lole25 commited on
Commit
134c76e
1 Parent(s): 77077a0

Model save

Browse files
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: peft
4
+ tags:
5
+ - trl
6
+ - dpo
7
+ - generated_from_trainer
8
+ base_model: DUAL-GPO/zephyr-7b-gpo-final-i0
9
+ model-index:
10
+ - name: zephyr-7b-gpo-v9-i1
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
+ # zephyr-7b-gpo-v9-i1
18
+
19
+ This model is a fine-tuned version of [DUAL-GPO/zephyr-7b-gpo-final-i0](https://huggingface.co/DUAL-GPO/zephyr-7b-gpo-final-i0) on the None dataset.
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 5e-06
39
+ - train_batch_size: 2
40
+ - eval_batch_size: 2
41
+ - seed: 42
42
+ - distributed_type: multi-GPU
43
+ - num_devices: 2
44
+ - gradient_accumulation_steps: 2
45
+ - total_train_batch_size: 8
46
+ - total_eval_batch_size: 4
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.1
50
+ - num_epochs: 1
51
+
52
+ ### Training results
53
+
54
+
55
+
56
+ ### Framework versions
57
+
58
+ - PEFT 0.7.1
59
+ - Transformers 4.36.2
60
+ - Pytorch 2.1.2+cu121
61
+ - Datasets 2.14.6
62
+ - Tokenizers 0.15.2
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:db79912969bf7ff02cc0f582c93d1b39f3ef503a5db031d8d34bb7faac52912e
3
  size 671150064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:acc0899428eb14214004a8ef35db80ff3329e8f36e3c95ef684b07f9dd104dca
3
  size 671150064
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.13990657812754312,
4
+ "train_runtime": 16010.7596,
5
+ "train_samples": 15000,
6
+ "train_samples_per_second": 0.937,
7
+ "train_steps_per_second": 0.117
8
+ }
runs/May08_17-41-32_gpu4-119-5/events.out.tfevents.1715154200.gpu4-119-5.292195.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e2611175afe3e7c902cdb3e4e52bf48e9a8a7816762a7b191f3b4418cee52d6b
3
- size 119170
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff1920744b3861a422af2fe34a9a162dc078cd5d7d957ee44c0c7aa9f9c9a9ed
3
+ size 123962
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.13990657812754312,
4
+ "train_runtime": 16010.7596,
5
+ "train_samples": 15000,
6
+ "train_samples_per_second": 0.937,
7
+ "train_steps_per_second": 0.117
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2662 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1875,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 2.6595744680851065e-08,
14
+ "logits/chosen": -1.7968215942382812,
15
+ "logits/rejected": -2.159090995788574,
16
+ "logps/chosen": -88.33059692382812,
17
+ "logps/rejected": -242.96200561523438,
18
+ "loss": 0.4322,
19
+ "rewards/accuracies": 0.0,
20
+ "rewards/chosen": 0.0,
21
+ "rewards/margins": 0.0,
22
+ "rewards/rejected": 0.0,
23
+ "step": 1
24
+ },
25
+ {
26
+ "epoch": 0.01,
27
+ "learning_rate": 2.6595744680851066e-07,
28
+ "logits/chosen": -2.003159999847412,
29
+ "logits/rejected": -1.3869916200637817,
30
+ "logps/chosen": -240.9772186279297,
31
+ "logps/rejected": -195.60606384277344,
32
+ "loss": 0.3319,
33
+ "rewards/accuracies": 0.3333333432674408,
34
+ "rewards/chosen": -3.270954766776413e-05,
35
+ "rewards/margins": -8.25071256258525e-05,
36
+ "rewards/rejected": 4.979758523404598e-05,
37
+ "step": 10
38
+ },
39
+ {
40
+ "epoch": 0.01,
41
+ "learning_rate": 5.319148936170213e-07,
42
+ "logits/chosen": -2.0388007164001465,
43
+ "logits/rejected": -1.5615094900131226,
44
+ "logps/chosen": -291.083740234375,
45
+ "logps/rejected": -277.5216369628906,
46
+ "loss": 0.3514,
47
+ "rewards/accuracies": 0.574999988079071,
48
+ "rewards/chosen": 1.1951732631132472e-05,
49
+ "rewards/margins": 0.00027519199647940695,
50
+ "rewards/rejected": -0.0002632402756717056,
51
+ "step": 20
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "learning_rate": 7.97872340425532e-07,
56
+ "logits/chosen": -1.860889196395874,
57
+ "logits/rejected": -1.5862194299697876,
58
+ "logps/chosen": -248.38510131835938,
59
+ "logps/rejected": -261.7816467285156,
60
+ "loss": 0.324,
61
+ "rewards/accuracies": 0.800000011920929,
62
+ "rewards/chosen": -0.0007015246083028615,
63
+ "rewards/margins": 0.004821115639060736,
64
+ "rewards/rejected": -0.005522639956325293,
65
+ "step": 30
66
+ },
67
+ {
68
+ "epoch": 0.02,
69
+ "learning_rate": 1.0638297872340427e-06,
70
+ "logits/chosen": -1.8764064311981201,
71
+ "logits/rejected": -1.2899483442306519,
72
+ "logps/chosen": -355.25958251953125,
73
+ "logps/rejected": -389.2695007324219,
74
+ "loss": 0.3286,
75
+ "rewards/accuracies": 0.8500000238418579,
76
+ "rewards/chosen": -0.009422576054930687,
77
+ "rewards/margins": 0.022184943780303,
78
+ "rewards/rejected": -0.03160751983523369,
79
+ "step": 40
80
+ },
81
+ {
82
+ "epoch": 0.03,
83
+ "learning_rate": 1.3297872340425533e-06,
84
+ "logits/chosen": -1.990142583847046,
85
+ "logits/rejected": -1.2961665391921997,
86
+ "logps/chosen": -316.00860595703125,
87
+ "logps/rejected": -277.88421630859375,
88
+ "loss": 0.2629,
89
+ "rewards/accuracies": 0.699999988079071,
90
+ "rewards/chosen": -0.05700983479619026,
91
+ "rewards/margins": 0.059757936745882034,
92
+ "rewards/rejected": -0.1167677640914917,
93
+ "step": 50
94
+ },
95
+ {
96
+ "epoch": 0.03,
97
+ "learning_rate": 1.595744680851064e-06,
98
+ "logits/chosen": -1.748492956161499,
99
+ "logits/rejected": -0.8994135856628418,
100
+ "logps/chosen": -389.3627624511719,
101
+ "logps/rejected": -579.7057495117188,
102
+ "loss": 0.1989,
103
+ "rewards/accuracies": 0.875,
104
+ "rewards/chosen": -0.09011684358119965,
105
+ "rewards/margins": 0.22812744975090027,
106
+ "rewards/rejected": -0.3182442784309387,
107
+ "step": 60
108
+ },
109
+ {
110
+ "epoch": 0.04,
111
+ "learning_rate": 1.8617021276595745e-06,
112
+ "logits/chosen": -1.6900399923324585,
113
+ "logits/rejected": -1.4010140895843506,
114
+ "logps/chosen": -420.5406799316406,
115
+ "logps/rejected": -859.8084716796875,
116
+ "loss": 0.1253,
117
+ "rewards/accuracies": 0.8999999761581421,
118
+ "rewards/chosen": -0.20870384573936462,
119
+ "rewards/margins": 0.3385527431964874,
120
+ "rewards/rejected": -0.547256588935852,
121
+ "step": 70
122
+ },
123
+ {
124
+ "epoch": 0.04,
125
+ "learning_rate": 2.1276595744680853e-06,
126
+ "logits/chosen": -1.7609751224517822,
127
+ "logits/rejected": -1.0384010076522827,
128
+ "logps/chosen": -474.48187255859375,
129
+ "logps/rejected": -747.34716796875,
130
+ "loss": 0.1309,
131
+ "rewards/accuracies": 0.8999999761581421,
132
+ "rewards/chosen": -0.19330377876758575,
133
+ "rewards/margins": 0.34078216552734375,
134
+ "rewards/rejected": -0.5340859293937683,
135
+ "step": 80
136
+ },
137
+ {
138
+ "epoch": 0.05,
139
+ "learning_rate": 2.393617021276596e-06,
140
+ "logits/chosen": -1.7291476726531982,
141
+ "logits/rejected": -1.2021540403366089,
142
+ "logps/chosen": -454.2134704589844,
143
+ "logps/rejected": -764.934326171875,
144
+ "loss": 0.16,
145
+ "rewards/accuracies": 0.8500000238418579,
146
+ "rewards/chosen": -0.1365814059972763,
147
+ "rewards/margins": 0.36457785964012146,
148
+ "rewards/rejected": -0.5011593103408813,
149
+ "step": 90
150
+ },
151
+ {
152
+ "epoch": 0.05,
153
+ "learning_rate": 2.6595744680851065e-06,
154
+ "logits/chosen": -1.5737159252166748,
155
+ "logits/rejected": -0.9248941540718079,
156
+ "logps/chosen": -482.3492126464844,
157
+ "logps/rejected": -792.2481689453125,
158
+ "loss": 0.1239,
159
+ "rewards/accuracies": 0.875,
160
+ "rewards/chosen": -0.19203761219978333,
161
+ "rewards/margins": 0.3564862310886383,
162
+ "rewards/rejected": -0.5485238432884216,
163
+ "step": 100
164
+ },
165
+ {
166
+ "epoch": 0.06,
167
+ "learning_rate": 2.9255319148936174e-06,
168
+ "logits/chosen": -1.7435375452041626,
169
+ "logits/rejected": -1.356065034866333,
170
+ "logps/chosen": -416.564208984375,
171
+ "logps/rejected": -796.4661254882812,
172
+ "loss": 0.1253,
173
+ "rewards/accuracies": 0.824999988079071,
174
+ "rewards/chosen": -0.1526903361082077,
175
+ "rewards/margins": 0.3349696397781372,
176
+ "rewards/rejected": -0.4876599907875061,
177
+ "step": 110
178
+ },
179
+ {
180
+ "epoch": 0.06,
181
+ "learning_rate": 3.191489361702128e-06,
182
+ "logits/chosen": -1.6976553201675415,
183
+ "logits/rejected": -1.0894078016281128,
184
+ "logps/chosen": -409.96258544921875,
185
+ "logps/rejected": -617.7588500976562,
186
+ "loss": 0.1948,
187
+ "rewards/accuracies": 0.7749999761581421,
188
+ "rewards/chosen": -0.13733306527137756,
189
+ "rewards/margins": 0.293459415435791,
190
+ "rewards/rejected": -0.4307924807071686,
191
+ "step": 120
192
+ },
193
+ {
194
+ "epoch": 0.07,
195
+ "learning_rate": 3.457446808510639e-06,
196
+ "logits/chosen": -1.7993590831756592,
197
+ "logits/rejected": -1.400632619857788,
198
+ "logps/chosen": -370.1565856933594,
199
+ "logps/rejected": -709.3056640625,
200
+ "loss": 0.2055,
201
+ "rewards/accuracies": 0.875,
202
+ "rewards/chosen": -0.11207763850688934,
203
+ "rewards/margins": 0.340470552444458,
204
+ "rewards/rejected": -0.45254817605018616,
205
+ "step": 130
206
+ },
207
+ {
208
+ "epoch": 0.07,
209
+ "learning_rate": 3.723404255319149e-06,
210
+ "logits/chosen": -1.495011806488037,
211
+ "logits/rejected": -0.9245948791503906,
212
+ "logps/chosen": -388.5771789550781,
213
+ "logps/rejected": -792.4680786132812,
214
+ "loss": 0.1088,
215
+ "rewards/accuracies": 0.925000011920929,
216
+ "rewards/chosen": -0.1575288623571396,
217
+ "rewards/margins": 0.3997672200202942,
218
+ "rewards/rejected": -0.557296097278595,
219
+ "step": 140
220
+ },
221
+ {
222
+ "epoch": 0.08,
223
+ "learning_rate": 3.98936170212766e-06,
224
+ "logits/chosen": -1.6491578817367554,
225
+ "logits/rejected": -1.2172632217407227,
226
+ "logps/chosen": -407.8502502441406,
227
+ "logps/rejected": -738.5733642578125,
228
+ "loss": 0.1397,
229
+ "rewards/accuracies": 0.800000011920929,
230
+ "rewards/chosen": -0.12080486118793488,
231
+ "rewards/margins": 0.32797589898109436,
232
+ "rewards/rejected": -0.44878071546554565,
233
+ "step": 150
234
+ },
235
+ {
236
+ "epoch": 0.09,
237
+ "learning_rate": 4.255319148936171e-06,
238
+ "logits/chosen": -1.6612653732299805,
239
+ "logits/rejected": -1.1705405712127686,
240
+ "logps/chosen": -353.0194396972656,
241
+ "logps/rejected": -689.8749389648438,
242
+ "loss": 0.1454,
243
+ "rewards/accuracies": 0.8500000238418579,
244
+ "rewards/chosen": -0.09746531397104263,
245
+ "rewards/margins": 0.3515530228614807,
246
+ "rewards/rejected": -0.44901829957962036,
247
+ "step": 160
248
+ },
249
+ {
250
+ "epoch": 0.09,
251
+ "learning_rate": 4.521276595744681e-06,
252
+ "logits/chosen": -1.54987370967865,
253
+ "logits/rejected": -1.1912695169448853,
254
+ "logps/chosen": -544.5787963867188,
255
+ "logps/rejected": -835.3132934570312,
256
+ "loss": 0.1048,
257
+ "rewards/accuracies": 0.7749999761581421,
258
+ "rewards/chosen": -0.23248295485973358,
259
+ "rewards/margins": 0.3334501087665558,
260
+ "rewards/rejected": -0.565933108329773,
261
+ "step": 170
262
+ },
263
+ {
264
+ "epoch": 0.1,
265
+ "learning_rate": 4.787234042553192e-06,
266
+ "logits/chosen": -1.7397425174713135,
267
+ "logits/rejected": -0.8725941777229309,
268
+ "logps/chosen": -510.69842529296875,
269
+ "logps/rejected": -840.5343017578125,
270
+ "loss": 0.1531,
271
+ "rewards/accuracies": 0.875,
272
+ "rewards/chosen": -0.2171137034893036,
273
+ "rewards/margins": 0.39573976397514343,
274
+ "rewards/rejected": -0.6128535270690918,
275
+ "step": 180
276
+ },
277
+ {
278
+ "epoch": 0.1,
279
+ "learning_rate": 4.999982660399688e-06,
280
+ "logits/chosen": -1.6966726779937744,
281
+ "logits/rejected": -1.09552800655365,
282
+ "logps/chosen": -514.5984497070312,
283
+ "logps/rejected": -911.4729614257812,
284
+ "loss": 0.1503,
285
+ "rewards/accuracies": 0.8500000238418579,
286
+ "rewards/chosen": -0.2481248676776886,
287
+ "rewards/margins": 0.3679044842720032,
288
+ "rewards/rejected": -0.6160293221473694,
289
+ "step": 190
290
+ },
291
+ {
292
+ "epoch": 0.11,
293
+ "learning_rate": 4.99937579964398e-06,
294
+ "logits/chosen": -1.4942667484283447,
295
+ "logits/rejected": -1.1419141292572021,
296
+ "logps/chosen": -432.5450134277344,
297
+ "logps/rejected": -730.1014404296875,
298
+ "loss": 0.1267,
299
+ "rewards/accuracies": 0.824999988079071,
300
+ "rewards/chosen": -0.21937580406665802,
301
+ "rewards/margins": 0.3185574412345886,
302
+ "rewards/rejected": -0.5379332304000854,
303
+ "step": 200
304
+ },
305
+ {
306
+ "epoch": 0.11,
307
+ "learning_rate": 4.9979021993870645e-06,
308
+ "logits/chosen": -1.571395993232727,
309
+ "logits/rejected": -0.9183829426765442,
310
+ "logps/chosen": -480.79644775390625,
311
+ "logps/rejected": -813.7987060546875,
312
+ "loss": 0.1624,
313
+ "rewards/accuracies": 0.824999988079071,
314
+ "rewards/chosen": -0.18962515890598297,
315
+ "rewards/margins": 0.3737575113773346,
316
+ "rewards/rejected": -0.563382625579834,
317
+ "step": 210
318
+ },
319
+ {
320
+ "epoch": 0.12,
321
+ "learning_rate": 4.995562370647553e-06,
322
+ "logits/chosen": -1.668015480041504,
323
+ "logits/rejected": -1.1087052822113037,
324
+ "logps/chosen": -517.7100219726562,
325
+ "logps/rejected": -838.1522216796875,
326
+ "loss": 0.1372,
327
+ "rewards/accuracies": 0.7749999761581421,
328
+ "rewards/chosen": -0.22473697364330292,
329
+ "rewards/margins": 0.35942238569259644,
330
+ "rewards/rejected": -0.5841594338417053,
331
+ "step": 220
332
+ },
333
+ {
334
+ "epoch": 0.12,
335
+ "learning_rate": 4.992357124836838e-06,
336
+ "logits/chosen": -1.3532911539077759,
337
+ "logits/rejected": -0.6337820291519165,
338
+ "logps/chosen": -458.74462890625,
339
+ "logps/rejected": -736.6771240234375,
340
+ "loss": 0.1419,
341
+ "rewards/accuracies": 0.800000011920929,
342
+ "rewards/chosen": -0.21437951922416687,
343
+ "rewards/margins": 0.35455334186553955,
344
+ "rewards/rejected": -0.5689328908920288,
345
+ "step": 230
346
+ },
347
+ {
348
+ "epoch": 0.13,
349
+ "learning_rate": 4.9882875734777044e-06,
350
+ "logits/chosen": -1.6833770275115967,
351
+ "logits/rejected": -1.0865981578826904,
352
+ "logps/chosen": -476.49578857421875,
353
+ "logps/rejected": -742.6441650390625,
354
+ "loss": 0.162,
355
+ "rewards/accuracies": 0.8500000238418579,
356
+ "rewards/chosen": -0.1685108244419098,
357
+ "rewards/margins": 0.33409184217453003,
358
+ "rewards/rejected": -0.5026026368141174,
359
+ "step": 240
360
+ },
361
+ {
362
+ "epoch": 0.13,
363
+ "learning_rate": 4.983355127818882e-06,
364
+ "logits/chosen": -1.4850168228149414,
365
+ "logits/rejected": -0.9603360295295715,
366
+ "logps/chosen": -400.22967529296875,
367
+ "logps/rejected": -569.9345703125,
368
+ "loss": 0.1919,
369
+ "rewards/accuracies": 0.625,
370
+ "rewards/chosen": -0.20730257034301758,
371
+ "rewards/margins": 0.22820453345775604,
372
+ "rewards/rejected": -0.4355071187019348,
373
+ "step": 250
374
+ },
375
+ {
376
+ "epoch": 0.14,
377
+ "learning_rate": 4.977561498345639e-06,
378
+ "logits/chosen": -1.544639229774475,
379
+ "logits/rejected": -1.1013596057891846,
380
+ "logps/chosen": -407.4434509277344,
381
+ "logps/rejected": -822.0099487304688,
382
+ "loss": 0.0966,
383
+ "rewards/accuracies": 0.925000011920929,
384
+ "rewards/chosen": -0.19301848113536835,
385
+ "rewards/margins": 0.41144537925720215,
386
+ "rewards/rejected": -0.6044638752937317,
387
+ "step": 260
388
+ },
389
+ {
390
+ "epoch": 0.14,
391
+ "learning_rate": 4.970908694186624e-06,
392
+ "logits/chosen": -1.5448771715164185,
393
+ "logits/rejected": -0.8540661931037903,
394
+ "logps/chosen": -542.1297607421875,
395
+ "logps/rejected": -889.5344848632812,
396
+ "loss": 0.1371,
397
+ "rewards/accuracies": 0.875,
398
+ "rewards/chosen": -0.25042372941970825,
399
+ "rewards/margins": 0.39966678619384766,
400
+ "rewards/rejected": -0.6500904560089111,
401
+ "step": 270
402
+ },
403
+ {
404
+ "epoch": 0.15,
405
+ "learning_rate": 4.9633990224171305e-06,
406
+ "logits/chosen": -1.4944156408309937,
407
+ "logits/rejected": -0.8036524057388306,
408
+ "logps/chosen": -671.7164306640625,
409
+ "logps/rejected": -922.0513916015625,
410
+ "loss": 0.1319,
411
+ "rewards/accuracies": 0.8500000238418579,
412
+ "rewards/chosen": -0.3747330605983734,
413
+ "rewards/margins": 0.3100079596042633,
414
+ "rewards/rejected": -0.6847410202026367,
415
+ "step": 280
416
+ },
417
+ {
418
+ "epoch": 0.15,
419
+ "learning_rate": 4.955035087259046e-06,
420
+ "logits/chosen": -1.4477952718734741,
421
+ "logits/rejected": -0.8218593597412109,
422
+ "logps/chosen": -612.0665893554688,
423
+ "logps/rejected": -863.3322143554688,
424
+ "loss": 0.1494,
425
+ "rewards/accuracies": 0.824999988079071,
426
+ "rewards/chosen": -0.3176848888397217,
427
+ "rewards/margins": 0.30745354294776917,
428
+ "rewards/rejected": -0.6251384019851685,
429
+ "step": 290
430
+ },
431
+ {
432
+ "epoch": 0.16,
433
+ "learning_rate": 4.945819789177756e-06,
434
+ "logits/chosen": -1.611026406288147,
435
+ "logits/rejected": -1.1793110370635986,
436
+ "logps/chosen": -518.3214111328125,
437
+ "logps/rejected": -892.6036987304688,
438
+ "loss": 0.1228,
439
+ "rewards/accuracies": 0.875,
440
+ "rewards/chosen": -0.2295423001050949,
441
+ "rewards/margins": 0.37014490365982056,
442
+ "rewards/rejected": -0.5996872186660767,
443
+ "step": 300
444
+ },
445
+ {
446
+ "epoch": 0.17,
447
+ "learning_rate": 4.935756323876306e-06,
448
+ "logits/chosen": -1.508418083190918,
449
+ "logits/rejected": -1.277306318283081,
450
+ "logps/chosen": -459.0326232910156,
451
+ "logps/rejected": -823.93017578125,
452
+ "loss": 0.1518,
453
+ "rewards/accuracies": 0.699999988079071,
454
+ "rewards/chosen": -0.27009934186935425,
455
+ "rewards/margins": 0.32507914304733276,
456
+ "rewards/rejected": -0.5951785445213318,
457
+ "step": 310
458
+ },
459
+ {
460
+ "epoch": 0.17,
461
+ "learning_rate": 4.924848181187199e-06,
462
+ "logits/chosen": -1.6442441940307617,
463
+ "logits/rejected": -1.1329659223556519,
464
+ "logps/chosen": -494.06097412109375,
465
+ "logps/rejected": -854.1019287109375,
466
+ "loss": 0.1422,
467
+ "rewards/accuracies": 0.824999988079071,
468
+ "rewards/chosen": -0.20919394493103027,
469
+ "rewards/margins": 0.39555859565734863,
470
+ "rewards/rejected": -0.6047526001930237,
471
+ "step": 320
472
+ },
473
+ {
474
+ "epoch": 0.18,
475
+ "learning_rate": 4.913099143862173e-06,
476
+ "logits/chosen": -1.3633651733398438,
477
+ "logits/rejected": -0.9123932123184204,
478
+ "logps/chosen": -474.5048828125,
479
+ "logps/rejected": -829.7545776367188,
480
+ "loss": 0.1498,
481
+ "rewards/accuracies": 0.8500000238418579,
482
+ "rewards/chosen": -0.28733277320861816,
483
+ "rewards/margins": 0.3500373959541321,
484
+ "rewards/rejected": -0.6373701095581055,
485
+ "step": 330
486
+ },
487
+ {
488
+ "epoch": 0.18,
489
+ "learning_rate": 4.900513286260416e-06,
490
+ "logits/chosen": -1.5975598096847534,
491
+ "logits/rejected": -1.2887117862701416,
492
+ "logps/chosen": -400.32781982421875,
493
+ "logps/rejected": -744.0382080078125,
494
+ "loss": 0.1053,
495
+ "rewards/accuracies": 0.800000011920929,
496
+ "rewards/chosen": -0.20992426574230194,
497
+ "rewards/margins": 0.3338248133659363,
498
+ "rewards/rejected": -0.5437491536140442,
499
+ "step": 340
500
+ },
501
+ {
502
+ "epoch": 0.19,
503
+ "learning_rate": 4.887094972935645e-06,
504
+ "logits/chosen": -1.764219045639038,
505
+ "logits/rejected": -0.9871004819869995,
506
+ "logps/chosen": -573.0086059570312,
507
+ "logps/rejected": -937.3956909179688,
508
+ "loss": 0.1504,
509
+ "rewards/accuracies": 0.875,
510
+ "rewards/chosen": -0.2734777331352234,
511
+ "rewards/margins": 0.37265342473983765,
512
+ "rewards/rejected": -0.646131157875061,
513
+ "step": 350
514
+ },
515
+ {
516
+ "epoch": 0.19,
517
+ "learning_rate": 4.87284885712256e-06,
518
+ "logits/chosen": -1.5166432857513428,
519
+ "logits/rejected": -0.8717886209487915,
520
+ "logps/chosen": -572.190673828125,
521
+ "logps/rejected": -876.5632934570312,
522
+ "loss": 0.1876,
523
+ "rewards/accuracies": 0.824999988079071,
524
+ "rewards/chosen": -0.32470396161079407,
525
+ "rewards/margins": 0.34118732810020447,
526
+ "rewards/rejected": -0.6658912897109985,
527
+ "step": 360
528
+ },
529
+ {
530
+ "epoch": 0.2,
531
+ "learning_rate": 4.857779879123181e-06,
532
+ "logits/chosen": -1.7403156757354736,
533
+ "logits/rejected": -0.9518265724182129,
534
+ "logps/chosen": -505.53387451171875,
535
+ "logps/rejected": -778.5391845703125,
536
+ "loss": 0.1207,
537
+ "rewards/accuracies": 0.7749999761581421,
538
+ "rewards/chosen": -0.20143766701221466,
539
+ "rewards/margins": 0.37924817204475403,
540
+ "rewards/rejected": -0.5806857943534851,
541
+ "step": 370
542
+ },
543
+ {
544
+ "epoch": 0.2,
545
+ "learning_rate": 4.841893264593643e-06,
546
+ "logits/chosen": -1.7110675573349,
547
+ "logits/rejected": -1.026960849761963,
548
+ "logps/chosen": -472.236328125,
549
+ "logps/rejected": -761.3172607421875,
550
+ "loss": 0.095,
551
+ "rewards/accuracies": 0.875,
552
+ "rewards/chosen": -0.2130139172077179,
553
+ "rewards/margins": 0.36098140478134155,
554
+ "rewards/rejected": -0.5739952325820923,
555
+ "step": 380
556
+ },
557
+ {
558
+ "epoch": 0.21,
559
+ "learning_rate": 4.825194522732023e-06,
560
+ "logits/chosen": -1.6140925884246826,
561
+ "logits/rejected": -1.1293842792510986,
562
+ "logps/chosen": -507.7935485839844,
563
+ "logps/rejected": -913.9110107421875,
564
+ "loss": 0.1277,
565
+ "rewards/accuracies": 0.8999999761581421,
566
+ "rewards/chosen": -0.26502326130867004,
567
+ "rewards/margins": 0.3756178915500641,
568
+ "rewards/rejected": -0.6406410932540894,
569
+ "step": 390
570
+ },
571
+ {
572
+ "epoch": 0.21,
573
+ "learning_rate": 4.807689444367853e-06,
574
+ "logits/chosen": -1.7682578563690186,
575
+ "logits/rejected": -1.3489004373550415,
576
+ "logps/chosen": -495.90869140625,
577
+ "logps/rejected": -782.658447265625,
578
+ "loss": 0.1251,
579
+ "rewards/accuracies": 0.7250000238418579,
580
+ "rewards/chosen": -0.260581910610199,
581
+ "rewards/margins": 0.3089093565940857,
582
+ "rewards/rejected": -0.5694912075996399,
583
+ "step": 400
584
+ },
585
+ {
586
+ "epoch": 0.22,
587
+ "learning_rate": 4.78938409995396e-06,
588
+ "logits/chosen": -1.5254316329956055,
589
+ "logits/rejected": -1.0945005416870117,
590
+ "logps/chosen": -463.646484375,
591
+ "logps/rejected": -902.4519653320312,
592
+ "loss": 0.1313,
593
+ "rewards/accuracies": 0.8500000238418579,
594
+ "rewards/chosen": -0.24164071679115295,
595
+ "rewards/margins": 0.3996545672416687,
596
+ "rewards/rejected": -0.6412952542304993,
597
+ "step": 410
598
+ },
599
+ {
600
+ "epoch": 0.22,
601
+ "learning_rate": 4.770284837461342e-06,
602
+ "logits/chosen": -1.5965580940246582,
603
+ "logits/rejected": -0.836743175983429,
604
+ "logps/chosen": -586.2058715820312,
605
+ "logps/rejected": -917.9168090820312,
606
+ "loss": 0.1124,
607
+ "rewards/accuracies": 0.925000011920929,
608
+ "rewards/chosen": -0.26902061700820923,
609
+ "rewards/margins": 0.3916351795196533,
610
+ "rewards/rejected": -0.6606558561325073,
611
+ "step": 420
612
+ },
613
+ {
614
+ "epoch": 0.23,
615
+ "learning_rate": 4.7503982801778015e-06,
616
+ "logits/chosen": -1.557950735092163,
617
+ "logits/rejected": -1.0112215280532837,
618
+ "logps/chosen": -501.4098205566406,
619
+ "logps/rejected": -789.2760009765625,
620
+ "loss": 0.154,
621
+ "rewards/accuracies": 0.800000011920929,
622
+ "rewards/chosen": -0.24787664413452148,
623
+ "rewards/margins": 0.3141789734363556,
624
+ "rewards/rejected": -0.5620556473731995,
625
+ "step": 430
626
+ },
627
+ {
628
+ "epoch": 0.23,
629
+ "learning_rate": 4.729731324411104e-06,
630
+ "logits/chosen": -1.7676448822021484,
631
+ "logits/rejected": -1.1603769063949585,
632
+ "logps/chosen": -429.96734619140625,
633
+ "logps/rejected": -750.7506103515625,
634
+ "loss": 0.1237,
635
+ "rewards/accuracies": 0.800000011920929,
636
+ "rewards/chosen": -0.1744639277458191,
637
+ "rewards/margins": 0.35027581453323364,
638
+ "rewards/rejected": -0.5247397422790527,
639
+ "step": 440
640
+ },
641
+ {
642
+ "epoch": 0.24,
643
+ "learning_rate": 4.7082911370974645e-06,
644
+ "logits/chosen": -1.8625621795654297,
645
+ "logits/rejected": -1.36086905002594,
646
+ "logps/chosen": -548.0135498046875,
647
+ "logps/rejected": -759.1170654296875,
648
+ "loss": 0.1707,
649
+ "rewards/accuracies": 0.7250000238418579,
650
+ "rewards/chosen": -0.2390761822462082,
651
+ "rewards/margins": 0.2892398238182068,
652
+ "rewards/rejected": -0.5283160209655762,
653
+ "step": 450
654
+ },
655
+ {
656
+ "epoch": 0.25,
657
+ "learning_rate": 4.68608515331618e-06,
658
+ "logits/chosen": -1.7188348770141602,
659
+ "logits/rejected": -1.187195062637329,
660
+ "logps/chosen": -492.1756286621094,
661
+ "logps/rejected": -859.0760498046875,
662
+ "loss": 0.1414,
663
+ "rewards/accuracies": 0.875,
664
+ "rewards/chosen": -0.23255252838134766,
665
+ "rewards/margins": 0.3831843137741089,
666
+ "rewards/rejected": -0.6157368421554565,
667
+ "step": 460
668
+ },
669
+ {
670
+ "epoch": 0.25,
671
+ "learning_rate": 4.663121073711269e-06,
672
+ "logits/chosen": -1.5974490642547607,
673
+ "logits/rejected": -1.2564659118652344,
674
+ "logps/chosen": -336.80487060546875,
675
+ "logps/rejected": -661.6661376953125,
676
+ "loss": 0.1196,
677
+ "rewards/accuracies": 0.800000011920929,
678
+ "rewards/chosen": -0.14996656775474548,
679
+ "rewards/margins": 0.32413381338119507,
680
+ "rewards/rejected": -0.47410035133361816,
681
+ "step": 470
682
+ },
683
+ {
684
+ "epoch": 0.26,
685
+ "learning_rate": 4.63940686182103e-06,
686
+ "logits/chosen": -1.6767423152923584,
687
+ "logits/rejected": -1.1938632726669312,
688
+ "logps/chosen": -505.0990295410156,
689
+ "logps/rejected": -846.8779296875,
690
+ "loss": 0.159,
691
+ "rewards/accuracies": 0.875,
692
+ "rewards/chosen": -0.2421807050704956,
693
+ "rewards/margins": 0.3599635660648346,
694
+ "rewards/rejected": -0.6021442413330078,
695
+ "step": 480
696
+ },
697
+ {
698
+ "epoch": 0.26,
699
+ "learning_rate": 4.614950741316425e-06,
700
+ "logits/chosen": -1.529900312423706,
701
+ "logits/rejected": -1.0826785564422607,
702
+ "logps/chosen": -421.31707763671875,
703
+ "logps/rejected": -685.7420654296875,
704
+ "loss": 0.1721,
705
+ "rewards/accuracies": 0.7250000238418579,
706
+ "rewards/chosen": -0.22861020267009735,
707
+ "rewards/margins": 0.28916865587234497,
708
+ "rewards/rejected": -0.5177788734436035,
709
+ "step": 490
710
+ },
711
+ {
712
+ "epoch": 0.27,
713
+ "learning_rate": 4.589761193149254e-06,
714
+ "logits/chosen": -1.6966304779052734,
715
+ "logits/rejected": -0.9312071800231934,
716
+ "logps/chosen": -535.8978271484375,
717
+ "logps/rejected": -940.1627197265625,
718
+ "loss": 0.1144,
719
+ "rewards/accuracies": 0.949999988079071,
720
+ "rewards/chosen": -0.24013669788837433,
721
+ "rewards/margins": 0.4429057240486145,
722
+ "rewards/rejected": -0.6830424070358276,
723
+ "step": 500
724
+ },
725
+ {
726
+ "epoch": 0.27,
727
+ "learning_rate": 4.563846952611112e-06,
728
+ "logits/chosen": -1.6221929788589478,
729
+ "logits/rejected": -0.9574362635612488,
730
+ "logps/chosen": -428.6238708496094,
731
+ "logps/rejected": -704.8244018554688,
732
+ "loss": 0.0844,
733
+ "rewards/accuracies": 0.824999988079071,
734
+ "rewards/chosen": -0.195401132106781,
735
+ "rewards/margins": 0.3248489797115326,
736
+ "rewards/rejected": -0.520250141620636,
737
+ "step": 510
738
+ },
739
+ {
740
+ "epoch": 0.28,
741
+ "learning_rate": 4.537217006304141e-06,
742
+ "logits/chosen": -1.4427409172058105,
743
+ "logits/rejected": -1.003901481628418,
744
+ "logps/chosen": -429.660400390625,
745
+ "logps/rejected": -772.5963134765625,
746
+ "loss": 0.1457,
747
+ "rewards/accuracies": 0.824999988079071,
748
+ "rewards/chosen": -0.15723460912704468,
749
+ "rewards/margins": 0.3122255206108093,
750
+ "rewards/rejected": -0.4694600999355316,
751
+ "step": 520
752
+ },
753
+ {
754
+ "epoch": 0.28,
755
+ "learning_rate": 4.50988058902464e-06,
756
+ "logits/chosen": -1.2717740535736084,
757
+ "logits/rejected": -0.8480876684188843,
758
+ "logps/chosen": -367.8202209472656,
759
+ "logps/rejected": -776.1926879882812,
760
+ "loss": 0.1152,
761
+ "rewards/accuracies": 0.8500000238418579,
762
+ "rewards/chosen": -0.14414677023887634,
763
+ "rewards/margins": 0.38212689757347107,
764
+ "rewards/rejected": -0.5262737274169922,
765
+ "step": 530
766
+ },
767
+ {
768
+ "epoch": 0.29,
769
+ "learning_rate": 4.481847180560593e-06,
770
+ "logits/chosen": -1.5822323560714722,
771
+ "logits/rejected": -0.9035153388977051,
772
+ "logps/chosen": -439.30816650390625,
773
+ "logps/rejected": -708.1422729492188,
774
+ "loss": 0.2078,
775
+ "rewards/accuracies": 0.75,
776
+ "rewards/chosen": -0.18015776574611664,
777
+ "rewards/margins": 0.32809919118881226,
778
+ "rewards/rejected": -0.5082569122314453,
779
+ "step": 540
780
+ },
781
+ {
782
+ "epoch": 0.29,
783
+ "learning_rate": 4.453126502404253e-06,
784
+ "logits/chosen": -1.6248279809951782,
785
+ "logits/rejected": -0.9642871022224426,
786
+ "logps/chosen": -561.463623046875,
787
+ "logps/rejected": -740.08935546875,
788
+ "loss": 0.1773,
789
+ "rewards/accuracies": 0.800000011920929,
790
+ "rewards/chosen": -0.27120378613471985,
791
+ "rewards/margins": 0.28620854020118713,
792
+ "rewards/rejected": -0.557412326335907,
793
+ "step": 550
794
+ },
795
+ {
796
+ "epoch": 0.3,
797
+ "learning_rate": 4.423728514380892e-06,
798
+ "logits/chosen": -1.4605586528778076,
799
+ "logits/rejected": -0.8407928347587585,
800
+ "logps/chosen": -514.7965087890625,
801
+ "logps/rejected": -831.8440551757812,
802
+ "loss": 0.12,
803
+ "rewards/accuracies": 0.8500000238418579,
804
+ "rewards/chosen": -0.24941739439964294,
805
+ "rewards/margins": 0.36968275904655457,
806
+ "rewards/rejected": -0.6191002130508423,
807
+ "step": 560
808
+ },
809
+ {
810
+ "epoch": 0.3,
811
+ "learning_rate": 4.393663411194918e-06,
812
+ "logits/chosen": -1.4048388004302979,
813
+ "logits/rejected": -1.0212897062301636,
814
+ "logps/chosen": -490.42431640625,
815
+ "logps/rejected": -855.1259765625,
816
+ "loss": 0.1416,
817
+ "rewards/accuracies": 0.800000011920929,
818
+ "rewards/chosen": -0.24018950760364532,
819
+ "rewards/margins": 0.3597009778022766,
820
+ "rewards/rejected": -0.5998905301094055,
821
+ "step": 570
822
+ },
823
+ {
824
+ "epoch": 0.31,
825
+ "learning_rate": 4.362941618894523e-06,
826
+ "logits/chosen": -1.3778400421142578,
827
+ "logits/rejected": -0.983964741230011,
828
+ "logps/chosen": -586.1995849609375,
829
+ "logps/rejected": -981.2742309570312,
830
+ "loss": 0.1205,
831
+ "rewards/accuracies": 0.8500000238418579,
832
+ "rewards/chosen": -0.30800628662109375,
833
+ "rewards/margins": 0.3807603120803833,
834
+ "rewards/rejected": -0.6887666583061218,
835
+ "step": 580
836
+ },
837
+ {
838
+ "epoch": 0.31,
839
+ "learning_rate": 4.331573791256116e-06,
840
+ "logits/chosen": -1.4645698070526123,
841
+ "logits/rejected": -0.9271195530891418,
842
+ "logps/chosen": -621.4105224609375,
843
+ "logps/rejected": -899.0559692382812,
844
+ "loss": 0.1273,
845
+ "rewards/accuracies": 0.824999988079071,
846
+ "rewards/chosen": -0.2714827358722687,
847
+ "rewards/margins": 0.362936794757843,
848
+ "rewards/rejected": -0.6344195604324341,
849
+ "step": 590
850
+ },
851
+ {
852
+ "epoch": 0.32,
853
+ "learning_rate": 4.299570806089786e-06,
854
+ "logits/chosen": -1.6326652765274048,
855
+ "logits/rejected": -0.9927080273628235,
856
+ "logps/chosen": -490.701904296875,
857
+ "logps/rejected": -842.052734375,
858
+ "loss": 0.1023,
859
+ "rewards/accuracies": 0.875,
860
+ "rewards/chosen": -0.18793320655822754,
861
+ "rewards/margins": 0.408639132976532,
862
+ "rewards/rejected": -0.5965723395347595,
863
+ "step": 600
864
+ },
865
+ {
866
+ "epoch": 0.33,
867
+ "learning_rate": 4.266943761467057e-06,
868
+ "logits/chosen": -1.2816569805145264,
869
+ "logits/rejected": -0.8941723704338074,
870
+ "logps/chosen": -367.02191162109375,
871
+ "logps/rejected": -760.0553588867188,
872
+ "loss": 0.1443,
873
+ "rewards/accuracies": 0.800000011920929,
874
+ "rewards/chosen": -0.11693791300058365,
875
+ "rewards/margins": 0.39976662397384644,
876
+ "rewards/rejected": -0.5167044997215271,
877
+ "step": 610
878
+ },
879
+ {
880
+ "epoch": 0.33,
881
+ "learning_rate": 4.233703971872287e-06,
882
+ "logits/chosen": -1.8729069232940674,
883
+ "logits/rejected": -1.0977063179016113,
884
+ "logps/chosen": -393.6733093261719,
885
+ "logps/rejected": -763.2752075195312,
886
+ "loss": 0.1335,
887
+ "rewards/accuracies": 0.875,
888
+ "rewards/chosen": -0.08343084156513214,
889
+ "rewards/margins": 0.4291655123233795,
890
+ "rewards/rejected": -0.5125963687896729,
891
+ "step": 620
892
+ },
893
+ {
894
+ "epoch": 0.34,
895
+ "learning_rate": 4.1998629642789925e-06,
896
+ "logits/chosen": -1.5668641328811646,
897
+ "logits/rejected": -1.1349601745605469,
898
+ "logps/chosen": -426.9754333496094,
899
+ "logps/rejected": -820.5556640625,
900
+ "loss": 0.1742,
901
+ "rewards/accuracies": 0.800000011920929,
902
+ "rewards/chosen": -0.1329190582036972,
903
+ "rewards/margins": 0.36928990483283997,
904
+ "rewards/rejected": -0.5022088885307312,
905
+ "step": 630
906
+ },
907
+ {
908
+ "epoch": 0.34,
909
+ "learning_rate": 4.165432474152505e-06,
910
+ "logits/chosen": -1.5012271404266357,
911
+ "logits/rejected": -1.278693675994873,
912
+ "logps/chosen": -365.3034973144531,
913
+ "logps/rejected": -678.4292602539062,
914
+ "loss": 0.1737,
915
+ "rewards/accuracies": 0.7250000238418579,
916
+ "rewards/chosen": -0.13991737365722656,
917
+ "rewards/margins": 0.3017304837703705,
918
+ "rewards/rejected": -0.44164785742759705,
919
+ "step": 640
920
+ },
921
+ {
922
+ "epoch": 0.35,
923
+ "learning_rate": 4.130424441380308e-06,
924
+ "logits/chosen": -1.42804753780365,
925
+ "logits/rejected": -0.967817485332489,
926
+ "logps/chosen": -411.77801513671875,
927
+ "logps/rejected": -691.2272338867188,
928
+ "loss": 0.1527,
929
+ "rewards/accuracies": 0.8500000238418579,
930
+ "rewards/chosen": -0.1383361965417862,
931
+ "rewards/margins": 0.33920183777809143,
932
+ "rewards/rejected": -0.47753801941871643,
933
+ "step": 650
934
+ },
935
+ {
936
+ "epoch": 0.35,
937
+ "learning_rate": 4.09485100613151e-06,
938
+ "logits/chosen": -1.5555391311645508,
939
+ "logits/rejected": -1.1440869569778442,
940
+ "logps/chosen": -434.00335693359375,
941
+ "logps/rejected": -744.3508911132812,
942
+ "loss": 0.168,
943
+ "rewards/accuracies": 0.824999988079071,
944
+ "rewards/chosen": -0.19056299328804016,
945
+ "rewards/margins": 0.30447274446487427,
946
+ "rewards/rejected": -0.49503573775291443,
947
+ "step": 660
948
+ },
949
+ {
950
+ "epoch": 0.36,
951
+ "learning_rate": 4.058724504646834e-06,
952
+ "logits/chosen": -1.8099536895751953,
953
+ "logits/rejected": -1.0837316513061523,
954
+ "logps/chosen": -427.7793884277344,
955
+ "logps/rejected": -729.915283203125,
956
+ "loss": 0.1229,
957
+ "rewards/accuracies": 0.824999988079071,
958
+ "rewards/chosen": -0.1386745572090149,
959
+ "rewards/margins": 0.3526052236557007,
960
+ "rewards/rejected": -0.4912797808647156,
961
+ "step": 670
962
+ },
963
+ {
964
+ "epoch": 0.36,
965
+ "learning_rate": 4.022057464960632e-06,
966
+ "logits/chosen": -1.6691503524780273,
967
+ "logits/rejected": -1.33521568775177,
968
+ "logps/chosen": -428.7286071777344,
969
+ "logps/rejected": -789.0191040039062,
970
+ "loss": 0.1607,
971
+ "rewards/accuracies": 0.8999999761581421,
972
+ "rewards/chosen": -0.1496300995349884,
973
+ "rewards/margins": 0.35229435563087463,
974
+ "rewards/rejected": -0.5019243955612183,
975
+ "step": 680
976
+ },
977
+ {
978
+ "epoch": 0.37,
979
+ "learning_rate": 3.984862602556383e-06,
980
+ "logits/chosen": -1.6232519149780273,
981
+ "logits/rejected": -1.197933554649353,
982
+ "logps/chosen": -460.2228088378906,
983
+ "logps/rejected": -696.0914306640625,
984
+ "loss": 0.1346,
985
+ "rewards/accuracies": 0.7749999761581421,
986
+ "rewards/chosen": -0.18895591795444489,
987
+ "rewards/margins": 0.28954973816871643,
988
+ "rewards/rejected": -0.4785057008266449,
989
+ "step": 690
990
+ },
991
+ {
992
+ "epoch": 0.37,
993
+ "learning_rate": 3.947152815957187e-06,
994
+ "logits/chosen": -1.5458933115005493,
995
+ "logits/rejected": -1.116236925125122,
996
+ "logps/chosen": -433.87322998046875,
997
+ "logps/rejected": -756.8858642578125,
998
+ "loss": 0.1492,
999
+ "rewards/accuracies": 0.8500000238418579,
1000
+ "rewards/chosen": -0.22537223994731903,
1001
+ "rewards/margins": 0.34959647059440613,
1002
+ "rewards/rejected": -0.5749687552452087,
1003
+ "step": 700
1004
+ },
1005
+ {
1006
+ "epoch": 0.38,
1007
+ "learning_rate": 3.908941182252785e-06,
1008
+ "logits/chosen": -1.5793603658676147,
1009
+ "logits/rejected": -0.9729734659194946,
1010
+ "logps/chosen": -458.96368408203125,
1011
+ "logps/rejected": -781.1962890625,
1012
+ "loss": 0.1615,
1013
+ "rewards/accuracies": 0.875,
1014
+ "rewards/chosen": -0.1985333412885666,
1015
+ "rewards/margins": 0.3719526529312134,
1016
+ "rewards/rejected": -0.5704860091209412,
1017
+ "step": 710
1018
+ },
1019
+ {
1020
+ "epoch": 0.38,
1021
+ "learning_rate": 3.8702409525646535e-06,
1022
+ "logits/chosen": -1.6880747079849243,
1023
+ "logits/rejected": -1.0946999788284302,
1024
+ "logps/chosen": -550.5426635742188,
1025
+ "logps/rejected": -861.6978759765625,
1026
+ "loss": 0.1362,
1027
+ "rewards/accuracies": 0.824999988079071,
1028
+ "rewards/chosen": -0.1908409297466278,
1029
+ "rewards/margins": 0.3778737485408783,
1030
+ "rewards/rejected": -0.5687146782875061,
1031
+ "step": 720
1032
+ },
1033
+ {
1034
+ "epoch": 0.39,
1035
+ "learning_rate": 3.8310655474507495e-06,
1036
+ "logits/chosen": -1.7694594860076904,
1037
+ "logits/rejected": -1.1918199062347412,
1038
+ "logps/chosen": -443.54736328125,
1039
+ "logps/rejected": -717.8020629882812,
1040
+ "loss": 0.1418,
1041
+ "rewards/accuracies": 0.7250000238418579,
1042
+ "rewards/chosen": -0.16424255073070526,
1043
+ "rewards/margins": 0.30146175622940063,
1044
+ "rewards/rejected": -0.4657043516635895,
1045
+ "step": 730
1046
+ },
1047
+ {
1048
+ "epoch": 0.39,
1049
+ "learning_rate": 3.7914285522515002e-06,
1050
+ "logits/chosen": -1.539620280265808,
1051
+ "logits/rejected": -1.3648602962493896,
1052
+ "logps/chosen": -468.6385192871094,
1053
+ "logps/rejected": -892.7566528320312,
1054
+ "loss": 0.1552,
1055
+ "rewards/accuracies": 0.875,
1056
+ "rewards/chosen": -0.22199459373950958,
1057
+ "rewards/margins": 0.36069172620773315,
1058
+ "rewards/rejected": -0.5826863050460815,
1059
+ "step": 740
1060
+ },
1061
+ {
1062
+ "epoch": 0.4,
1063
+ "learning_rate": 3.751343712378639e-06,
1064
+ "logits/chosen": -1.68185555934906,
1065
+ "logits/rejected": -1.0438605546951294,
1066
+ "logps/chosen": -377.9205627441406,
1067
+ "logps/rejected": -688.3480834960938,
1068
+ "loss": 0.1417,
1069
+ "rewards/accuracies": 0.75,
1070
+ "rewards/chosen": -0.15213271975517273,
1071
+ "rewards/margins": 0.3400834798812866,
1072
+ "rewards/rejected": -0.49221619963645935,
1073
+ "step": 750
1074
+ },
1075
+ {
1076
+ "epoch": 0.41,
1077
+ "learning_rate": 3.710824928548546e-06,
1078
+ "logits/chosen": -1.7241179943084717,
1079
+ "logits/rejected": -1.1749062538146973,
1080
+ "logps/chosen": -398.90521240234375,
1081
+ "logps/rejected": -792.80078125,
1082
+ "loss": 0.1218,
1083
+ "rewards/accuracies": 0.8999999761581421,
1084
+ "rewards/chosen": -0.14098913967609406,
1085
+ "rewards/margins": 0.4070391058921814,
1086
+ "rewards/rejected": -0.5480281710624695,
1087
+ "step": 760
1088
+ },
1089
+ {
1090
+ "epoch": 0.41,
1091
+ "learning_rate": 3.6698862519617225e-06,
1092
+ "logits/chosen": -1.862091064453125,
1093
+ "logits/rejected": -1.0774017572402954,
1094
+ "logps/chosen": -380.6012878417969,
1095
+ "logps/rejected": -803.9888916015625,
1096
+ "loss": 0.1009,
1097
+ "rewards/accuracies": 0.8999999761581421,
1098
+ "rewards/chosen": -0.12483291327953339,
1099
+ "rewards/margins": 0.46553611755371094,
1100
+ "rewards/rejected": -0.5903691053390503,
1101
+ "step": 770
1102
+ },
1103
+ {
1104
+ "epoch": 0.42,
1105
+ "learning_rate": 3.6285418794300793e-06,
1106
+ "logits/chosen": -1.416322946548462,
1107
+ "logits/rejected": -0.8399195671081543,
1108
+ "logps/chosen": -444.05523681640625,
1109
+ "logps/rejected": -761.3108520507812,
1110
+ "loss": 0.1571,
1111
+ "rewards/accuracies": 0.75,
1112
+ "rewards/chosen": -0.20516355335712433,
1113
+ "rewards/margins": 0.3872140049934387,
1114
+ "rewards/rejected": -0.5923775434494019,
1115
+ "step": 780
1116
+ },
1117
+ {
1118
+ "epoch": 0.42,
1119
+ "learning_rate": 3.5868061484537365e-06,
1120
+ "logits/chosen": -1.39794921875,
1121
+ "logits/rejected": -0.8267971873283386,
1122
+ "logps/chosen": -507.1766052246094,
1123
+ "logps/rejected": -875.86962890625,
1124
+ "loss": 0.1497,
1125
+ "rewards/accuracies": 0.8999999761581421,
1126
+ "rewards/chosen": -0.23690223693847656,
1127
+ "rewards/margins": 0.40496787428855896,
1128
+ "rewards/rejected": -0.6418701410293579,
1129
+ "step": 790
1130
+ },
1131
+ {
1132
+ "epoch": 0.43,
1133
+ "learning_rate": 3.5446935322490285e-06,
1134
+ "logits/chosen": -1.7719318866729736,
1135
+ "logits/rejected": -0.9355955123901367,
1136
+ "logps/chosen": -544.9541015625,
1137
+ "logps/rejected": -865.5302734375,
1138
+ "loss": 0.1963,
1139
+ "rewards/accuracies": 0.7749999761581421,
1140
+ "rewards/chosen": -0.2105661928653717,
1141
+ "rewards/margins": 0.3866081237792969,
1142
+ "rewards/rejected": -0.5971742868423462,
1143
+ "step": 800
1144
+ },
1145
+ {
1146
+ "epoch": 0.43,
1147
+ "learning_rate": 3.502218634729447e-06,
1148
+ "logits/chosen": -1.6639026403427124,
1149
+ "logits/rejected": -1.067781925201416,
1150
+ "logps/chosen": -575.9091796875,
1151
+ "logps/rejected": -838.0983276367188,
1152
+ "loss": 0.1233,
1153
+ "rewards/accuracies": 0.875,
1154
+ "rewards/chosen": -0.23653562366962433,
1155
+ "rewards/margins": 0.3239360749721527,
1156
+ "rewards/rejected": -0.5604716539382935,
1157
+ "step": 810
1158
+ },
1159
+ {
1160
+ "epoch": 0.44,
1161
+ "learning_rate": 3.459396185441265e-06,
1162
+ "logits/chosen": -1.7067358493804932,
1163
+ "logits/rejected": -1.0498546361923218,
1164
+ "logps/chosen": -398.35516357421875,
1165
+ "logps/rejected": -626.8757934570312,
1166
+ "loss": 0.1608,
1167
+ "rewards/accuracies": 0.7749999761581421,
1168
+ "rewards/chosen": -0.1356094628572464,
1169
+ "rewards/margins": 0.3172938823699951,
1170
+ "rewards/rejected": -0.4529033601284027,
1171
+ "step": 820
1172
+ },
1173
+ {
1174
+ "epoch": 0.44,
1175
+ "learning_rate": 3.4162410344555834e-06,
1176
+ "logits/chosen": -1.9210001230239868,
1177
+ "logits/rejected": -1.1206413507461548,
1178
+ "logps/chosen": -405.9615783691406,
1179
+ "logps/rejected": -725.4310913085938,
1180
+ "loss": 0.125,
1181
+ "rewards/accuracies": 0.8500000238418579,
1182
+ "rewards/chosen": -0.10851490497589111,
1183
+ "rewards/margins": 0.38589829206466675,
1184
+ "rewards/rejected": -0.49441319704055786,
1185
+ "step": 830
1186
+ },
1187
+ {
1188
+ "epoch": 0.45,
1189
+ "learning_rate": 3.3727681472185937e-06,
1190
+ "logits/chosen": -1.6562303304672241,
1191
+ "logits/rejected": -1.19851553440094,
1192
+ "logps/chosen": -486.35107421875,
1193
+ "logps/rejected": -963.0572509765625,
1194
+ "loss": 0.1105,
1195
+ "rewards/accuracies": 0.8999999761581421,
1196
+ "rewards/chosen": -0.22348380088806152,
1197
+ "rewards/margins": 0.4227561056613922,
1198
+ "rewards/rejected": -0.6462398767471313,
1199
+ "step": 840
1200
+ },
1201
+ {
1202
+ "epoch": 0.45,
1203
+ "learning_rate": 3.3289925993618217e-06,
1204
+ "logits/chosen": -1.5856201648712158,
1205
+ "logits/rejected": -1.0767395496368408,
1206
+ "logps/chosen": -526.1747436523438,
1207
+ "logps/rejected": -797.7916870117188,
1208
+ "loss": 0.137,
1209
+ "rewards/accuracies": 0.800000011920929,
1210
+ "rewards/chosen": -0.2687681019306183,
1211
+ "rewards/margins": 0.2989320755004883,
1212
+ "rewards/rejected": -0.567700207233429,
1213
+ "step": 850
1214
+ },
1215
+ {
1216
+ "epoch": 0.46,
1217
+ "learning_rate": 3.2849295714741643e-06,
1218
+ "logits/chosen": -1.7678568363189697,
1219
+ "logits/rejected": -1.2151532173156738,
1220
+ "logps/chosen": -597.7952880859375,
1221
+ "logps/rejected": -848.87841796875,
1222
+ "loss": 0.1308,
1223
+ "rewards/accuracies": 0.800000011920929,
1224
+ "rewards/chosen": -0.27867773175239563,
1225
+ "rewards/margins": 0.30206385254859924,
1226
+ "rewards/rejected": -0.5807415843009949,
1227
+ "step": 860
1228
+ },
1229
+ {
1230
+ "epoch": 0.46,
1231
+ "learning_rate": 3.2405943438375287e-06,
1232
+ "logits/chosen": -1.7643588781356812,
1233
+ "logits/rejected": -1.099827527999878,
1234
+ "logps/chosen": -423.9742126464844,
1235
+ "logps/rejected": -774.4637451171875,
1236
+ "loss": 0.0974,
1237
+ "rewards/accuracies": 0.875,
1238
+ "rewards/chosen": -0.1482265591621399,
1239
+ "rewards/margins": 0.410900741815567,
1240
+ "rewards/rejected": -0.5591272711753845,
1241
+ "step": 870
1242
+ },
1243
+ {
1244
+ "epoch": 0.47,
1245
+ "learning_rate": 3.1960022911279036e-06,
1246
+ "logits/chosen": -1.5414252281188965,
1247
+ "logits/rejected": -1.1484423875808716,
1248
+ "logps/chosen": -493.69464111328125,
1249
+ "logps/rejected": -835.8029174804688,
1250
+ "loss": 0.1526,
1251
+ "rewards/accuracies": 0.824999988079071,
1252
+ "rewards/chosen": -0.19744431972503662,
1253
+ "rewards/margins": 0.36680763959884644,
1254
+ "rewards/rejected": -0.5642520189285278,
1255
+ "step": 880
1256
+ },
1257
+ {
1258
+ "epoch": 0.47,
1259
+ "learning_rate": 3.1511688770836844e-06,
1260
+ "logits/chosen": -1.511249303817749,
1261
+ "logits/rejected": -1.3401678800582886,
1262
+ "logps/chosen": -404.75933837890625,
1263
+ "logps/rejected": -805.8262939453125,
1264
+ "loss": 0.1035,
1265
+ "rewards/accuracies": 0.800000011920929,
1266
+ "rewards/chosen": -0.18373355269432068,
1267
+ "rewards/margins": 0.3318132758140564,
1268
+ "rewards/rejected": -0.5155468583106995,
1269
+ "step": 890
1270
+ },
1271
+ {
1272
+ "epoch": 0.48,
1273
+ "learning_rate": 3.1061096491431307e-06,
1274
+ "logits/chosen": -1.8831459283828735,
1275
+ "logits/rejected": -1.2160544395446777,
1276
+ "logps/chosen": -447.65032958984375,
1277
+ "logps/rejected": -843.8660278320312,
1278
+ "loss": 0.1345,
1279
+ "rewards/accuracies": 0.875,
1280
+ "rewards/chosen": -0.13529552519321442,
1281
+ "rewards/margins": 0.40966707468032837,
1282
+ "rewards/rejected": -0.5449625849723816,
1283
+ "step": 900
1284
+ },
1285
+ {
1286
+ "epoch": 0.49,
1287
+ "learning_rate": 3.0608402330527796e-06,
1288
+ "logits/chosen": -1.6770378351211548,
1289
+ "logits/rejected": -0.9972168803215027,
1290
+ "logps/chosen": -379.8583984375,
1291
+ "logps/rejected": -719.3693237304688,
1292
+ "loss": 0.1765,
1293
+ "rewards/accuracies": 0.824999988079071,
1294
+ "rewards/chosen": -0.15062110126018524,
1295
+ "rewards/margins": 0.3817873001098633,
1296
+ "rewards/rejected": -0.5324083566665649,
1297
+ "step": 910
1298
+ },
1299
+ {
1300
+ "epoch": 0.49,
1301
+ "learning_rate": 3.0153763274487176e-06,
1302
+ "logits/chosen": -1.4134846925735474,
1303
+ "logits/rejected": -0.966874897480011,
1304
+ "logps/chosen": -441.3450622558594,
1305
+ "logps/rejected": -707.3884887695312,
1306
+ "loss": 0.1235,
1307
+ "rewards/accuracies": 0.75,
1308
+ "rewards/chosen": -0.21538302302360535,
1309
+ "rewards/margins": 0.3049529790878296,
1310
+ "rewards/rejected": -0.5203360319137573,
1311
+ "step": 920
1312
+ },
1313
+ {
1314
+ "epoch": 0.5,
1315
+ "learning_rate": 2.9697336984125683e-06,
1316
+ "logits/chosen": -1.6667283773422241,
1317
+ "logits/rejected": -1.0133411884307861,
1318
+ "logps/chosen": -401.2959899902344,
1319
+ "logps/rejected": -851.93701171875,
1320
+ "loss": 0.1206,
1321
+ "rewards/accuracies": 0.925000011920929,
1322
+ "rewards/chosen": -0.15221676230430603,
1323
+ "rewards/margins": 0.45876413583755493,
1324
+ "rewards/rejected": -0.6109809279441833,
1325
+ "step": 930
1326
+ },
1327
+ {
1328
+ "epoch": 0.5,
1329
+ "learning_rate": 2.923928174004094e-06,
1330
+ "logits/chosen": -1.8437814712524414,
1331
+ "logits/rejected": -1.0747764110565186,
1332
+ "logps/chosen": -470.7169494628906,
1333
+ "logps/rejected": -732.7559814453125,
1334
+ "loss": 0.1247,
1335
+ "rewards/accuracies": 0.875,
1336
+ "rewards/chosen": -0.13192041218280792,
1337
+ "rewards/margins": 0.3781585991382599,
1338
+ "rewards/rejected": -0.5100789666175842,
1339
+ "step": 940
1340
+ },
1341
+ {
1342
+ "epoch": 0.51,
1343
+ "learning_rate": 2.8779756387723036e-06,
1344
+ "logits/chosen": -1.7663402557373047,
1345
+ "logits/rejected": -1.3018739223480225,
1346
+ "logps/chosen": -446.77490234375,
1347
+ "logps/rejected": -766.7832641601562,
1348
+ "loss": 0.1146,
1349
+ "rewards/accuracies": 0.875,
1350
+ "rewards/chosen": -0.1646779477596283,
1351
+ "rewards/margins": 0.36457663774490356,
1352
+ "rewards/rejected": -0.5292545557022095,
1353
+ "step": 950
1354
+ },
1355
+ {
1356
+ "epoch": 0.51,
1357
+ "learning_rate": 2.831892028246968e-06,
1358
+ "logits/chosen": -1.848724603652954,
1359
+ "logits/rejected": -1.216956377029419,
1360
+ "logps/chosen": -418.67645263671875,
1361
+ "logps/rejected": -703.2694702148438,
1362
+ "loss": 0.1209,
1363
+ "rewards/accuracies": 0.800000011920929,
1364
+ "rewards/chosen": -0.14982689917087555,
1365
+ "rewards/margins": 0.36974358558654785,
1366
+ "rewards/rejected": -0.5195704698562622,
1367
+ "step": 960
1368
+ },
1369
+ {
1370
+ "epoch": 0.52,
1371
+ "learning_rate": 2.7856933234124617e-06,
1372
+ "logits/chosen": -1.7911808490753174,
1373
+ "logits/rejected": -1.0922878980636597,
1374
+ "logps/chosen": -448.37603759765625,
1375
+ "logps/rejected": -834.2364501953125,
1376
+ "loss": 0.1538,
1377
+ "rewards/accuracies": 0.8500000238418579,
1378
+ "rewards/chosen": -0.17878659069538116,
1379
+ "rewards/margins": 0.3954610228538513,
1380
+ "rewards/rejected": -0.5742476582527161,
1381
+ "step": 970
1382
+ },
1383
+ {
1384
+ "epoch": 0.52,
1385
+ "learning_rate": 2.7393955451658387e-06,
1386
+ "logits/chosen": -1.7210479974746704,
1387
+ "logits/rejected": -1.2294584512710571,
1388
+ "logps/chosen": -514.4754028320312,
1389
+ "logps/rejected": -868.5929565429688,
1390
+ "loss": 0.1626,
1391
+ "rewards/accuracies": 0.875,
1392
+ "rewards/chosen": -0.21787652373313904,
1393
+ "rewards/margins": 0.3894199728965759,
1394
+ "rewards/rejected": -0.6072965264320374,
1395
+ "step": 980
1396
+ },
1397
+ {
1398
+ "epoch": 0.53,
1399
+ "learning_rate": 2.6930147487610667e-06,
1400
+ "logits/chosen": -1.5907623767852783,
1401
+ "logits/rejected": -0.78331458568573,
1402
+ "logps/chosen": -462.7984313964844,
1403
+ "logps/rejected": -805.7174072265625,
1404
+ "loss": 0.1373,
1405
+ "rewards/accuracies": 0.824999988079071,
1406
+ "rewards/chosen": -0.17765957117080688,
1407
+ "rewards/margins": 0.40347957611083984,
1408
+ "rewards/rejected": -0.581139087677002,
1409
+ "step": 990
1410
+ },
1411
+ {
1412
+ "epoch": 0.53,
1413
+ "learning_rate": 2.6465670182413487e-06,
1414
+ "logits/chosen": -1.6310056447982788,
1415
+ "logits/rejected": -1.0298982858657837,
1416
+ "logps/chosen": -411.04937744140625,
1417
+ "logps/rejected": -758.7462158203125,
1418
+ "loss": 0.1237,
1419
+ "rewards/accuracies": 0.824999988079071,
1420
+ "rewards/chosen": -0.1506483405828476,
1421
+ "rewards/margins": 0.36691543459892273,
1422
+ "rewards/rejected": -0.5175637602806091,
1423
+ "step": 1000
1424
+ },
1425
+ {
1426
+ "epoch": 0.54,
1427
+ "learning_rate": 2.6000684608614594e-06,
1428
+ "logits/chosen": -1.6570842266082764,
1429
+ "logits/rejected": -0.8277125358581543,
1430
+ "logps/chosen": -506.580810546875,
1431
+ "logps/rejected": -801.989990234375,
1432
+ "loss": 0.1436,
1433
+ "rewards/accuracies": 0.8999999761581421,
1434
+ "rewards/chosen": -0.1983393281698227,
1435
+ "rewards/margins": 0.38503485918045044,
1436
+ "rewards/rejected": -0.5833742022514343,
1437
+ "step": 1010
1438
+ },
1439
+ {
1440
+ "epoch": 0.54,
1441
+ "learning_rate": 2.5535352015020338e-06,
1442
+ "logits/chosen": -1.528637170791626,
1443
+ "logits/rejected": -0.8484199643135071,
1444
+ "logps/chosen": -470.8020935058594,
1445
+ "logps/rejected": -820.8448486328125,
1446
+ "loss": 0.1363,
1447
+ "rewards/accuracies": 0.875,
1448
+ "rewards/chosen": -0.20915472507476807,
1449
+ "rewards/margins": 0.38422003388404846,
1450
+ "rewards/rejected": -0.5933747887611389,
1451
+ "step": 1020
1452
+ },
1453
+ {
1454
+ "epoch": 0.55,
1455
+ "learning_rate": 2.506983377077741e-06,
1456
+ "logits/chosen": -1.3463196754455566,
1457
+ "logits/rejected": -1.018822193145752,
1458
+ "logps/chosen": -464.81524658203125,
1459
+ "logps/rejected": -807.076171875,
1460
+ "loss": 0.1584,
1461
+ "rewards/accuracies": 0.8500000238418579,
1462
+ "rewards/chosen": -0.21939115226268768,
1463
+ "rewards/margins": 0.3329920172691345,
1464
+ "rewards/rejected": -0.5523831844329834,
1465
+ "step": 1030
1466
+ },
1467
+ {
1468
+ "epoch": 0.55,
1469
+ "learning_rate": 2.460429130941289e-06,
1470
+ "logits/chosen": -1.4068031311035156,
1471
+ "logits/rejected": -0.9966346621513367,
1472
+ "logps/chosen": -443.41583251953125,
1473
+ "logps/rejected": -826.1185302734375,
1474
+ "loss": 0.1182,
1475
+ "rewards/accuracies": 0.875,
1476
+ "rewards/chosen": -0.1990918219089508,
1477
+ "rewards/margins": 0.39130455255508423,
1478
+ "rewards/rejected": -0.5903963446617126,
1479
+ "step": 1040
1480
+ },
1481
+ {
1482
+ "epoch": 0.56,
1483
+ "learning_rate": 2.413888607285192e-06,
1484
+ "logits/chosen": -1.2919907569885254,
1485
+ "logits/rejected": -0.9193531274795532,
1486
+ "logps/chosen": -496.358642578125,
1487
+ "logps/rejected": -845.7939453125,
1488
+ "loss": 0.173,
1489
+ "rewards/accuracies": 0.8500000238418579,
1490
+ "rewards/chosen": -0.24995890259742737,
1491
+ "rewards/margins": 0.3503498435020447,
1492
+ "rewards/rejected": -0.6003087162971497,
1493
+ "step": 1050
1494
+ },
1495
+ {
1496
+ "epoch": 0.57,
1497
+ "learning_rate": 2.367377945543249e-06,
1498
+ "logits/chosen": -1.6841480731964111,
1499
+ "logits/rejected": -0.907370924949646,
1500
+ "logps/chosen": -446.6328125,
1501
+ "logps/rejected": -884.2018432617188,
1502
+ "loss": 0.1068,
1503
+ "rewards/accuracies": 0.925000011920929,
1504
+ "rewards/chosen": -0.15798960626125336,
1505
+ "rewards/margins": 0.4673282206058502,
1506
+ "rewards/rejected": -0.6253177523612976,
1507
+ "step": 1060
1508
+ },
1509
+ {
1510
+ "epoch": 0.57,
1511
+ "learning_rate": 2.320913274793676e-06,
1512
+ "logits/chosen": -1.7113037109375,
1513
+ "logits/rejected": -1.1816798448562622,
1514
+ "logps/chosen": -410.67645263671875,
1515
+ "logps/rejected": -770.4984741210938,
1516
+ "loss": 0.1423,
1517
+ "rewards/accuracies": 0.800000011920929,
1518
+ "rewards/chosen": -0.1650415062904358,
1519
+ "rewards/margins": 0.37022119760513306,
1520
+ "rewards/rejected": -0.5352627038955688,
1521
+ "step": 1070
1522
+ },
1523
+ {
1524
+ "epoch": 0.58,
1525
+ "learning_rate": 2.27451070816582e-06,
1526
+ "logits/chosen": -1.6226348876953125,
1527
+ "logits/rejected": -0.9200002551078796,
1528
+ "logps/chosen": -518.1405029296875,
1529
+ "logps/rejected": -891.6884765625,
1530
+ "loss": 0.1105,
1531
+ "rewards/accuracies": 0.8500000238418579,
1532
+ "rewards/chosen": -0.16074618697166443,
1533
+ "rewards/margins": 0.47465044260025024,
1534
+ "rewards/rejected": -0.6353966593742371,
1535
+ "step": 1080
1536
+ },
1537
+ {
1538
+ "epoch": 0.58,
1539
+ "learning_rate": 2.228186337252414e-06,
1540
+ "logits/chosen": -1.7294307947158813,
1541
+ "logits/rejected": -0.8779215812683105,
1542
+ "logps/chosen": -516.14013671875,
1543
+ "logps/rejected": -824.5764770507812,
1544
+ "loss": 0.1407,
1545
+ "rewards/accuracies": 0.875,
1546
+ "rewards/chosen": -0.17551277577877045,
1547
+ "rewards/margins": 0.40844354033470154,
1548
+ "rewards/rejected": -0.583956241607666,
1549
+ "step": 1090
1550
+ },
1551
+ {
1552
+ "epoch": 0.59,
1553
+ "learning_rate": 2.1819562265292946e-06,
1554
+ "logits/chosen": -1.5813789367675781,
1555
+ "logits/rejected": -1.001509666442871,
1556
+ "logps/chosen": -429.26593017578125,
1557
+ "logps/rejected": -789.7249755859375,
1558
+ "loss": 0.1574,
1559
+ "rewards/accuracies": 0.875,
1560
+ "rewards/chosen": -0.21274442970752716,
1561
+ "rewards/margins": 0.3669392764568329,
1562
+ "rewards/rejected": -0.5796837210655212,
1563
+ "step": 1100
1564
+ },
1565
+ {
1566
+ "epoch": 0.59,
1567
+ "learning_rate": 2.1358364077845236e-06,
1568
+ "logits/chosen": -1.533307671546936,
1569
+ "logits/rejected": -0.9590204954147339,
1570
+ "logps/chosen": -387.72381591796875,
1571
+ "logps/rejected": -828.5607299804688,
1572
+ "loss": 0.1014,
1573
+ "rewards/accuracies": 0.8999999761581421,
1574
+ "rewards/chosen": -0.14084286987781525,
1575
+ "rewards/margins": 0.4476155638694763,
1576
+ "rewards/rejected": -0.5884584188461304,
1577
+ "step": 1110
1578
+ },
1579
+ {
1580
+ "epoch": 0.6,
1581
+ "learning_rate": 2.089842874558849e-06,
1582
+ "logits/chosen": -1.3846327066421509,
1583
+ "logits/rejected": -1.0313770771026611,
1584
+ "logps/chosen": -479.6465759277344,
1585
+ "logps/rejected": -906.1482543945312,
1586
+ "loss": 0.1133,
1587
+ "rewards/accuracies": 0.875,
1588
+ "rewards/chosen": -0.22880907356739044,
1589
+ "rewards/margins": 0.4139330983161926,
1590
+ "rewards/rejected": -0.6427421569824219,
1591
+ "step": 1120
1592
+ },
1593
+ {
1594
+ "epoch": 0.6,
1595
+ "learning_rate": 2.0439915765994242e-06,
1596
+ "logits/chosen": -1.5441999435424805,
1597
+ "logits/rejected": -0.8765427470207214,
1598
+ "logps/chosen": -374.98504638671875,
1599
+ "logps/rejected": -726.3242797851562,
1600
+ "loss": 0.1078,
1601
+ "rewards/accuracies": 0.8500000238418579,
1602
+ "rewards/chosen": -0.1631808578968048,
1603
+ "rewards/margins": 0.38433948159217834,
1604
+ "rewards/rejected": -0.5475203394889832,
1605
+ "step": 1130
1606
+ },
1607
+ {
1608
+ "epoch": 0.61,
1609
+ "learning_rate": 1.9982984143287186e-06,
1610
+ "logits/chosen": -1.7160451412200928,
1611
+ "logits/rejected": -0.9389771223068237,
1612
+ "logps/chosen": -419.886962890625,
1613
+ "logps/rejected": -727.5339965820312,
1614
+ "loss": 0.1209,
1615
+ "rewards/accuracies": 0.8500000238418579,
1616
+ "rewards/chosen": -0.1307600438594818,
1617
+ "rewards/margins": 0.4173372685909271,
1618
+ "rewards/rejected": -0.5480973720550537,
1619
+ "step": 1140
1620
+ },
1621
+ {
1622
+ "epoch": 0.61,
1623
+ "learning_rate": 1.95277923333053e-06,
1624
+ "logits/chosen": -1.5577538013458252,
1625
+ "logits/rejected": -0.9766386151313782,
1626
+ "logps/chosen": -432.010498046875,
1627
+ "logps/rejected": -772.4149780273438,
1628
+ "loss": 0.1015,
1629
+ "rewards/accuracies": 0.75,
1630
+ "rewards/chosen": -0.1316412091255188,
1631
+ "rewards/margins": 0.40984097123146057,
1632
+ "rewards/rejected": -0.541482150554657,
1633
+ "step": 1150
1634
+ },
1635
+ {
1636
+ "epoch": 0.62,
1637
+ "learning_rate": 1.9074498188550156e-06,
1638
+ "logits/chosen": -1.6150667667388916,
1639
+ "logits/rejected": -1.0481829643249512,
1640
+ "logps/chosen": -460.06781005859375,
1641
+ "logps/rejected": -748.8250122070312,
1642
+ "loss": 0.1577,
1643
+ "rewards/accuracies": 0.7749999761581421,
1644
+ "rewards/chosen": -0.18149954080581665,
1645
+ "rewards/margins": 0.34171923995018005,
1646
+ "rewards/rejected": -0.5232187509536743,
1647
+ "step": 1160
1648
+ },
1649
+ {
1650
+ "epoch": 0.62,
1651
+ "learning_rate": 1.862325890344643e-06,
1652
+ "logits/chosen": -1.3022327423095703,
1653
+ "logits/rejected": -0.9266065359115601,
1654
+ "logps/chosen": -367.62823486328125,
1655
+ "logps/rejected": -806.5985107421875,
1656
+ "loss": 0.1656,
1657
+ "rewards/accuracies": 0.875,
1658
+ "rewards/chosen": -0.17992374300956726,
1659
+ "rewards/margins": 0.42552104592323303,
1660
+ "rewards/rejected": -0.6054448485374451,
1661
+ "step": 1170
1662
+ },
1663
+ {
1664
+ "epoch": 0.63,
1665
+ "learning_rate": 1.817423095982972e-06,
1666
+ "logits/chosen": -1.3970632553100586,
1667
+ "logits/rejected": -0.9412476420402527,
1668
+ "logps/chosen": -451.11883544921875,
1669
+ "logps/rejected": -783.0731201171875,
1670
+ "loss": 0.102,
1671
+ "rewards/accuracies": 0.8500000238418579,
1672
+ "rewards/chosen": -0.20501062273979187,
1673
+ "rewards/margins": 0.37199467420578003,
1674
+ "rewards/rejected": -0.5770053267478943,
1675
+ "step": 1180
1676
+ },
1677
+ {
1678
+ "epoch": 0.63,
1679
+ "learning_rate": 1.7727570072681293e-06,
1680
+ "logits/chosen": -1.4293451309204102,
1681
+ "logits/rejected": -0.8616847991943359,
1682
+ "logps/chosen": -406.25042724609375,
1683
+ "logps/rejected": -737.0385131835938,
1684
+ "loss": 0.1378,
1685
+ "rewards/accuracies": 0.824999988079071,
1686
+ "rewards/chosen": -0.16194511950016022,
1687
+ "rewards/margins": 0.3638822138309479,
1688
+ "rewards/rejected": -0.5258272886276245,
1689
+ "step": 1190
1690
+ },
1691
+ {
1692
+ "epoch": 0.64,
1693
+ "learning_rate": 1.7283431136128961e-06,
1694
+ "logits/chosen": -1.6002616882324219,
1695
+ "logits/rejected": -1.116288423538208,
1696
+ "logps/chosen": -449.384521484375,
1697
+ "logps/rejected": -805.2763061523438,
1698
+ "loss": 0.1454,
1699
+ "rewards/accuracies": 0.875,
1700
+ "rewards/chosen": -0.20747177302837372,
1701
+ "rewards/margins": 0.3553561270236969,
1702
+ "rewards/rejected": -0.5628278851509094,
1703
+ "step": 1200
1704
+ },
1705
+ {
1706
+ "epoch": 0.65,
1707
+ "learning_rate": 1.6841968169732478e-06,
1708
+ "logits/chosen": -1.5592294931411743,
1709
+ "logits/rejected": -1.054216742515564,
1710
+ "logps/chosen": -448.9071350097656,
1711
+ "logps/rejected": -851.6107177734375,
1712
+ "loss": 0.1204,
1713
+ "rewards/accuracies": 0.8500000238418579,
1714
+ "rewards/chosen": -0.18522223830223083,
1715
+ "rewards/margins": 0.4141850471496582,
1716
+ "rewards/rejected": -0.5994073152542114,
1717
+ "step": 1210
1718
+ },
1719
+ {
1720
+ "epoch": 0.65,
1721
+ "learning_rate": 1.6403334265072284e-06,
1722
+ "logits/chosen": -1.6474437713623047,
1723
+ "logits/rejected": -0.8614113926887512,
1724
+ "logps/chosen": -453.735107421875,
1725
+ "logps/rejected": -801.1546630859375,
1726
+ "loss": 0.1081,
1727
+ "rewards/accuracies": 0.8999999761581421,
1728
+ "rewards/chosen": -0.1977526694536209,
1729
+ "rewards/margins": 0.38285189867019653,
1730
+ "rewards/rejected": -0.5806045532226562,
1731
+ "step": 1220
1732
+ },
1733
+ {
1734
+ "epoch": 0.66,
1735
+ "learning_rate": 1.5967681532660066e-06,
1736
+ "logits/chosen": -1.2708427906036377,
1737
+ "logits/rejected": -0.9732850790023804,
1738
+ "logps/chosen": -437.337890625,
1739
+ "logps/rejected": -822.8092041015625,
1740
+ "loss": 0.1336,
1741
+ "rewards/accuracies": 0.8999999761581421,
1742
+ "rewards/chosen": -0.20598828792572021,
1743
+ "rewards/margins": 0.38752201199531555,
1744
+ "rewards/rejected": -0.5935102701187134,
1745
+ "step": 1230
1746
+ },
1747
+ {
1748
+ "epoch": 0.66,
1749
+ "learning_rate": 1.5535161049189463e-06,
1750
+ "logits/chosen": -1.5570838451385498,
1751
+ "logits/rejected": -1.1252386569976807,
1752
+ "logps/chosen": -500.2212829589844,
1753
+ "logps/rejected": -786.1821899414062,
1754
+ "loss": 0.1145,
1755
+ "rewards/accuracies": 0.800000011920929,
1756
+ "rewards/chosen": -0.1828458607196808,
1757
+ "rewards/margins": 0.3151377737522125,
1758
+ "rewards/rejected": -0.4979836046695709,
1759
+ "step": 1240
1760
+ },
1761
+ {
1762
+ "epoch": 0.67,
1763
+ "learning_rate": 1.5105922805145356e-06,
1764
+ "logits/chosen": -1.8010812997817993,
1765
+ "logits/rejected": -1.2702046632766724,
1766
+ "logps/chosen": -434.25421142578125,
1767
+ "logps/rejected": -807.052001953125,
1768
+ "loss": 0.1215,
1769
+ "rewards/accuracies": 0.8500000238418579,
1770
+ "rewards/chosen": -0.15841101109981537,
1771
+ "rewards/margins": 0.35736268758773804,
1772
+ "rewards/rejected": -0.5157736539840698,
1773
+ "step": 1250
1774
+ },
1775
+ {
1776
+ "epoch": 0.67,
1777
+ "learning_rate": 1.4680115652789823e-06,
1778
+ "logits/chosen": -1.856612205505371,
1779
+ "logits/rejected": -1.147216558456421,
1780
+ "logps/chosen": -523.8411865234375,
1781
+ "logps/rejected": -821.1082763671875,
1782
+ "loss": 0.1727,
1783
+ "rewards/accuracies": 0.824999988079071,
1784
+ "rewards/chosen": -0.2023150473833084,
1785
+ "rewards/margins": 0.35308974981307983,
1786
+ "rewards/rejected": -0.555404782295227,
1787
+ "step": 1260
1788
+ },
1789
+ {
1790
+ "epoch": 0.68,
1791
+ "learning_rate": 1.4257887254542767e-06,
1792
+ "logits/chosen": -1.5119379758834839,
1793
+ "logits/rejected": -1.0702050924301147,
1794
+ "logps/chosen": -511.7137756347656,
1795
+ "logps/rejected": -906.3107299804688,
1796
+ "loss": 0.1025,
1797
+ "rewards/accuracies": 0.8999999761581421,
1798
+ "rewards/chosen": -0.21919742226600647,
1799
+ "rewards/margins": 0.3765312731266022,
1800
+ "rewards/rejected": -0.5957286953926086,
1801
+ "step": 1270
1802
+ },
1803
+ {
1804
+ "epoch": 0.68,
1805
+ "learning_rate": 1.3839384031775227e-06,
1806
+ "logits/chosen": -1.6945511102676392,
1807
+ "logits/rejected": -0.8750427961349487,
1808
+ "logps/chosen": -440.59552001953125,
1809
+ "logps/rejected": -766.9216918945312,
1810
+ "loss": 0.1519,
1811
+ "rewards/accuracies": 0.824999988079071,
1812
+ "rewards/chosen": -0.15435707569122314,
1813
+ "rewards/margins": 0.4113141894340515,
1814
+ "rewards/rejected": -0.5656712651252747,
1815
+ "step": 1280
1816
+ },
1817
+ {
1818
+ "epoch": 0.69,
1819
+ "learning_rate": 1.342475111403298e-06,
1820
+ "logits/chosen": -1.4833415746688843,
1821
+ "logits/rejected": -1.0713919401168823,
1822
+ "logps/chosen": -438.8766174316406,
1823
+ "logps/rejected": -720.0028076171875,
1824
+ "loss": 0.1574,
1825
+ "rewards/accuracies": 0.7250000238418579,
1826
+ "rewards/chosen": -0.20507605373859406,
1827
+ "rewards/margins": 0.28728824853897095,
1828
+ "rewards/rejected": -0.4923642575740814,
1829
+ "step": 1290
1830
+ },
1831
+ {
1832
+ "epoch": 0.69,
1833
+ "learning_rate": 1.3014132288708209e-06,
1834
+ "logits/chosen": -1.5766406059265137,
1835
+ "logits/rejected": -1.0825704336166382,
1836
+ "logps/chosen": -438.3309020996094,
1837
+ "logps/rejected": -823.6751708984375,
1838
+ "loss": 0.166,
1839
+ "rewards/accuracies": 0.875,
1840
+ "rewards/chosen": -0.19768479466438293,
1841
+ "rewards/margins": 0.3687485158443451,
1842
+ "rewards/rejected": -0.566433310508728,
1843
+ "step": 1300
1844
+ },
1845
+ {
1846
+ "epoch": 0.7,
1847
+ "learning_rate": 1.2607669951176549e-06,
1848
+ "logits/chosen": -1.4940482378005981,
1849
+ "logits/rejected": -1.2070845365524292,
1850
+ "logps/chosen": -389.45343017578125,
1851
+ "logps/rejected": -773.9241333007812,
1852
+ "loss": 0.1574,
1853
+ "rewards/accuracies": 0.824999988079071,
1854
+ "rewards/chosen": -0.15667062997817993,
1855
+ "rewards/margins": 0.3334207832813263,
1856
+ "rewards/rejected": -0.49009138345718384,
1857
+ "step": 1310
1858
+ },
1859
+ {
1860
+ "epoch": 0.7,
1861
+ "learning_rate": 1.2205505055416891e-06,
1862
+ "logits/chosen": -1.5122394561767578,
1863
+ "logits/rejected": -1.3955858945846558,
1864
+ "logps/chosen": -338.9855651855469,
1865
+ "logps/rejected": -748.5198364257812,
1866
+ "loss": 0.1404,
1867
+ "rewards/accuracies": 0.875,
1868
+ "rewards/chosen": -0.16774006187915802,
1869
+ "rewards/margins": 0.34086841344833374,
1870
+ "rewards/rejected": -0.5086084604263306,
1871
+ "step": 1320
1872
+ },
1873
+ {
1874
+ "epoch": 0.71,
1875
+ "learning_rate": 1.1807777065131002e-06,
1876
+ "logits/chosen": -1.5142749547958374,
1877
+ "logits/rejected": -1.0132977962493896,
1878
+ "logps/chosen": -410.44879150390625,
1879
+ "logps/rejected": -810.9103393554688,
1880
+ "loss": 0.1108,
1881
+ "rewards/accuracies": 0.925000011920929,
1882
+ "rewards/chosen": -0.14515772461891174,
1883
+ "rewards/margins": 0.36618533730506897,
1884
+ "rewards/rejected": -0.5113429427146912,
1885
+ "step": 1330
1886
+ },
1887
+ {
1888
+ "epoch": 0.71,
1889
+ "learning_rate": 1.1414623905380012e-06,
1890
+ "logits/chosen": -1.756066083908081,
1891
+ "logits/rejected": -1.1571279764175415,
1892
+ "logps/chosen": -441.978515625,
1893
+ "logps/rejected": -786.6061401367188,
1894
+ "loss": 0.1217,
1895
+ "rewards/accuracies": 0.824999988079071,
1896
+ "rewards/chosen": -0.1463043987751007,
1897
+ "rewards/margins": 0.36096060276031494,
1898
+ "rewards/rejected": -0.5072649717330933,
1899
+ "step": 1340
1900
+ },
1901
+ {
1902
+ "epoch": 0.72,
1903
+ "learning_rate": 1.1026181914754388e-06,
1904
+ "logits/chosen": -1.784054160118103,
1905
+ "logits/rejected": -1.0276035070419312,
1906
+ "logps/chosen": -506.1011657714844,
1907
+ "logps/rejected": -819.1619873046875,
1908
+ "loss": 0.1352,
1909
+ "rewards/accuracies": 0.875,
1910
+ "rewards/chosen": -0.1743244081735611,
1911
+ "rewards/margins": 0.38723859190940857,
1912
+ "rewards/rejected": -0.5615630149841309,
1913
+ "step": 1350
1914
+ },
1915
+ {
1916
+ "epoch": 0.73,
1917
+ "learning_rate": 1.0642585798094136e-06,
1918
+ "logits/chosen": -1.5410611629486084,
1919
+ "logits/rejected": -1.0178577899932861,
1920
+ "logps/chosen": -377.84197998046875,
1921
+ "logps/rejected": -720.7569580078125,
1922
+ "loss": 0.1264,
1923
+ "rewards/accuracies": 0.7749999761581421,
1924
+ "rewards/chosen": -0.12278805673122406,
1925
+ "rewards/margins": 0.38729211688041687,
1926
+ "rewards/rejected": -0.5100802183151245,
1927
+ "step": 1360
1928
+ },
1929
+ {
1930
+ "epoch": 0.73,
1931
+ "learning_rate": 1.0263968579775522e-06,
1932
+ "logits/chosen": -1.5256543159484863,
1933
+ "logits/rejected": -0.9656683802604675,
1934
+ "logps/chosen": -458.48089599609375,
1935
+ "logps/rejected": -791.9251708984375,
1936
+ "loss": 0.1401,
1937
+ "rewards/accuracies": 0.824999988079071,
1938
+ "rewards/chosen": -0.16513575613498688,
1939
+ "rewards/margins": 0.38074809312820435,
1940
+ "rewards/rejected": -0.54588383436203,
1941
+ "step": 1370
1942
+ },
1943
+ {
1944
+ "epoch": 0.74,
1945
+ "learning_rate": 9.89046155758058e-07,
1946
+ "logits/chosen": -1.6825745105743408,
1947
+ "logits/rejected": -0.8826824426651001,
1948
+ "logps/chosen": -455.65594482421875,
1949
+ "logps/rejected": -802.0789794921875,
1950
+ "loss": 0.1228,
1951
+ "rewards/accuracies": 0.925000011920929,
1952
+ "rewards/chosen": -0.1435212790966034,
1953
+ "rewards/margins": 0.41453132033348083,
1954
+ "rewards/rejected": -0.558052659034729,
1955
+ "step": 1380
1956
+ },
1957
+ {
1958
+ "epoch": 0.74,
1959
+ "learning_rate": 9.52219425716534e-07,
1960
+ "logits/chosen": -1.4951298236846924,
1961
+ "logits/rejected": -0.8258262872695923,
1962
+ "logps/chosen": -515.0365600585938,
1963
+ "logps/rejected": -771.9305419921875,
1964
+ "loss": 0.1429,
1965
+ "rewards/accuracies": 0.824999988079071,
1966
+ "rewards/chosen": -0.22707219421863556,
1967
+ "rewards/margins": 0.3406526446342468,
1968
+ "rewards/rejected": -0.5677248239517212,
1969
+ "step": 1390
1970
+ },
1971
+ {
1972
+ "epoch": 0.75,
1973
+ "learning_rate": 9.15929438714262e-07,
1974
+ "logits/chosen": -1.6602566242218018,
1975
+ "logits/rejected": -0.9937980771064758,
1976
+ "logps/chosen": -368.70684814453125,
1977
+ "logps/rejected": -689.556884765625,
1978
+ "loss": 0.1528,
1979
+ "rewards/accuracies": 0.800000011920929,
1980
+ "rewards/chosen": -0.12013135105371475,
1981
+ "rewards/margins": 0.37646666169166565,
1982
+ "rewards/rejected": -0.4965980052947998,
1983
+ "step": 1400
1984
+ },
1985
+ {
1986
+ "epoch": 0.75,
1987
+ "learning_rate": 8.801887794794911e-07,
1988
+ "logits/chosen": -1.4943420886993408,
1989
+ "logits/rejected": -0.9112469553947449,
1990
+ "logps/chosen": -379.4705810546875,
1991
+ "logps/rejected": -716.5015258789062,
1992
+ "loss": 0.1407,
1993
+ "rewards/accuracies": 0.7749999761581421,
1994
+ "rewards/chosen": -0.13394254446029663,
1995
+ "rewards/margins": 0.3894422650337219,
1996
+ "rewards/rejected": -0.5233848690986633,
1997
+ "step": 1410
1998
+ },
1999
+ {
2000
+ "epoch": 0.76,
2001
+ "learning_rate": 8.450098422432787e-07,
2002
+ "logits/chosen": -1.7622817754745483,
2003
+ "logits/rejected": -0.7207467555999756,
2004
+ "logps/chosen": -537.2728271484375,
2005
+ "logps/rejected": -854.8095703125,
2006
+ "loss": 0.1203,
2007
+ "rewards/accuracies": 0.925000011920929,
2008
+ "rewards/chosen": -0.1893191635608673,
2009
+ "rewards/margins": 0.4274328649044037,
2010
+ "rewards/rejected": -0.616752028465271,
2011
+ "step": 1420
2012
+ },
2013
+ {
2014
+ "epoch": 0.76,
2015
+ "learning_rate": 8.104048264413858e-07,
2016
+ "logits/chosen": -1.5849692821502686,
2017
+ "logits/rejected": -0.9879060983657837,
2018
+ "logps/chosen": -451.66802978515625,
2019
+ "logps/rejected": -812.3735961914062,
2020
+ "loss": 0.1162,
2021
+ "rewards/accuracies": 0.8999999761581421,
2022
+ "rewards/chosen": -0.1757660210132599,
2023
+ "rewards/margins": 0.4036192297935486,
2024
+ "rewards/rejected": -0.5793852806091309,
2025
+ "step": 1430
2026
+ },
2027
+ {
2028
+ "epoch": 0.77,
2029
+ "learning_rate": 7.763857324837321e-07,
2030
+ "logits/chosen": -1.7880465984344482,
2031
+ "logits/rejected": -1.1138683557510376,
2032
+ "logps/chosen": -470.4102478027344,
2033
+ "logps/rejected": -782.1883544921875,
2034
+ "loss": 0.1273,
2035
+ "rewards/accuracies": 0.824999988079071,
2036
+ "rewards/chosen": -0.18462924659252167,
2037
+ "rewards/margins": 0.37033870816230774,
2038
+ "rewards/rejected": -0.5549679398536682,
2039
+ "step": 1440
2040
+ },
2041
+ {
2042
+ "epoch": 0.77,
2043
+ "learning_rate": 7.429643575928605e-07,
2044
+ "logits/chosen": -1.688932180404663,
2045
+ "logits/rejected": -1.1515108346939087,
2046
+ "logps/chosen": -416.93896484375,
2047
+ "logps/rejected": -748.1307373046875,
2048
+ "loss": 0.1284,
2049
+ "rewards/accuracies": 0.7749999761581421,
2050
+ "rewards/chosen": -0.14920882880687714,
2051
+ "rewards/margins": 0.345571368932724,
2052
+ "rewards/rejected": -0.4947802424430847,
2053
+ "step": 1450
2054
+ },
2055
+ {
2056
+ "epoch": 0.78,
2057
+ "learning_rate": 7.101522917128709e-07,
2058
+ "logits/chosen": -1.3505184650421143,
2059
+ "logits/rejected": -0.8502361178398132,
2060
+ "logps/chosen": -453.3301696777344,
2061
+ "logps/rejected": -852.1624755859375,
2062
+ "loss": 0.149,
2063
+ "rewards/accuracies": 0.875,
2064
+ "rewards/chosen": -0.18861651420593262,
2065
+ "rewards/margins": 0.39376121759414673,
2066
+ "rewards/rejected": -0.5823776721954346,
2067
+ "step": 1460
2068
+ },
2069
+ {
2070
+ "epoch": 0.78,
2071
+ "learning_rate": 6.779609134902312e-07,
2072
+ "logits/chosen": -1.4756485223770142,
2073
+ "logits/rejected": -0.8883223533630371,
2074
+ "logps/chosen": -409.55029296875,
2075
+ "logps/rejected": -707.3751831054688,
2076
+ "loss": 0.1383,
2077
+ "rewards/accuracies": 0.7250000238418579,
2078
+ "rewards/chosen": -0.17147330939769745,
2079
+ "rewards/margins": 0.3226475715637207,
2080
+ "rewards/rejected": -0.49412089586257935,
2081
+ "step": 1470
2082
+ },
2083
+ {
2084
+ "epoch": 0.79,
2085
+ "learning_rate": 6.464013863278629e-07,
2086
+ "logits/chosen": -1.593145728111267,
2087
+ "logits/rejected": -0.8717827796936035,
2088
+ "logps/chosen": -429.87725830078125,
2089
+ "logps/rejected": -856.5046997070312,
2090
+ "loss": 0.1131,
2091
+ "rewards/accuracies": 0.925000011920929,
2092
+ "rewards/chosen": -0.12572301924228668,
2093
+ "rewards/margins": 0.46143823862075806,
2094
+ "rewards/rejected": -0.5871611833572388,
2095
+ "step": 1480
2096
+ },
2097
+ {
2098
+ "epoch": 0.79,
2099
+ "learning_rate": 6.154846545138696e-07,
2100
+ "logits/chosen": -1.556706190109253,
2101
+ "logits/rejected": -1.1209014654159546,
2102
+ "logps/chosen": -434.39813232421875,
2103
+ "logps/rejected": -873.4528198242188,
2104
+ "loss": 0.1184,
2105
+ "rewards/accuracies": 0.949999988079071,
2106
+ "rewards/chosen": -0.151644766330719,
2107
+ "rewards/margins": 0.44497567415237427,
2108
+ "rewards/rejected": -0.5966204404830933,
2109
+ "step": 1490
2110
+ },
2111
+ {
2112
+ "epoch": 0.8,
2113
+ "learning_rate": 5.852214394262515e-07,
2114
+ "logits/chosen": -1.5190951824188232,
2115
+ "logits/rejected": -1.1570379734039307,
2116
+ "logps/chosen": -394.3932189941406,
2117
+ "logps/rejected": -781.7257690429688,
2118
+ "loss": 0.1364,
2119
+ "rewards/accuracies": 0.8999999761581421,
2120
+ "rewards/chosen": -0.16030506789684296,
2121
+ "rewards/margins": 0.3787681758403778,
2122
+ "rewards/rejected": -0.5390732884407043,
2123
+ "step": 1500
2124
+ },
2125
+ {
2126
+ "epoch": 0.81,
2127
+ "learning_rate": 5.556222358149191e-07,
2128
+ "logits/chosen": -1.5962765216827393,
2129
+ "logits/rejected": -0.9932464361190796,
2130
+ "logps/chosen": -390.4809875488281,
2131
+ "logps/rejected": -714.4888916015625,
2132
+ "loss": 0.1652,
2133
+ "rewards/accuracies": 0.800000011920929,
2134
+ "rewards/chosen": -0.17372211813926697,
2135
+ "rewards/margins": 0.361335813999176,
2136
+ "rewards/rejected": -0.5350579023361206,
2137
+ "step": 1510
2138
+ },
2139
+ {
2140
+ "epoch": 0.81,
2141
+ "learning_rate": 5.266973081622992e-07,
2142
+ "logits/chosen": -1.4811457395553589,
2143
+ "logits/rejected": -1.0426948070526123,
2144
+ "logps/chosen": -451.40069580078125,
2145
+ "logps/rejected": -772.6881713867188,
2146
+ "loss": 0.1611,
2147
+ "rewards/accuracies": 0.7749999761581421,
2148
+ "rewards/chosen": -0.18898162245750427,
2149
+ "rewards/margins": 0.3498608469963074,
2150
+ "rewards/rejected": -0.5388425588607788,
2151
+ "step": 1520
2152
+ },
2153
+ {
2154
+ "epoch": 0.82,
2155
+ "learning_rate": 4.984566871237942e-07,
2156
+ "logits/chosen": -1.4943921566009521,
2157
+ "logits/rejected": -1.000528335571289,
2158
+ "logps/chosen": -393.3673095703125,
2159
+ "logps/rejected": -763.3333129882812,
2160
+ "loss": 0.1387,
2161
+ "rewards/accuracies": 0.875,
2162
+ "rewards/chosen": -0.15240536630153656,
2163
+ "rewards/margins": 0.3935711681842804,
2164
+ "rewards/rejected": -0.5459765195846558,
2165
+ "step": 1530
2166
+ },
2167
+ {
2168
+ "epoch": 0.82,
2169
+ "learning_rate": 4.709101660493251e-07,
2170
+ "logits/chosen": -1.4344061613082886,
2171
+ "logits/rejected": -0.8900424838066101,
2172
+ "logps/chosen": -454.6851501464844,
2173
+ "logps/rejected": -862.0211181640625,
2174
+ "loss": 0.1148,
2175
+ "rewards/accuracies": 0.9750000238418579,
2176
+ "rewards/chosen": -0.2144501656293869,
2177
+ "rewards/margins": 0.39736613631248474,
2178
+ "rewards/rejected": -0.6118162870407104,
2179
+ "step": 1540
2180
+ },
2181
+ {
2182
+ "epoch": 0.83,
2183
+ "learning_rate": 4.440672975871743e-07,
2184
+ "logits/chosen": -1.6005455255508423,
2185
+ "logits/rejected": -1.2345631122589111,
2186
+ "logps/chosen": -454.6659240722656,
2187
+ "logps/rejected": -910.2664794921875,
2188
+ "loss": 0.0851,
2189
+ "rewards/accuracies": 0.9750000238418579,
2190
+ "rewards/chosen": -0.1945376694202423,
2191
+ "rewards/margins": 0.4257555603981018,
2192
+ "rewards/rejected": -0.6202932000160217,
2193
+ "step": 1550
2194
+ },
2195
+ {
2196
+ "epoch": 0.83,
2197
+ "learning_rate": 4.1793739037129134e-07,
2198
+ "logits/chosen": -1.755613923072815,
2199
+ "logits/rejected": -0.9976798892021179,
2200
+ "logps/chosen": -426.450927734375,
2201
+ "logps/rejected": -827.8946533203125,
2202
+ "loss": 0.1088,
2203
+ "rewards/accuracies": 0.875,
2204
+ "rewards/chosen": -0.1098506897687912,
2205
+ "rewards/margins": 0.4619103968143463,
2206
+ "rewards/rejected": -0.5717611908912659,
2207
+ "step": 1560
2208
+ },
2209
+ {
2210
+ "epoch": 0.84,
2211
+ "learning_rate": 3.9252950579322405e-07,
2212
+ "logits/chosen": -1.7585302591323853,
2213
+ "logits/rejected": -0.9437012672424316,
2214
+ "logps/chosen": -617.3839111328125,
2215
+ "logps/rejected": -886.2396240234375,
2216
+ "loss": 0.1537,
2217
+ "rewards/accuracies": 0.824999988079071,
2218
+ "rewards/chosen": -0.2242734432220459,
2219
+ "rewards/margins": 0.37233808636665344,
2220
+ "rewards/rejected": -0.5966114401817322,
2221
+ "step": 1570
2222
+ },
2223
+ {
2224
+ "epoch": 0.84,
2225
+ "learning_rate": 3.6785245485978864e-07,
2226
+ "logits/chosen": -1.5823638439178467,
2227
+ "logits/rejected": -1.016841173171997,
2228
+ "logps/chosen": -453.40643310546875,
2229
+ "logps/rejected": -826.7568359375,
2230
+ "loss": 0.1052,
2231
+ "rewards/accuracies": 0.875,
2232
+ "rewards/chosen": -0.14566640555858612,
2233
+ "rewards/margins": 0.41016706824302673,
2234
+ "rewards/rejected": -0.5558334589004517,
2235
+ "step": 1580
2236
+ },
2237
+ {
2238
+ "epoch": 0.85,
2239
+ "learning_rate": 3.43914795137566e-07,
2240
+ "logits/chosen": -1.3402397632598877,
2241
+ "logits/rejected": -0.6611793041229248,
2242
+ "logps/chosen": -491.9454040527344,
2243
+ "logps/rejected": -827.7058715820312,
2244
+ "loss": 0.1243,
2245
+ "rewards/accuracies": 0.824999988079071,
2246
+ "rewards/chosen": -0.1830025464296341,
2247
+ "rewards/margins": 0.3839171230792999,
2248
+ "rewards/rejected": -0.5669196844100952,
2249
+ "step": 1590
2250
+ },
2251
+ {
2252
+ "epoch": 0.85,
2253
+ "learning_rate": 3.207248277852901e-07,
2254
+ "logits/chosen": -1.3319523334503174,
2255
+ "logits/rejected": -1.2467930316925049,
2256
+ "logps/chosen": -415.2613220214844,
2257
+ "logps/rejected": -794.6478271484375,
2258
+ "loss": 0.175,
2259
+ "rewards/accuracies": 0.8500000238418579,
2260
+ "rewards/chosen": -0.18765749037265778,
2261
+ "rewards/margins": 0.33323392271995544,
2262
+ "rewards/rejected": -0.5208913683891296,
2263
+ "step": 1600
2264
+ },
2265
+ {
2266
+ "epoch": 0.86,
2267
+ "learning_rate": 2.9829059467515074e-07,
2268
+ "logits/chosen": -1.6862188577651978,
2269
+ "logits/rejected": -1.0607928037643433,
2270
+ "logps/chosen": -466.0138244628906,
2271
+ "logps/rejected": -823.7083129882812,
2272
+ "loss": 0.1264,
2273
+ "rewards/accuracies": 0.875,
2274
+ "rewards/chosen": -0.1697189062833786,
2275
+ "rewards/margins": 0.39212626218795776,
2276
+ "rewards/rejected": -0.5618451833724976,
2277
+ "step": 1610
2278
+ },
2279
+ {
2280
+ "epoch": 0.86,
2281
+ "learning_rate": 2.766198756040153e-07,
2282
+ "logits/chosen": -1.5529918670654297,
2283
+ "logits/rejected": -1.1102968454360962,
2284
+ "logps/chosen": -514.1727294921875,
2285
+ "logps/rejected": -934.7576293945312,
2286
+ "loss": 0.0946,
2287
+ "rewards/accuracies": 0.925000011920929,
2288
+ "rewards/chosen": -0.2095489799976349,
2289
+ "rewards/margins": 0.39688506722450256,
2290
+ "rewards/rejected": -0.6064340472221375,
2291
+ "step": 1620
2292
+ },
2293
+ {
2294
+ "epoch": 0.87,
2295
+ "learning_rate": 2.5572018559553155e-07,
2296
+ "logits/chosen": -1.4525808095932007,
2297
+ "logits/rejected": -1.114332675933838,
2298
+ "logps/chosen": -429.51336669921875,
2299
+ "logps/rejected": -814.693115234375,
2300
+ "loss": 0.1319,
2301
+ "rewards/accuracies": 0.800000011920929,
2302
+ "rewards/chosen": -0.20347242057323456,
2303
+ "rewards/margins": 0.3701416254043579,
2304
+ "rewards/rejected": -0.5736140012741089,
2305
+ "step": 1630
2306
+ },
2307
+ {
2308
+ "epoch": 0.87,
2309
+ "learning_rate": 2.3559877229404864e-07,
2310
+ "logits/chosen": -1.5984094142913818,
2311
+ "logits/rejected": -1.1003965139389038,
2312
+ "logps/chosen": -458.2529296875,
2313
+ "logps/rejected": -795.9619140625,
2314
+ "loss": 0.1294,
2315
+ "rewards/accuracies": 0.824999988079071,
2316
+ "rewards/chosen": -0.17800959944725037,
2317
+ "rewards/margins": 0.363762229681015,
2318
+ "rewards/rejected": -0.5417717695236206,
2319
+ "step": 1640
2320
+ },
2321
+ {
2322
+ "epoch": 0.88,
2323
+ "learning_rate": 2.1626261345126576e-07,
2324
+ "logits/chosen": -1.4350886344909668,
2325
+ "logits/rejected": -1.0259506702423096,
2326
+ "logps/chosen": -415.7510681152344,
2327
+ "logps/rejected": -919.9736328125,
2328
+ "loss": 0.076,
2329
+ "rewards/accuracies": 0.9750000238418579,
2330
+ "rewards/chosen": -0.15845449268817902,
2331
+ "rewards/margins": 0.4928809702396393,
2332
+ "rewards/rejected": -0.6513354182243347,
2333
+ "step": 1650
2334
+ },
2335
+ {
2336
+ "epoch": 0.89,
2337
+ "learning_rate": 1.9771841450646505e-07,
2338
+ "logits/chosen": -1.6430625915527344,
2339
+ "logits/rejected": -0.9447630643844604,
2340
+ "logps/chosen": -506.1864318847656,
2341
+ "logps/rejected": -795.38134765625,
2342
+ "loss": 0.1602,
2343
+ "rewards/accuracies": 0.800000011920929,
2344
+ "rewards/chosen": -0.23687663674354553,
2345
+ "rewards/margins": 0.33186858892440796,
2346
+ "rewards/rejected": -0.5687452554702759,
2347
+ "step": 1660
2348
+ },
2349
+ {
2350
+ "epoch": 0.89,
2351
+ "learning_rate": 1.7997260626118758e-07,
2352
+ "logits/chosen": -1.898046851158142,
2353
+ "logits/rejected": -1.3102858066558838,
2354
+ "logps/chosen": -514.0572509765625,
2355
+ "logps/rejected": -825.0703125,
2356
+ "loss": 0.1456,
2357
+ "rewards/accuracies": 0.800000011920929,
2358
+ "rewards/chosen": -0.17340177297592163,
2359
+ "rewards/margins": 0.38971638679504395,
2360
+ "rewards/rejected": -0.5631181597709656,
2361
+ "step": 1670
2362
+ },
2363
+ {
2364
+ "epoch": 0.9,
2365
+ "learning_rate": 1.6303134264914365e-07,
2366
+ "logits/chosen": -1.6851441860198975,
2367
+ "logits/rejected": -1.0963430404663086,
2368
+ "logps/chosen": -480.8072204589844,
2369
+ "logps/rejected": -728.0396728515625,
2370
+ "loss": 0.1273,
2371
+ "rewards/accuracies": 0.824999988079071,
2372
+ "rewards/chosen": -0.1836322844028473,
2373
+ "rewards/margins": 0.3386848270893097,
2374
+ "rewards/rejected": -0.5223170518875122,
2375
+ "step": 1680
2376
+ },
2377
+ {
2378
+ "epoch": 0.9,
2379
+ "learning_rate": 1.469004986021355e-07,
2380
+ "logits/chosen": -1.414111852645874,
2381
+ "logits/rejected": -0.8712374567985535,
2382
+ "logps/chosen": -443.46728515625,
2383
+ "logps/rejected": -897.1246948242188,
2384
+ "loss": 0.0899,
2385
+ "rewards/accuracies": 0.925000011920929,
2386
+ "rewards/chosen": -0.17597445845603943,
2387
+ "rewards/margins": 0.4499644339084625,
2388
+ "rewards/rejected": -0.6259388327598572,
2389
+ "step": 1690
2390
+ },
2391
+ {
2392
+ "epoch": 0.91,
2393
+ "learning_rate": 1.315856680127367e-07,
2394
+ "logits/chosen": -1.4355229139328003,
2395
+ "logits/rejected": -0.8268268704414368,
2396
+ "logps/chosen": -411.8287658691406,
2397
+ "logps/rejected": -796.5527954101562,
2398
+ "loss": 0.1013,
2399
+ "rewards/accuracies": 0.875,
2400
+ "rewards/chosen": -0.16030281782150269,
2401
+ "rewards/margins": 0.4271472990512848,
2402
+ "rewards/rejected": -0.5874501466751099,
2403
+ "step": 1700
2404
+ },
2405
+ {
2406
+ "epoch": 0.91,
2407
+ "learning_rate": 1.1709216179442817e-07,
2408
+ "logits/chosen": -1.5936983823776245,
2409
+ "logits/rejected": -0.9012172818183899,
2410
+ "logps/chosen": -452.53155517578125,
2411
+ "logps/rejected": -878.1297607421875,
2412
+ "loss": 0.1108,
2413
+ "rewards/accuracies": 0.925000011920929,
2414
+ "rewards/chosen": -0.17777523398399353,
2415
+ "rewards/margins": 0.4133872389793396,
2416
+ "rewards/rejected": -0.5911625623703003,
2417
+ "step": 1710
2418
+ },
2419
+ {
2420
+ "epoch": 0.92,
2421
+ "learning_rate": 1.0342500603986421e-07,
2422
+ "logits/chosen": -1.441282033920288,
2423
+ "logits/rejected": -0.9638457298278809,
2424
+ "logps/chosen": -417.11895751953125,
2425
+ "logps/rejected": -741.5521850585938,
2426
+ "loss": 0.1466,
2427
+ "rewards/accuracies": 0.75,
2428
+ "rewards/chosen": -0.16872674226760864,
2429
+ "rewards/margins": 0.3362739682197571,
2430
+ "rewards/rejected": -0.5050007104873657,
2431
+ "step": 1720
2432
+ },
2433
+ {
2434
+ "epoch": 0.92,
2435
+ "learning_rate": 9.058894027791643e-08,
2436
+ "logits/chosen": -1.4651381969451904,
2437
+ "logits/rejected": -0.9410767555236816,
2438
+ "logps/chosen": -497.6310119628906,
2439
+ "logps/rejected": -866.1295776367188,
2440
+ "loss": 0.1057,
2441
+ "rewards/accuracies": 0.8999999761581421,
2442
+ "rewards/chosen": -0.23562383651733398,
2443
+ "rewards/margins": 0.3830471634864807,
2444
+ "rewards/rejected": -0.6186710596084595,
2445
+ "step": 1730
2446
+ },
2447
+ {
2448
+ "epoch": 0.93,
2449
+ "learning_rate": 7.858841583008592e-08,
2450
+ "logits/chosen": -1.6138349771499634,
2451
+ "logits/rejected": -1.0234501361846924,
2452
+ "logps/chosen": -425.06610107421875,
2453
+ "logps/rejected": -700.060791015625,
2454
+ "loss": 0.1241,
2455
+ "rewards/accuracies": 0.7749999761581421,
2456
+ "rewards/chosen": -0.16232439875602722,
2457
+ "rewards/margins": 0.3415161669254303,
2458
+ "rewards/rejected": -0.5038405656814575,
2459
+ "step": 1740
2460
+ },
2461
+ {
2462
+ "epoch": 0.93,
2463
+ "learning_rate": 6.742759426686313e-08,
2464
+ "logits/chosen": -1.5296719074249268,
2465
+ "logits/rejected": -1.15841543674469,
2466
+ "logps/chosen": -541.86083984375,
2467
+ "logps/rejected": -857.0759887695312,
2468
+ "loss": 0.1324,
2469
+ "rewards/accuracies": 0.949999988079071,
2470
+ "rewards/chosen": -0.22835755348205566,
2471
+ "rewards/margins": 0.37031129002571106,
2472
+ "rewards/rejected": -0.5986688733100891,
2473
+ "step": 1750
2474
+ },
2475
+ {
2476
+ "epoch": 0.94,
2477
+ "learning_rate": 5.7110345964571104e-08,
2478
+ "logits/chosen": -1.6711105108261108,
2479
+ "logits/rejected": -1.0233453512191772,
2480
+ "logps/chosen": -445.75762939453125,
2481
+ "logps/rejected": -772.8753662109375,
2482
+ "loss": 0.1172,
2483
+ "rewards/accuracies": 0.8500000238418579,
2484
+ "rewards/chosen": -0.18701913952827454,
2485
+ "rewards/margins": 0.3782210052013397,
2486
+ "rewards/rejected": -0.565240204334259,
2487
+ "step": 1760
2488
+ },
2489
+ {
2490
+ "epoch": 0.94,
2491
+ "learning_rate": 4.764024876318357e-08,
2492
+ "logits/chosen": -1.5489776134490967,
2493
+ "logits/rejected": -0.8348779678344727,
2494
+ "logps/chosen": -509.6427307128906,
2495
+ "logps/rejected": -782.4371337890625,
2496
+ "loss": 0.1146,
2497
+ "rewards/accuracies": 0.8500000238418579,
2498
+ "rewards/chosen": -0.20378637313842773,
2499
+ "rewards/margins": 0.37221604585647583,
2500
+ "rewards/rejected": -0.5760024189949036,
2501
+ "step": 1770
2502
+ },
2503
+ {
2504
+ "epoch": 0.95,
2505
+ "learning_rate": 3.902058672559633e-08,
2506
+ "logits/chosen": -1.8395429849624634,
2507
+ "logits/rejected": -1.2655082941055298,
2508
+ "logps/chosen": -375.2162780761719,
2509
+ "logps/rejected": -805.35302734375,
2510
+ "loss": 0.1244,
2511
+ "rewards/accuracies": 0.8500000238418579,
2512
+ "rewards/chosen": -0.11881232261657715,
2513
+ "rewards/margins": 0.4330004155635834,
2514
+ "rewards/rejected": -0.5518127679824829,
2515
+ "step": 1780
2516
+ },
2517
+ {
2518
+ "epoch": 0.95,
2519
+ "learning_rate": 3.125434899876933e-08,
2520
+ "logits/chosen": -1.5633362531661987,
2521
+ "logits/rejected": -1.1406381130218506,
2522
+ "logps/chosen": -356.45098876953125,
2523
+ "logps/rejected": -783.3472900390625,
2524
+ "loss": 0.1001,
2525
+ "rewards/accuracies": 0.824999988079071,
2526
+ "rewards/chosen": -0.1347019374370575,
2527
+ "rewards/margins": 0.42905181646347046,
2528
+ "rewards/rejected": -0.5637537837028503,
2529
+ "step": 1790
2530
+ },
2531
+ {
2532
+ "epoch": 0.96,
2533
+ "learning_rate": 2.4344228777145873e-08,
2534
+ "logits/chosen": -1.6571776866912842,
2535
+ "logits/rejected": -0.7649690508842468,
2536
+ "logps/chosen": -587.1907348632812,
2537
+ "logps/rejected": -933.9886474609375,
2538
+ "loss": 0.132,
2539
+ "rewards/accuracies": 0.875,
2540
+ "rewards/chosen": -0.25137990713119507,
2541
+ "rewards/margins": 0.42299261689186096,
2542
+ "rewards/rejected": -0.6743724942207336,
2543
+ "step": 1800
2544
+ },
2545
+ {
2546
+ "epoch": 0.97,
2547
+ "learning_rate": 1.829262236869772e-08,
2548
+ "logits/chosen": -1.541998267173767,
2549
+ "logits/rejected": -0.8689600229263306,
2550
+ "logps/chosen": -483.3575134277344,
2551
+ "logps/rejected": -698.212158203125,
2552
+ "loss": 0.1768,
2553
+ "rewards/accuracies": 0.7749999761581421,
2554
+ "rewards/chosen": -0.22255222499370575,
2555
+ "rewards/margins": 0.27523303031921387,
2556
+ "rewards/rejected": -0.4977852404117584,
2557
+ "step": 1810
2558
+ },
2559
+ {
2560
+ "epoch": 0.97,
2561
+ "learning_rate": 1.3101628363929586e-08,
2562
+ "logits/chosen": -1.5238444805145264,
2563
+ "logits/rejected": -0.7508775591850281,
2564
+ "logps/chosen": -520.79296875,
2565
+ "logps/rejected": -767.9632568359375,
2566
+ "loss": 0.1203,
2567
+ "rewards/accuracies": 0.875,
2568
+ "rewards/chosen": -0.1896631270647049,
2569
+ "rewards/margins": 0.36830946803092957,
2570
+ "rewards/rejected": -0.5579725503921509,
2571
+ "step": 1820
2572
+ },
2573
+ {
2574
+ "epoch": 0.98,
2575
+ "learning_rate": 8.773046908123195e-09,
2576
+ "logits/chosen": -1.6025253534317017,
2577
+ "logits/rejected": -1.304527997970581,
2578
+ "logps/chosen": -375.14874267578125,
2579
+ "logps/rejected": -767.8821411132812,
2580
+ "loss": 0.1284,
2581
+ "rewards/accuracies": 0.824999988079071,
2582
+ "rewards/chosen": -0.1781584918498993,
2583
+ "rewards/margins": 0.33938026428222656,
2584
+ "rewards/rejected": -0.5175387263298035,
2585
+ "step": 1830
2586
+ },
2587
+ {
2588
+ "epoch": 0.98,
2589
+ "learning_rate": 5.308379077080817e-09,
2590
+ "logits/chosen": -1.6030333042144775,
2591
+ "logits/rejected": -1.3066356182098389,
2592
+ "logps/chosen": -397.94873046875,
2593
+ "logps/rejected": -825.3441162109375,
2594
+ "loss": 0.1111,
2595
+ "rewards/accuracies": 0.875,
2596
+ "rewards/chosen": -0.20664629340171814,
2597
+ "rewards/margins": 0.3704259991645813,
2598
+ "rewards/rejected": -0.5770723819732666,
2599
+ "step": 1840
2600
+ },
2601
+ {
2602
+ "epoch": 0.99,
2603
+ "learning_rate": 2.7088263565760996e-09,
2604
+ "logits/chosen": -1.6151325702667236,
2605
+ "logits/rejected": -0.9792189598083496,
2606
+ "logps/chosen": -399.3708801269531,
2607
+ "logps/rejected": -748.7066650390625,
2608
+ "loss": 0.1181,
2609
+ "rewards/accuracies": 0.875,
2610
+ "rewards/chosen": -0.1316554844379425,
2611
+ "rewards/margins": 0.42211928963661194,
2612
+ "rewards/rejected": -0.5537747740745544,
2613
+ "step": 1850
2614
+ },
2615
+ {
2616
+ "epoch": 0.99,
2617
+ "learning_rate": 9.752902257023633e-10,
2618
+ "logits/chosen": -1.6095302104949951,
2619
+ "logits/rejected": -1.1830781698226929,
2620
+ "logps/chosen": -393.78350830078125,
2621
+ "logps/rejected": -796.3955078125,
2622
+ "loss": 0.0928,
2623
+ "rewards/accuracies": 0.875,
2624
+ "rewards/chosen": -0.15096323192119598,
2625
+ "rewards/margins": 0.4236125349998474,
2626
+ "rewards/rejected": -0.5745757818222046,
2627
+ "step": 1860
2628
+ },
2629
+ {
2630
+ "epoch": 1.0,
2631
+ "learning_rate": 1.083718442532189e-10,
2632
+ "logits/chosen": -1.4612399339675903,
2633
+ "logits/rejected": -0.8474820256233215,
2634
+ "logps/chosen": -456.6351623535156,
2635
+ "logps/rejected": -785.6075439453125,
2636
+ "loss": 0.1354,
2637
+ "rewards/accuracies": 0.800000011920929,
2638
+ "rewards/chosen": -0.18163737654685974,
2639
+ "rewards/margins": 0.3639640808105469,
2640
+ "rewards/rejected": -0.545601487159729,
2641
+ "step": 1870
2642
+ },
2643
+ {
2644
+ "epoch": 1.0,
2645
+ "step": 1875,
2646
+ "total_flos": 0.0,
2647
+ "train_loss": 0.13990657812754312,
2648
+ "train_runtime": 16010.7596,
2649
+ "train_samples_per_second": 0.937,
2650
+ "train_steps_per_second": 0.117
2651
+ }
2652
+ ],
2653
+ "logging_steps": 10,
2654
+ "max_steps": 1875,
2655
+ "num_input_tokens_seen": 0,
2656
+ "num_train_epochs": 1,
2657
+ "save_steps": 100,
2658
+ "total_flos": 0.0,
2659
+ "train_batch_size": 2,
2660
+ "trial_name": null,
2661
+ "trial_params": null
2662
+ }