863
28
- - Logits/chosen: 92.7360
29
 
30
  ## Model description
31
 
 
2
  license: other
3
  base_model: lewtun/zephyr-7b-gemma-sft
4
  tags:
5
+ - alignment-handbook
6
  - trl
7
  - dpo
8
  - generated_from_trainer
9
+ - trl
10
+ - dpo
11
+ - generated_from_trainer
12
+ datasets:
13
+ - argilla/dpo-mix-7k
14
  model-index:
15
  - name: zephyr-7b-gemma-dpo
16
  results: []
 
21
 
22
  # zephyr-7b-gemma-dpo
23
 
24
+ This model is a fine-tuned version of [lewtun/zephyr-7b-gemma-sft](https://huggingface.co/lewtun/zephyr-7b-gemma-sft) on the argilla/dpo-mix-7k dataset.
25
  It achieves the following results on the evaluation set:
26
+ - Loss: 0.4695
27
+ - Rewards/chosen: -3.3746
28
+ - Rewards/rejected: -4.9715
29
+ - Rewards/accuracies: 0.7188
30
+ - Rewards/margins: 1.5970
31
+ - Logps/rejected: -459.4853
32
+ - Logps/chosen: -429.9115
33
+ - Logits/rejected: 86.4684
34
+ - Logits/chosen: 92.8200
35
 
36
  ## Model description
37
 
all_results.json CHANGED
@@ -1,5 +1,18 @@
1
  {
2
  "epoch": 1.97,
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  "train_loss": 0.38887147261546207,
4
  "train_runtime": 1183.8142,
5
  "train_samples": 6750,
 
1
  {
2
  "epoch": 1.97,
3
+ "eval_logits/chosen": 92.81997680664062,
4
+ "eval_logits/rejected": 86.46841430664062,
5
+ "eval_logps/chosen": -429.9114685058594,
6
+ "eval_logps/rejected": -459.4852600097656,
7
+ "eval_loss": 0.4695254862308502,
8
+ "eval_rewards/accuracies": 0.71875,
9
+ "eval_rewards/chosen": -3.3745555877685547,
10
+ "eval_rewards/margins": 1.5969535112380981,
11
+ "eval_rewards/rejected": -4.9715094566345215,
12
+ "eval_runtime": 52.4051,
13
+ "eval_samples": 750,
14
+ "eval_samples_per_second": 14.312,
15
+ "eval_steps_per_second": 0.458,
16
  "train_loss": 0.38887147261546207,
17
  "train_runtime": 1183.8142,
18
  "train_samples": 6750,
config.json CHANGED
@@ -23,6 +23,6 @@
23
  "rope_theta": 10000.0,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.39.0.dev0",
26
- "use_cache": false,
27
  "vocab_size": 256000
28
  }
 
23
  "rope_theta": 10000.0,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.39.0.dev0",
26
+ "use_cache": true,
27
  "vocab_size": 256000
28
  }
eval_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.97,
3
+ "eval_logits/chosen": 92.81997680664062,
4
+ "eval_logits/rejected": 86.46841430664062,
5
+ "eval_logps/chosen": -429.9114685058594,
6
+ "eval_logps/rejected": -459.4852600097656,
7
+ "eval_loss": 0.4695254862308502,
8
+ "eval_rewards/accuracies": 0.71875,
9
+ "eval_rewards/chosen": -3.3745555877685547,
10
+ "eval_rewards/margins": 1.5969535112380981,
11
+ "eval_rewards/rejected": -4.9715094566345215,
12
+ "eval_runtime": 52.4051,
13
+ "eval_samples": 750,
14
+ "eval_samples_per_second": 14.312,
15
+ "eval_steps_per_second": 0.458
16
+ }
runs/Mar01_11-14-37_ip-26-0-170-132/events.out.tfevents.1709293097.ip-26-0-170-132.2208298.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:056a7f6db597040a25a4449e5cc73eb025c0f7c3a5b436f829fbf4df0000983a
3
+ size 815