Model save
Browse files- README.md +20 -19
- adapter_model.safetensors +1 -1
- all_results.json +16 -16
- eval_results.json +12 -12
- runs/Jan23_20-43-15_dpo-tests-2-85c5c9854f-p9wdp/events.out.tfevents.1706043566.dpo-tests-2-85c5c9854f-p9wdp.12217.0 +2 -2
- runs/Jan23_20-43-15_dpo-tests-2-85c5c9854f-p9wdp/events.out.tfevents.1706056801.dpo-tests-2-85c5c9854f-p9wdp.12217.1 +3 -0
- train_results.json +4 -4
- trainer_state.json +1049 -969
README.md
CHANGED
|
@@ -2,13 +2,9 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
| 5 |
-
- alignment-handbook
|
| 6 |
-
- generated_from_trainer
|
| 7 |
- trl
|
| 8 |
- dpo
|
| 9 |
- generated_from_trainer
|
| 10 |
-
datasets:
|
| 11 |
-
- sablo/HelpSteer_binarized
|
| 12 |
base_model: sablo/sablo-pebble-mistral
|
| 13 |
model-index:
|
| 14 |
- name: sablo-pebble-mistral-dpo-lora-HelpSteer_binarized
|
|
@@ -20,17 +16,17 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 20 |
|
| 21 |
# sablo-pebble-mistral-dpo-lora-HelpSteer_binarized
|
| 22 |
|
| 23 |
-
This model is a fine-tuned version of [sablo/sablo-pebble-mistral](https://huggingface.co/sablo/sablo-pebble-mistral) on the
|
| 24 |
It achieves the following results on the evaluation set:
|
| 25 |
-
- Loss: 0.
|
| 26 |
-
- Rewards/chosen: -
|
| 27 |
-
- Rewards/rejected: -
|
| 28 |
-
- Rewards/accuracies: 0.
|
| 29 |
-
- Rewards/margins: 0.
|
| 30 |
-
- Logps/rejected: -
|
| 31 |
-
- Logps/chosen: -
|
| 32 |
-
- Logits/rejected: -2.
|
| 33 |
-
- Logits/chosen: -2.
|
| 34 |
|
| 35 |
## Model description
|
| 36 |
|
|
@@ -65,11 +61,16 @@ The following hyperparameters were used during training:
|
|
| 65 |
|
| 66 |
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 67 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 68 |
-
| 0.
|
| 69 |
-
| 0.
|
| 70 |
-
| 0.
|
| 71 |
-
| 0.
|
| 72 |
-
| 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
|
| 74 |
|
| 75 |
### Framework versions
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- trl
|
| 6 |
- dpo
|
| 7 |
- generated_from_trainer
|
|
|
|
|
|
|
| 8 |
base_model: sablo/sablo-pebble-mistral
|
| 9 |
model-index:
|
| 10 |
- name: sablo-pebble-mistral-dpo-lora-HelpSteer_binarized
|
|
|
|
| 16 |
|
| 17 |
# sablo-pebble-mistral-dpo-lora-HelpSteer_binarized
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [sablo/sablo-pebble-mistral](https://huggingface.co/sablo/sablo-pebble-mistral) on the None dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.5371
|
| 22 |
+
- Rewards/chosen: -0.9335
|
| 23 |
+
- Rewards/rejected: -1.6455
|
| 24 |
+
- Rewards/accuracies: 0.7264
|
| 25 |
+
- Rewards/margins: 0.7121
|
| 26 |
+
- Logps/rejected: -298.0735
|
| 27 |
+
- Logps/chosen: -253.4149
|
| 28 |
+
- Logits/rejected: -2.4554
|
| 29 |
+
- Logits/chosen: -2.5093
|
| 30 |
|
| 31 |
## Model description
|
| 32 |
|
|
|
|
| 61 |
|
| 62 |
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 63 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 64 |
+
| 0.6874 | 0.1 | 100 | 0.6892 | 0.0213 | 0.0133 | 0.6698 | 0.0080 | -132.1924 | -157.9395 | -2.4463 | -2.4843 |
|
| 65 |
+
| 0.6592 | 0.2 | 200 | 0.6594 | 0.0055 | -0.0704 | 0.6698 | 0.0759 | -140.5588 | -159.5180 | -2.4922 | -2.5370 |
|
| 66 |
+
| 0.5451 | 0.3 | 300 | 0.5867 | -0.4490 | -0.7587 | 0.6863 | 0.3097 | -209.3938 | -204.9713 | -2.5128 | -2.5620 |
|
| 67 |
+
| 0.4933 | 0.39 | 400 | 0.5591 | -0.6060 | -1.1029 | 0.7146 | 0.4968 | -243.8062 | -220.6713 | -2.4868 | -2.5386 |
|
| 68 |
+
| 0.5271 | 0.49 | 500 | 0.5488 | -0.6712 | -1.2738 | 0.7193 | 0.6026 | -260.8958 | -227.1889 | -2.4784 | -2.5312 |
|
| 69 |
+
| 0.4594 | 0.59 | 600 | 0.5418 | -0.7977 | -1.4672 | 0.7311 | 0.6695 | -280.2420 | -239.8430 | -2.4672 | -2.5200 |
|
| 70 |
+
| 0.5444 | 0.69 | 700 | 0.5358 | -0.7688 | -1.4528 | 0.7335 | 0.6840 | -278.8014 | -236.9531 | -2.4594 | -2.5127 |
|
| 71 |
+
| 0.5755 | 0.79 | 800 | 0.5405 | -1.0672 | -1.7631 | 0.7311 | 0.6959 | -309.8293 | -266.7906 | -2.4585 | -2.5118 |
|
| 72 |
+
| 0.5495 | 0.89 | 900 | 0.5371 | -0.9321 | -1.6450 | 0.7288 | 0.7129 | -298.0242 | -253.2804 | -2.4558 | -2.5096 |
|
| 73 |
+
| 0.5948 | 0.98 | 1000 | 0.5371 | -0.9335 | -1.6455 | 0.7264 | 0.7121 | -298.0735 | -253.4149 | -2.4554 | -2.5093 |
|
| 74 |
|
| 75 |
|
| 76 |
### Framework versions
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 83945744
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0aab87c6f6a113f5aa86d5285796151d3bfe9b2bac65489ec2d14f61e072556b
|
| 3 |
size 83945744
|
all_results.json
CHANGED
|
@@ -1,21 +1,21 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_logits/chosen": -2.
|
| 4 |
-
"eval_logits/rejected": -2.
|
| 5 |
-
"eval_logps/chosen": -
|
| 6 |
-
"eval_logps/rejected": -
|
| 7 |
-
"eval_loss": 0.
|
| 8 |
-
"eval_rewards/accuracies": 0.
|
| 9 |
-
"eval_rewards/chosen": -
|
| 10 |
-
"eval_rewards/margins": 0.
|
| 11 |
-
"eval_rewards/rejected": -
|
| 12 |
-
"eval_runtime":
|
| 13 |
"eval_samples": 418,
|
| 14 |
-
"eval_samples_per_second":
|
| 15 |
-
"eval_steps_per_second": 0.
|
| 16 |
-
"train_loss": 0.
|
| 17 |
-
"train_runtime":
|
| 18 |
"train_samples": 8130,
|
| 19 |
-
"train_samples_per_second": 0.
|
| 20 |
-
"train_steps_per_second": 0.
|
| 21 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_logits/chosen": -2.5093116760253906,
|
| 4 |
+
"eval_logits/rejected": -2.4554219245910645,
|
| 5 |
+
"eval_logps/chosen": -253.41494750976562,
|
| 6 |
+
"eval_logps/rejected": -298.073486328125,
|
| 7 |
+
"eval_loss": 0.5370525121688843,
|
| 8 |
+
"eval_rewards/accuracies": 0.7264150977134705,
|
| 9 |
+
"eval_rewards/chosen": -0.9334509372711182,
|
| 10 |
+
"eval_rewards/margins": 0.7120789289474487,
|
| 11 |
+
"eval_rewards/rejected": -1.6455297470092773,
|
| 12 |
+
"eval_runtime": 287.3881,
|
| 13 |
"eval_samples": 418,
|
| 14 |
+
"eval_samples_per_second": 1.454,
|
| 15 |
+
"eval_steps_per_second": 0.184,
|
| 16 |
+
"train_loss": 0.5672297686573089,
|
| 17 |
+
"train_runtime": 12946.9794,
|
| 18 |
"train_samples": 8130,
|
| 19 |
+
"train_samples_per_second": 0.628,
|
| 20 |
+
"train_steps_per_second": 0.078
|
| 21 |
}
|
eval_results.json
CHANGED
|
@@ -1,16 +1,16 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_logits/chosen": -2.
|
| 4 |
-
"eval_logits/rejected": -2.
|
| 5 |
-
"eval_logps/chosen": -
|
| 6 |
-
"eval_logps/rejected": -
|
| 7 |
-
"eval_loss": 0.
|
| 8 |
-
"eval_rewards/accuracies": 0.
|
| 9 |
-
"eval_rewards/chosen": -
|
| 10 |
-
"eval_rewards/margins": 0.
|
| 11 |
-
"eval_rewards/rejected": -
|
| 12 |
-
"eval_runtime":
|
| 13 |
"eval_samples": 418,
|
| 14 |
-
"eval_samples_per_second":
|
| 15 |
-
"eval_steps_per_second": 0.
|
| 16 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_logits/chosen": -2.5093116760253906,
|
| 4 |
+
"eval_logits/rejected": -2.4554219245910645,
|
| 5 |
+
"eval_logps/chosen": -253.41494750976562,
|
| 6 |
+
"eval_logps/rejected": -298.073486328125,
|
| 7 |
+
"eval_loss": 0.5370525121688843,
|
| 8 |
+
"eval_rewards/accuracies": 0.7264150977134705,
|
| 9 |
+
"eval_rewards/chosen": -0.9334509372711182,
|
| 10 |
+
"eval_rewards/margins": 0.7120789289474487,
|
| 11 |
+
"eval_rewards/rejected": -1.6455297470092773,
|
| 12 |
+
"eval_runtime": 287.3881,
|
| 13 |
"eval_samples": 418,
|
| 14 |
+
"eval_samples_per_second": 1.454,
|
| 15 |
+
"eval_steps_per_second": 0.184
|
| 16 |
}
|
runs/Jan23_20-43-15_dpo-tests-2-85c5c9854f-p9wdp/events.out.tfevents.1706043566.dpo-tests-2-85c5c9854f-p9wdp.12217.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5aeeb87042389af97fa4aa208f10a7104d74e30697eeb0119c27556309334926
|
| 3 |
+
size 76962
|
runs/Jan23_20-43-15_dpo-tests-2-85c5c9854f-p9wdp/events.out.tfevents.1706056801.dpo-tests-2-85c5c9854f-p9wdp.12217.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e5494e9c2e7e625933707008ec53066d79dc966ce77e55f3a8482e6a4cee63f
|
| 3 |
+
size 828
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 8130,
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.5672297686573089,
|
| 4 |
+
"train_runtime": 12946.9794,
|
| 5 |
"train_samples": 8130,
|
| 6 |
+
"train_samples_per_second": 0.628,
|
| 7 |
+
"train_steps_per_second": 0.078
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 0.999508116084604,
|
| 5 |
-
"eval_steps":
|
| 6 |
"global_step": 1016,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
|
@@ -11,10 +11,10 @@
|
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
"learning_rate": 4.901960784313726e-08,
|
| 14 |
-
"logits/chosen": -2.
|
| 15 |
-
"logits/rejected": -2.
|
| 16 |
-
"logps/chosen": -
|
| 17 |
-
"logps/rejected": -
|
| 18 |
"loss": 0.6931,
|
| 19 |
"rewards/accuracies": 0.0,
|
| 20 |
"rewards/chosen": 0.0,
|
|
@@ -25,1505 +25,1585 @@
|
|
| 25 |
{
|
| 26 |
"epoch": 0.01,
|
| 27 |
"learning_rate": 4.901960784313725e-07,
|
| 28 |
-
"logits/chosen": -2.
|
| 29 |
-
"logits/rejected": -2.
|
| 30 |
-
"logps/chosen": -
|
| 31 |
-
"logps/rejected": -
|
| 32 |
-
"loss": 0.
|
| 33 |
-
"rewards/accuracies": 0.
|
| 34 |
-
"rewards/chosen": 0.
|
| 35 |
-
"rewards/margins": 0.
|
| 36 |
-
"rewards/rejected": -
|
| 37 |
"step": 10
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.02,
|
| 41 |
"learning_rate": 9.80392156862745e-07,
|
| 42 |
-
"logits/chosen": -2.
|
| 43 |
-
"logits/rejected": -2.
|
| 44 |
-
"logps/chosen": -
|
| 45 |
-
"logps/rejected": -
|
| 46 |
-
"loss": 0.
|
| 47 |
-
"rewards/accuracies": 0.
|
| 48 |
-
"rewards/chosen": -0.
|
| 49 |
-
"rewards/margins": -0.
|
| 50 |
-
"rewards/rejected":
|
| 51 |
"step": 20
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.03,
|
| 55 |
"learning_rate": 1.4705882352941177e-06,
|
| 56 |
-
"logits/chosen": -2.
|
| 57 |
-
"logits/rejected": -2.
|
| 58 |
-
"logps/chosen": -
|
| 59 |
-
"logps/rejected": -
|
| 60 |
-
"loss": 0.
|
| 61 |
-
"rewards/accuracies": 0.
|
| 62 |
-
"rewards/chosen":
|
| 63 |
-
"rewards/margins": -0.
|
| 64 |
-
"rewards/rejected":
|
| 65 |
"step": 30
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.04,
|
| 69 |
"learning_rate": 1.96078431372549e-06,
|
| 70 |
-
"logits/chosen": -2.
|
| 71 |
-
"logits/rejected": -2.
|
| 72 |
-
"logps/chosen": -
|
| 73 |
-
"logps/rejected": -
|
| 74 |
-
"loss": 0.
|
| 75 |
-
"rewards/accuracies": 0.
|
| 76 |
-
"rewards/chosen": 0.
|
| 77 |
-
"rewards/margins": 0.
|
| 78 |
-
"rewards/rejected":
|
| 79 |
"step": 40
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.05,
|
| 83 |
"learning_rate": 2.450980392156863e-06,
|
| 84 |
-
"logits/chosen": -2.
|
| 85 |
-
"logits/rejected": -2.
|
| 86 |
-
"logps/chosen": -
|
| 87 |
-
"logps/rejected": -
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"rewards/accuracies": 0.
|
| 90 |
-
"rewards/chosen": 0.
|
| 91 |
-
"rewards/margins": 0.
|
| 92 |
-
"rewards/rejected":
|
| 93 |
"step": 50
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.06,
|
| 97 |
"learning_rate": 2.9411764705882355e-06,
|
| 98 |
-
"logits/chosen": -2.
|
| 99 |
-
"logits/rejected": -2.
|
| 100 |
-
"logps/chosen": -
|
| 101 |
-
"logps/rejected": -
|
| 102 |
-
"loss": 0.
|
| 103 |
-
"rewards/accuracies": 0.
|
| 104 |
-
"rewards/chosen": 0.
|
| 105 |
-
"rewards/margins": 0.
|
| 106 |
-
"rewards/rejected": 0.
|
| 107 |
"step": 60
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.07,
|
| 111 |
"learning_rate": 3.431372549019608e-06,
|
| 112 |
-
"logits/chosen": -2.
|
| 113 |
-
"logits/rejected": -2.
|
| 114 |
-
"logps/chosen": -
|
| 115 |
-
"logps/rejected": -
|
| 116 |
-
"loss": 0.
|
| 117 |
-
"rewards/accuracies": 0.
|
| 118 |
-
"rewards/chosen": 0.
|
| 119 |
-
"rewards/margins": 0.
|
| 120 |
-
"rewards/rejected": 0.
|
| 121 |
"step": 70
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.08,
|
| 125 |
"learning_rate": 3.92156862745098e-06,
|
| 126 |
-
"logits/chosen": -2.
|
| 127 |
-
"logits/rejected": -2.
|
| 128 |
-
"logps/chosen": -
|
| 129 |
-
"logps/rejected": -
|
| 130 |
-
"loss": 0.
|
| 131 |
"rewards/accuracies": 0.637499988079071,
|
| 132 |
-
"rewards/chosen": 0.
|
| 133 |
-
"rewards/margins": 0.
|
| 134 |
-
"rewards/rejected": 0.
|
| 135 |
"step": 80
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.09,
|
| 139 |
"learning_rate": 4.411764705882353e-06,
|
| 140 |
-
"logits/chosen": -2.
|
| 141 |
-
"logits/rejected": -
|
| 142 |
-
"logps/chosen": -
|
| 143 |
-
"logps/rejected": -
|
| 144 |
-
"loss": 0.
|
| 145 |
-
"rewards/accuracies": 0.
|
| 146 |
-
"rewards/chosen": 0.
|
| 147 |
-
"rewards/margins": 0.
|
| 148 |
-
"rewards/rejected": 0.
|
| 149 |
"step": 90
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.1,
|
| 153 |
"learning_rate": 4.901960784313726e-06,
|
| 154 |
-
"logits/chosen": -2.
|
| 155 |
-
"logits/rejected": -
|
| 156 |
-
"logps/chosen": -
|
| 157 |
-
"logps/rejected": -
|
| 158 |
-
"loss": 0.
|
| 159 |
-
"rewards/accuracies": 0.
|
| 160 |
-
"rewards/chosen": 0.
|
| 161 |
-
"rewards/margins": 0.
|
| 162 |
-
"rewards/rejected": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 163 |
"step": 100
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.11,
|
| 167 |
"learning_rate": 4.9990549169459415e-06,
|
| 168 |
-
"logits/chosen": -2.
|
| 169 |
-
"logits/rejected": -2.
|
| 170 |
-
"logps/chosen": -
|
| 171 |
-
"logps/rejected": -
|
| 172 |
-
"loss": 0.
|
| 173 |
-
"rewards/accuracies": 0.
|
| 174 |
-
"rewards/chosen": 0.
|
| 175 |
-
"rewards/margins": 0.
|
| 176 |
-
"rewards/rejected": 0.
|
| 177 |
"step": 110
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.12,
|
| 181 |
"learning_rate": 4.995216741642263e-06,
|
| 182 |
-
"logits/chosen": -2.
|
| 183 |
-
"logits/rejected": -2.
|
| 184 |
-
"logps/chosen": -
|
| 185 |
-
"logps/rejected": -
|
| 186 |
-
"loss": 0.
|
| 187 |
-
"rewards/accuracies": 0.
|
| 188 |
-
"rewards/chosen": 0.
|
| 189 |
-
"rewards/margins": 0.
|
| 190 |
-
"rewards/rejected": 0.
|
| 191 |
"step": 120
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.13,
|
| 195 |
"learning_rate": 4.988430936991089e-06,
|
| 196 |
-
"logits/chosen": -2.
|
| 197 |
-
"logits/rejected": -2.
|
| 198 |
-
"logps/chosen": -
|
| 199 |
-
"logps/rejected": -
|
| 200 |
-
"loss": 0.
|
| 201 |
-
"rewards/accuracies": 0.
|
| 202 |
-
"rewards/chosen": 0.
|
| 203 |
-
"rewards/margins": 0.
|
| 204 |
-
"rewards/rejected": 0.
|
| 205 |
"step": 130
|
| 206 |
},
|
| 207 |
{
|
| 208 |
"epoch": 0.14,
|
| 209 |
"learning_rate": 4.978705519144525e-06,
|
| 210 |
-
"logits/chosen": -2.
|
| 211 |
-
"logits/rejected": -2.
|
| 212 |
-
"logps/chosen": -
|
| 213 |
-
"logps/rejected": -
|
| 214 |
-
"loss": 0.
|
| 215 |
-
"rewards/accuracies": 0.
|
| 216 |
-
"rewards/chosen": 0.
|
| 217 |
-
"rewards/margins": 0.
|
| 218 |
-
"rewards/rejected": 0.
|
| 219 |
"step": 140
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.15,
|
| 223 |
"learning_rate": 4.966051976854862e-06,
|
| 224 |
-
"logits/chosen": -2.
|
| 225 |
-
"logits/rejected": -2.
|
| 226 |
-
"logps/chosen": -
|
| 227 |
-
"logps/rejected": -
|
| 228 |
-
"loss": 0.
|
| 229 |
-
"rewards/accuracies": 0.
|
| 230 |
-
"rewards/chosen": 0.
|
| 231 |
-
"rewards/margins": 0.
|
| 232 |
-
"rewards/rejected": 0.
|
| 233 |
"step": 150
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.16,
|
| 237 |
"learning_rate": 4.950485257902782e-06,
|
| 238 |
-
"logits/chosen": -2.
|
| 239 |
-
"logits/rejected": -2.
|
| 240 |
-
"logps/chosen": -
|
| 241 |
-
"logps/rejected": -
|
| 242 |
-
"loss": 0.
|
| 243 |
-
"rewards/accuracies": 0.
|
| 244 |
-
"rewards/chosen": 0.
|
| 245 |
-
"rewards/margins": 0.
|
| 246 |
-
"rewards/rejected": 0.
|
| 247 |
"step": 160
|
| 248 |
},
|
| 249 |
{
|
| 250 |
"epoch": 0.17,
|
| 251 |
"learning_rate": 4.932023751439358e-06,
|
| 252 |
-
"logits/chosen": -2.
|
| 253 |
-
"logits/rejected": -2.
|
| 254 |
-
"logps/chosen": -
|
| 255 |
-
"logps/rejected": -
|
| 256 |
-
"loss": 0.
|
| 257 |
-
"rewards/accuracies": 0.
|
| 258 |
-
"rewards/chosen": 0.
|
| 259 |
-
"rewards/margins": 0.
|
| 260 |
-
"rewards/rejected": 0.
|
| 261 |
"step": 170
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.18,
|
| 265 |
"learning_rate": 4.9106892662627395e-06,
|
| 266 |
-
"logits/chosen": -2.
|
| 267 |
-
"logits/rejected": -2.
|
| 268 |
-
"logps/chosen": -
|
| 269 |
-
"logps/rejected": -
|
| 270 |
-
"loss": 0.
|
| 271 |
-
"rewards/accuracies": 0.
|
| 272 |
-
"rewards/chosen": 0.
|
| 273 |
-
"rewards/margins": 0.
|
| 274 |
-
"rewards/rejected": 0.
|
| 275 |
"step": 180
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 0.19,
|
| 279 |
"learning_rate": 4.886507005055149e-06,
|
| 280 |
-
"logits/chosen": -2.
|
| 281 |
-
"logits/rejected": -2.
|
| 282 |
-
"logps/chosen": -
|
| 283 |
-
"logps/rejected": -
|
| 284 |
-
"loss": 0.
|
| 285 |
-
"rewards/accuracies": 0.
|
| 286 |
-
"rewards/chosen": 0.
|
| 287 |
-
"rewards/margins": 0.
|
| 288 |
-
"rewards/rejected": 0.
|
| 289 |
"step": 190
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 0.2,
|
| 293 |
"learning_rate": 4.859505534610658e-06,
|
| 294 |
-
"logits/chosen": -2.
|
| 295 |
-
"logits/rejected": -2.
|
| 296 |
-
"logps/chosen": -
|
| 297 |
-
"logps/rejected": -
|
| 298 |
-
"loss": 0.
|
| 299 |
-
"rewards/accuracies": 0.
|
| 300 |
-
"rewards/chosen": 0.
|
| 301 |
-
"rewards/margins": 0.
|
| 302 |
-
"rewards/rejected": -0.
|
| 303 |
"step": 200
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.2,
|
| 307 |
-
"eval_logits/chosen": -2.
|
| 308 |
-
"eval_logits/rejected": -2.
|
| 309 |
-
"eval_logps/chosen": -
|
| 310 |
-
"eval_logps/rejected": -
|
| 311 |
-
"eval_loss": 0.
|
| 312 |
-
"eval_rewards/accuracies": 0.
|
| 313 |
-
"eval_rewards/chosen": 0.
|
| 314 |
-
"eval_rewards/margins": 0.
|
| 315 |
-
"eval_rewards/rejected": -0.
|
| 316 |
-
"eval_runtime":
|
| 317 |
-
"eval_samples_per_second": 1.
|
| 318 |
-
"eval_steps_per_second": 0.
|
| 319 |
"step": 200
|
| 320 |
},
|
| 321 |
{
|
| 322 |
"epoch": 0.21,
|
| 323 |
"learning_rate": 4.829716752088893e-06,
|
| 324 |
-
"logits/chosen": -2.
|
| 325 |
-
"logits/rejected": -2.
|
| 326 |
-
"logps/chosen": -
|
| 327 |
-
"logps/rejected": -
|
| 328 |
-
"loss": 0.
|
| 329 |
-
"rewards/accuracies": 0.
|
| 330 |
-
"rewards/chosen": 0.
|
| 331 |
-
"rewards/margins": 0.
|
| 332 |
-
"rewards/rejected": -0.
|
| 333 |
"step": 210
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 0.22,
|
| 337 |
"learning_rate": 4.797175847334535e-06,
|
| 338 |
-
"logits/chosen": -2.
|
| 339 |
-
"logits/rejected": -2.
|
| 340 |
-
"logps/chosen": -
|
| 341 |
-
"logps/rejected": -
|
| 342 |
-
"loss": 0.
|
| 343 |
-
"rewards/accuracies": 0.
|
| 344 |
-
"rewards/chosen": 0.
|
| 345 |
-
"rewards/margins": 0.
|
| 346 |
-
"rewards/rejected": -0.
|
| 347 |
"step": 220
|
| 348 |
},
|
| 349 |
{
|
| 350 |
"epoch": 0.23,
|
| 351 |
"learning_rate": 4.761921261307143e-06,
|
| 352 |
-
"logits/chosen": -2.
|
| 353 |
-
"logits/rejected": -2.
|
| 354 |
-
"logps/chosen": -
|
| 355 |
-
"logps/rejected": -
|
| 356 |
-
"loss": 0.
|
| 357 |
-
"rewards/accuracies": 0.
|
| 358 |
-
"rewards/chosen": -0.
|
| 359 |
-
"rewards/margins": 0.
|
| 360 |
-
"rewards/rejected": -0.
|
| 361 |
"step": 230
|
| 362 |
},
|
| 363 |
{
|
| 364 |
"epoch": 0.24,
|
| 365 |
"learning_rate": 4.723994640670377e-06,
|
| 366 |
-
"logits/chosen": -2.
|
| 367 |
-
"logits/rejected": -2.
|
| 368 |
-
"logps/chosen": -
|
| 369 |
-
"logps/rejected": -
|
| 370 |
-
"loss": 0.
|
| 371 |
-
"rewards/accuracies": 0.
|
| 372 |
-
"rewards/chosen": -0.
|
| 373 |
-
"rewards/margins": 0.
|
| 374 |
-
"rewards/rejected": -0.
|
| 375 |
"step": 240
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 0.25,
|
| 379 |
"learning_rate": 4.68344078859431e-06,
|
| 380 |
-
"logits/chosen": -2.
|
| 381 |
-
"logits/rejected": -2.
|
| 382 |
-
"logps/chosen": -
|
| 383 |
-
"logps/rejected": -
|
| 384 |
-
"loss": 0.
|
| 385 |
-
"rewards/accuracies": 0.
|
| 386 |
-
"rewards/chosen": -0.
|
| 387 |
-
"rewards/margins": 0.
|
| 388 |
-
"rewards/rejected": -0.
|
| 389 |
"step": 250
|
| 390 |
},
|
| 391 |
{
|
| 392 |
"epoch": 0.26,
|
| 393 |
"learning_rate": 4.6403076118289006e-06,
|
| 394 |
-
"logits/chosen": -2.
|
| 395 |
-
"logits/rejected": -2.
|
| 396 |
-
"logps/chosen": -
|
| 397 |
-
"logps/rejected": -
|
| 398 |
-
"loss": 0.
|
| 399 |
-
"rewards/accuracies": 0.
|
| 400 |
-
"rewards/chosen": -0.
|
| 401 |
-
"rewards/margins": 0.
|
| 402 |
-
"rewards/rejected": -0.
|
| 403 |
"step": 260
|
| 404 |
},
|
| 405 |
{
|
| 406 |
"epoch": 0.27,
|
| 407 |
"learning_rate": 4.5946460641111776e-06,
|
| 408 |
-
"logits/chosen": -2.
|
| 409 |
-
"logits/rejected": -2.
|
| 410 |
-
"logps/chosen": -
|
| 411 |
-
"logps/rejected": -
|
| 412 |
-
"loss": 0.
|
| 413 |
-
"rewards/accuracies": 0.
|
| 414 |
-
"rewards/chosen": -0.
|
| 415 |
-
"rewards/margins": 0.
|
| 416 |
-
"rewards/rejected": -0.
|
| 417 |
"step": 270
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 0.28,
|
| 421 |
"learning_rate": 4.546510085972983e-06,
|
| 422 |
-
"logits/chosen": -2.
|
| 423 |
-
"logits/rejected": -2.
|
| 424 |
-
"logps/chosen": -
|
| 425 |
-
"logps/rejected": -
|
| 426 |
-
"loss": 0.
|
| 427 |
-
"rewards/accuracies": 0.
|
| 428 |
-
"rewards/chosen": -0.
|
| 429 |
-
"rewards/margins": 0.
|
| 430 |
-
"rewards/rejected": -0.
|
| 431 |
"step": 280
|
| 432 |
},
|
| 433 |
{
|
| 434 |
"epoch": 0.29,
|
| 435 |
"learning_rate": 4.495956541020376e-06,
|
| 436 |
-
"logits/chosen": -2.
|
| 437 |
-
"logits/rejected": -2.
|
| 438 |
-
"logps/chosen": -
|
| 439 |
-
"logps/rejected": -
|
| 440 |
-
"loss": 0.
|
| 441 |
-
"rewards/accuracies": 0.
|
| 442 |
-
"rewards/chosen": -0.
|
| 443 |
-
"rewards/margins": 0.
|
| 444 |
-
"rewards/rejected": -0.
|
| 445 |
"step": 290
|
| 446 |
},
|
| 447 |
{
|
| 448 |
"epoch": 0.3,
|
| 449 |
"learning_rate": 4.443045148759978e-06,
|
| 450 |
-
"logits/chosen": -2.
|
| 451 |
-
"logits/rejected": -2.
|
| 452 |
-
"logps/chosen": -
|
| 453 |
-
"logps/rejected": -
|
| 454 |
-
"loss": 0.
|
| 455 |
-
"rewards/accuracies": 0.
|
| 456 |
-
"rewards/chosen": -0.
|
| 457 |
-
"rewards/margins": 0.
|
| 458 |
-
"rewards/rejected": -0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 459 |
"step": 300
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 0.3,
|
| 463 |
"learning_rate": 4.3878384140516025e-06,
|
| 464 |
-
"logits/chosen": -2.
|
| 465 |
-
"logits/rejected": -2.
|
| 466 |
-
"logps/chosen": -
|
| 467 |
-
"logps/rejected": -
|
| 468 |
-
"loss": 0.
|
| 469 |
-
"rewards/accuracies": 0.
|
| 470 |
-
"rewards/chosen": -0.
|
| 471 |
-
"rewards/margins": 0.
|
| 472 |
-
"rewards/rejected": -0.
|
| 473 |
"step": 310
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 0.31,
|
| 477 |
"learning_rate": 4.330401553270522e-06,
|
| 478 |
-
"logits/chosen": -2.
|
| 479 |
-
"logits/rejected": -2.
|
| 480 |
-
"logps/chosen": -
|
| 481 |
-
"logps/rejected": -
|
| 482 |
-
"loss": 0.
|
| 483 |
-
"rewards/accuracies": 0.
|
| 484 |
-
"rewards/chosen": -0.
|
| 485 |
-
"rewards/margins": 0.
|
| 486 |
-
"rewards/rejected": -0.
|
| 487 |
"step": 320
|
| 488 |
},
|
| 489 |
{
|
| 490 |
"epoch": 0.32,
|
| 491 |
"learning_rate": 4.2708024172665795e-06,
|
| 492 |
-
"logits/chosen": -2.
|
| 493 |
-
"logits/rejected": -2.
|
| 494 |
-
"logps/chosen": -
|
| 495 |
-
"logps/rejected": -
|
| 496 |
-
"loss": 0.
|
| 497 |
-
"rewards/accuracies": 0.
|
| 498 |
-
"rewards/chosen": -0.
|
| 499 |
-
"rewards/margins": 0.
|
| 500 |
-
"rewards/rejected": -
|
| 501 |
"step": 330
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 0.33,
|
| 505 |
"learning_rate": 4.209111411211174e-06,
|
| 506 |
-
"logits/chosen": -2.
|
| 507 |
-
"logits/rejected": -2.
|
| 508 |
-
"logps/chosen": -
|
| 509 |
-
"logps/rejected": -
|
| 510 |
-
"loss": 0.
|
| 511 |
"rewards/accuracies": 0.75,
|
| 512 |
-
"rewards/chosen": -0.
|
| 513 |
-
"rewards/margins": 0.
|
| 514 |
-
"rewards/rejected": -
|
| 515 |
"step": 340
|
| 516 |
},
|
| 517 |
{
|
| 518 |
"epoch": 0.34,
|
| 519 |
"learning_rate": 4.145401411426788e-06,
|
| 520 |
-
"logits/chosen": -2.
|
| 521 |
-
"logits/rejected": -2.
|
| 522 |
-
"logps/chosen": -
|
| 523 |
-
"logps/rejected": -
|
| 524 |
-
"loss": 0.
|
| 525 |
-
"rewards/accuracies": 0.
|
| 526 |
-
"rewards/chosen": -0.
|
| 527 |
-
"rewards/margins": 0.
|
| 528 |
-
"rewards/rejected": -
|
| 529 |
"step": 350
|
| 530 |
},
|
| 531 |
{
|
| 532 |
"epoch": 0.35,
|
| 533 |
"learning_rate": 4.079747679297314e-06,
|
| 534 |
-
"logits/chosen": -2.
|
| 535 |
-
"logits/rejected": -2.
|
| 536 |
-
"logps/chosen": -
|
| 537 |
-
"logps/rejected": -
|
| 538 |
-
"loss": 0.
|
| 539 |
-
"rewards/accuracies": 0.
|
| 540 |
-
"rewards/chosen": -0.
|
| 541 |
-
"rewards/margins": 0.
|
| 542 |
-
"rewards/rejected": -
|
| 543 |
"step": 360
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 0.36,
|
| 547 |
"learning_rate": 4.012227772360889e-06,
|
| 548 |
-
"logits/chosen": -2.
|
| 549 |
-
"logits/rejected": -2.
|
| 550 |
-
"logps/chosen": -
|
| 551 |
-
"logps/rejected": -
|
| 552 |
-
"loss": 0.
|
| 553 |
-
"rewards/accuracies": 0.
|
| 554 |
-
"rewards/chosen": -0.
|
| 555 |
-
"rewards/margins": 0.
|
| 556 |
-
"rewards/rejected": -1.
|
| 557 |
"step": 370
|
| 558 |
},
|
| 559 |
{
|
| 560 |
"epoch": 0.37,
|
| 561 |
"learning_rate": 3.942921452690245e-06,
|
| 562 |
-
"logits/chosen": -2.
|
| 563 |
-
"logits/rejected": -2.
|
| 564 |
-
"logps/chosen": -
|
| 565 |
-
"logps/rejected": -
|
| 566 |
-
"loss": 0.
|
| 567 |
-
"rewards/accuracies": 0.
|
| 568 |
-
"rewards/chosen": -
|
| 569 |
-
"rewards/margins": 0.
|
| 570 |
-
"rewards/rejected": -
|
| 571 |
"step": 380
|
| 572 |
},
|
| 573 |
{
|
| 574 |
"epoch": 0.38,
|
| 575 |
"learning_rate": 3.871910592668817e-06,
|
| 576 |
-
"logits/chosen": -2.
|
| 577 |
-
"logits/rejected": -2.
|
| 578 |
-
"logps/chosen": -
|
| 579 |
-
"logps/rejected": -
|
| 580 |
-
"loss": 0.
|
| 581 |
-
"rewards/accuracies": 0.
|
| 582 |
-
"rewards/chosen": -
|
| 583 |
-
"rewards/margins": 0.
|
| 584 |
-
"rewards/rejected": -1.
|
| 585 |
"step": 390
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 0.39,
|
| 589 |
"learning_rate": 3.799279078273921e-06,
|
| 590 |
-
"logits/chosen": -2.
|
| 591 |
-
"logits/rejected": -2.
|
| 592 |
-
"logps/chosen": -
|
| 593 |
-
"logps/rejected": -
|
| 594 |
-
"loss": 0.
|
| 595 |
-
"rewards/accuracies": 0.
|
| 596 |
-
"rewards/chosen": -
|
| 597 |
-
"rewards/margins": 0.
|
| 598 |
-
"rewards/rejected": -
|
| 599 |
"step": 400
|
| 600 |
},
|
| 601 |
{
|
| 602 |
"epoch": 0.39,
|
| 603 |
-
"eval_logits/chosen": -2.
|
| 604 |
-
"eval_logits/rejected": -2.
|
| 605 |
-
"eval_logps/chosen": -
|
| 606 |
-
"eval_logps/rejected": -
|
| 607 |
-
"eval_loss": 0.
|
| 608 |
-
"eval_rewards/accuracies": 0.
|
| 609 |
-
"eval_rewards/chosen": -
|
| 610 |
-
"eval_rewards/margins": 0.
|
| 611 |
-
"eval_rewards/rejected": -
|
| 612 |
-
"eval_runtime":
|
| 613 |
-
"eval_samples_per_second": 1.
|
| 614 |
-
"eval_steps_per_second": 0.
|
| 615 |
"step": 400
|
| 616 |
},
|
| 617 |
{
|
| 618 |
"epoch": 0.4,
|
| 619 |
"learning_rate": 3.725112709981249e-06,
|
| 620 |
-
"logits/chosen": -2.
|
| 621 |
-
"logits/rejected": -2.
|
| 622 |
-
"logps/chosen": -
|
| 623 |
-
"logps/rejected": -
|
| 624 |
-
"loss": 0.
|
| 625 |
-
"rewards/accuracies": 0.
|
| 626 |
-
"rewards/chosen": -
|
| 627 |
-
"rewards/margins": 0.
|
| 628 |
-
"rewards/rejected": -
|
| 629 |
"step": 410
|
| 630 |
},
|
| 631 |
{
|
| 632 |
"epoch": 0.41,
|
| 633 |
"learning_rate": 3.649499101407737e-06,
|
| 634 |
-
"logits/chosen": -2.
|
| 635 |
-
"logits/rejected": -2.
|
| 636 |
-
"logps/chosen": -
|
| 637 |
-
"logps/rejected": -
|
| 638 |
-
"loss": 0.
|
| 639 |
"rewards/accuracies": 0.6875,
|
| 640 |
-
"rewards/chosen": -
|
| 641 |
-
"rewards/margins": 0.
|
| 642 |
-
"rewards/rejected": -
|
| 643 |
"step": 420
|
| 644 |
},
|
| 645 |
{
|
| 646 |
"epoch": 0.42,
|
| 647 |
"learning_rate": 3.5725275758125564e-06,
|
| 648 |
-
"logits/chosen": -2.
|
| 649 |
-
"logits/rejected": -2.
|
| 650 |
-
"logps/chosen": -
|
| 651 |
-
"logps/rejected": -
|
| 652 |
-
"loss": 0.
|
| 653 |
-
"rewards/accuracies": 0.
|
| 654 |
-
"rewards/chosen": -
|
| 655 |
-
"rewards/margins": 0.
|
| 656 |
-
"rewards/rejected": -
|
| 657 |
"step": 430
|
| 658 |
},
|
| 659 |
{
|
| 660 |
"epoch": 0.43,
|
| 661 |
"learning_rate": 3.494289060578478e-06,
|
| 662 |
-
"logits/chosen": -2.
|
| 663 |
-
"logits/rejected": -2.
|
| 664 |
-
"logps/chosen": -
|
| 665 |
-
"logps/rejected": -
|
| 666 |
-
"loss": 0.
|
| 667 |
-
"rewards/accuracies": 0.
|
| 668 |
-
"rewards/chosen": -
|
| 669 |
-
"rewards/margins": 0.
|
| 670 |
-
"rewards/rejected": -
|
| 671 |
"step": 440
|
| 672 |
},
|
| 673 |
{
|
| 674 |
"epoch": 0.44,
|
| 675 |
"learning_rate": 3.414875979798272e-06,
|
| 676 |
-
"logits/chosen": -2.
|
| 677 |
-
"logits/rejected": -2.
|
| 678 |
-
"logps/chosen": -
|
| 679 |
-
"logps/rejected": -
|
| 680 |
-
"loss": 0.
|
| 681 |
-
"rewards/accuracies": 0.
|
| 682 |
-
"rewards/chosen": -
|
| 683 |
-
"rewards/margins": 0.
|
| 684 |
-
"rewards/rejected": -
|
| 685 |
"step": 450
|
| 686 |
},
|
| 687 |
{
|
| 688 |
"epoch": 0.45,
|
| 689 |
"learning_rate": 3.3343821450930196e-06,
|
| 690 |
-
"logits/chosen": -2.
|
| 691 |
-
"logits/rejected": -2.
|
| 692 |
-
"logps/chosen": -
|
| 693 |
-
"logps/rejected": -
|
| 694 |
-
"loss": 0.
|
| 695 |
-
"rewards/accuracies": 0.
|
| 696 |
-
"rewards/chosen": -
|
| 697 |
-
"rewards/margins": 0.
|
| 698 |
-
"rewards/rejected": -
|
| 699 |
"step": 460
|
| 700 |
},
|
| 701 |
{
|
| 702 |
"epoch": 0.46,
|
| 703 |
"learning_rate": 3.252902644791325e-06,
|
| 704 |
-
"logits/chosen": -2.
|
| 705 |
-
"logits/rejected": -2.
|
| 706 |
-
"logps/chosen": -
|
| 707 |
-
"logps/rejected": -
|
| 708 |
-
"loss": 0.
|
| 709 |
-
"rewards/accuracies": 0.
|
| 710 |
-
"rewards/chosen": -
|
| 711 |
-
"rewards/margins": 0.
|
| 712 |
-
"rewards/rejected": -
|
| 713 |
"step": 470
|
| 714 |
},
|
| 715 |
{
|
| 716 |
"epoch": 0.47,
|
| 717 |
"learning_rate": 3.170533731600339e-06,
|
| 718 |
-
"logits/chosen": -2.
|
| 719 |
-
"logits/rejected": -2.
|
| 720 |
-
"logps/chosen": -
|
| 721 |
-
"logps/rejected": -
|
| 722 |
-
"loss": 0.
|
| 723 |
-
"rewards/accuracies": 0.
|
| 724 |
-
"rewards/chosen": -
|
| 725 |
-
"rewards/margins": 0.
|
| 726 |
-
"rewards/rejected": -
|
| 727 |
"step": 480
|
| 728 |
},
|
| 729 |
{
|
| 730 |
"epoch": 0.48,
|
| 731 |
"learning_rate": 3.0873727089012816e-06,
|
| 732 |
-
"logits/chosen": -2.
|
| 733 |
-
"logits/rejected": -2.
|
| 734 |
-
"logps/chosen": -
|
| 735 |
-
"logps/rejected": -
|
| 736 |
-
"loss": 0.
|
| 737 |
-
"rewards/accuracies": 0.
|
| 738 |
-
"rewards/chosen": -
|
| 739 |
-
"rewards/margins": 0.
|
| 740 |
-
"rewards/rejected": -
|
| 741 |
"step": 490
|
| 742 |
},
|
| 743 |
{
|
| 744 |
"epoch": 0.49,
|
| 745 |
"learning_rate": 3.0035178158038026e-06,
|
| 746 |
-
"logits/chosen": -2.
|
| 747 |
-
"logits/rejected": -2.
|
| 748 |
-
"logps/chosen": -
|
| 749 |
-
"logps/rejected": -
|
| 750 |
-
"loss": 0.
|
| 751 |
-
"rewards/accuracies": 0.
|
| 752 |
-
"rewards/chosen": -
|
| 753 |
-
"rewards/margins": 0.
|
| 754 |
-
"rewards/rejected": -
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 755 |
"step": 500
|
| 756 |
},
|
| 757 |
{
|
| 758 |
"epoch": 0.5,
|
| 759 |
"learning_rate": 2.919068111094937e-06,
|
| 760 |
-
"logits/chosen": -2.
|
| 761 |
-
"logits/rejected": -2.
|
| 762 |
-
"logps/chosen": -
|
| 763 |
-
"logps/rejected": -
|
| 764 |
-
"loss": 0.
|
| 765 |
-
"rewards/accuracies": 0.
|
| 766 |
-
"rewards/chosen": -
|
| 767 |
-
"rewards/margins": 0.
|
| 768 |
-
"rewards/rejected": -
|
| 769 |
"step": 510
|
| 770 |
},
|
| 771 |
{
|
| 772 |
"epoch": 0.51,
|
| 773 |
"learning_rate": 2.8341233562197895e-06,
|
| 774 |
-
"logits/chosen": -2.
|
| 775 |
-
"logits/rejected": -2.
|
| 776 |
-
"logps/chosen": -
|
| 777 |
-
"logps/rejected": -
|
| 778 |
-
"loss": 0.
|
| 779 |
-
"rewards/accuracies": 0.
|
| 780 |
-
"rewards/chosen": -
|
| 781 |
-
"rewards/margins": 0.
|
| 782 |
-
"rewards/rejected": -
|
| 783 |
"step": 520
|
| 784 |
},
|
| 785 |
{
|
| 786 |
"epoch": 0.52,
|
| 787 |
"learning_rate": 2.7487838974321352e-06,
|
| 788 |
-
"logits/chosen": -2.
|
| 789 |
-
"logits/rejected": -2.
|
| 790 |
-
"logps/chosen": -
|
| 791 |
-
"logps/rejected": -
|
| 792 |
-
"loss": 0.
|
| 793 |
-
"rewards/accuracies": 0.
|
| 794 |
-
"rewards/chosen": -
|
| 795 |
-
"rewards/margins": 0.
|
| 796 |
-
"rewards/rejected": -
|
| 797 |
"step": 530
|
| 798 |
},
|
| 799 |
{
|
| 800 |
"epoch": 0.53,
|
| 801 |
"learning_rate": 2.6631505472541997e-06,
|
| 802 |
-
"logits/chosen": -2.
|
| 803 |
-
"logits/rejected": -2.
|
| 804 |
-
"logps/chosen": -
|
| 805 |
-
"logps/rejected": -
|
| 806 |
-
"loss": 0.
|
| 807 |
-
"rewards/accuracies": 0.
|
| 808 |
-
"rewards/chosen": -
|
| 809 |
-
"rewards/margins":
|
| 810 |
-
"rewards/rejected": -
|
| 811 |
"step": 540
|
| 812 |
},
|
| 813 |
{
|
| 814 |
"epoch": 0.54,
|
| 815 |
"learning_rate": 2.5773244653856173e-06,
|
| 816 |
-
"logits/chosen": -2.
|
| 817 |
-
"logits/rejected": -2.
|
| 818 |
-
"logps/chosen": -
|
| 819 |
-
"logps/rejected": -
|
| 820 |
-
"loss": 0.
|
| 821 |
-
"rewards/accuracies": 0.
|
| 822 |
-
"rewards/chosen": -
|
| 823 |
-
"rewards/margins": 0.
|
| 824 |
-
"rewards/rejected": -
|
| 825 |
"step": 550
|
| 826 |
},
|
| 827 |
{
|
| 828 |
"epoch": 0.55,
|
| 829 |
"learning_rate": 2.4914070392022717e-06,
|
| 830 |
-
"logits/chosen": -2.
|
| 831 |
-
"logits/rejected": -2.
|
| 832 |
-
"logps/chosen": -
|
| 833 |
-
"logps/rejected": -
|
| 834 |
-
"loss": 0.
|
| 835 |
-
"rewards/accuracies": 0.
|
| 836 |
-
"rewards/chosen": -
|
| 837 |
-
"rewards/margins": 0.
|
| 838 |
-
"rewards/rejected": -
|
| 839 |
"step": 560
|
| 840 |
},
|
| 841 |
{
|
| 842 |
"epoch": 0.56,
|
| 843 |
"learning_rate": 2.4054997639861778e-06,
|
| 844 |
-
"logits/chosen": -2.
|
| 845 |
-
"logits/rejected": -2.
|
| 846 |
-
"logps/chosen": -
|
| 847 |
-
"logps/rejected": -
|
| 848 |
-
"loss": 0.
|
| 849 |
-
"rewards/accuracies": 0.
|
| 850 |
-
"rewards/chosen": -
|
| 851 |
-
"rewards/margins":
|
| 852 |
-
"rewards/rejected": -
|
| 853 |
"step": 570
|
| 854 |
},
|
| 855 |
{
|
| 856 |
"epoch": 0.57,
|
| 857 |
"learning_rate": 2.3197041230278905e-06,
|
| 858 |
-
"logits/chosen": -2.
|
| 859 |
-
"logits/rejected": -2.
|
| 860 |
-
"logps/chosen": -
|
| 861 |
-
"logps/rejected": -
|
| 862 |
-
"loss": 0.
|
| 863 |
-
"rewards/accuracies": 0.
|
| 864 |
-
"rewards/chosen": -
|
| 865 |
-
"rewards/margins":
|
| 866 |
-
"rewards/rejected": -
|
| 867 |
"step": 580
|
| 868 |
},
|
| 869 |
{
|
| 870 |
"epoch": 0.58,
|
| 871 |
"learning_rate": 2.234121467743082e-06,
|
| 872 |
-
"logits/chosen": -2.
|
| 873 |
-
"logits/rejected": -2.
|
| 874 |
-
"logps/chosen": -
|
| 875 |
-
"logps/rejected": -
|
| 876 |
-
"loss": 0.
|
| 877 |
-
"rewards/accuracies": 0.
|
| 878 |
-
"rewards/chosen": -
|
| 879 |
-
"rewards/margins": 0.
|
| 880 |
-
"rewards/rejected": -
|
| 881 |
"step": 590
|
| 882 |
},
|
| 883 |
{
|
| 884 |
"epoch": 0.59,
|
| 885 |
"learning_rate": 2.148852897944905e-06,
|
| 886 |
-
"logits/chosen": -2.
|
| 887 |
-
"logits/rejected": -2.
|
| 888 |
-
"logps/chosen": -
|
| 889 |
-
"logps/rejected": -
|
| 890 |
-
"loss": 0.
|
| 891 |
-
"rewards/accuracies": 0.
|
| 892 |
-
"rewards/chosen": -
|
| 893 |
-
"rewards/margins":
|
| 894 |
-
"rewards/rejected": -
|
| 895 |
"step": 600
|
| 896 |
},
|
| 897 |
{
|
| 898 |
"epoch": 0.59,
|
| 899 |
-
"eval_logits/chosen": -2.
|
| 900 |
-
"eval_logits/rejected": -2.
|
| 901 |
-
"eval_logps/chosen": -
|
| 902 |
-
"eval_logps/rejected": -
|
| 903 |
-
"eval_loss": 0.
|
| 904 |
-
"eval_rewards/accuracies": 0.
|
| 905 |
-
"eval_rewards/chosen": -
|
| 906 |
-
"eval_rewards/margins": 0.
|
| 907 |
-
"eval_rewards/rejected": -
|
| 908 |
-
"eval_runtime":
|
| 909 |
-
"eval_samples_per_second":
|
| 910 |
-
"eval_steps_per_second": 0.
|
| 911 |
"step": 600
|
| 912 |
},
|
| 913 |
{
|
| 914 |
"epoch": 0.6,
|
| 915 |
"learning_rate": 2.063999142413574e-06,
|
| 916 |
-
"logits/chosen": -2.
|
| 917 |
-
"logits/rejected": -2.
|
| 918 |
-
"logps/chosen": -
|
| 919 |
-
"logps/rejected": -
|
| 920 |
-
"loss": 0.
|
| 921 |
"rewards/accuracies": 0.762499988079071,
|
| 922 |
-
"rewards/chosen": -
|
| 923 |
-
"rewards/margins": 0.
|
| 924 |
-
"rewards/rejected": -
|
| 925 |
"step": 610
|
| 926 |
},
|
| 927 |
{
|
| 928 |
"epoch": 0.61,
|
| 929 |
"learning_rate": 1.9796604399042547e-06,
|
| 930 |
-
"logits/chosen": -2.
|
| 931 |
-
"logits/rejected": -2.
|
| 932 |
-
"logps/chosen": -
|
| 933 |
-
"logps/rejected": -
|
| 934 |
-
"loss": 0.
|
| 935 |
-
"rewards/accuracies": 0.
|
| 936 |
-
"rewards/chosen": -
|
| 937 |
-
"rewards/margins":
|
| 938 |
-
"rewards/rejected": -
|
| 939 |
"step": 620
|
| 940 |
},
|
| 941 |
{
|
| 942 |
"epoch": 0.62,
|
| 943 |
"learning_rate": 1.8959364207338216e-06,
|
| 944 |
-
"logits/chosen": -2.
|
| 945 |
-
"logits/rejected": -2.
|
| 946 |
-
"logps/chosen": -
|
| 947 |
-
"logps/rejected": -
|
| 948 |
-
"loss": 0.
|
| 949 |
-
"rewards/accuracies": 0.
|
| 950 |
-
"rewards/chosen": -
|
| 951 |
-
"rewards/margins":
|
| 952 |
-
"rewards/rejected": -
|
| 953 |
"step": 630
|
| 954 |
},
|
| 955 |
{
|
| 956 |
"epoch": 0.63,
|
| 957 |
"learning_rate": 1.8129259890863825e-06,
|
| 958 |
-
"logits/chosen": -2.
|
| 959 |
-
"logits/rejected": -2.
|
| 960 |
-
"logps/chosen": -
|
| 961 |
-
"logps/rejected": -
|
| 962 |
-
"loss": 0.
|
| 963 |
-
"rewards/accuracies": 0.
|
| 964 |
-
"rewards/chosen": -
|
| 965 |
-
"rewards/margins":
|
| 966 |
-
"rewards/rejected": -
|
| 967 |
"step": 640
|
| 968 |
},
|
| 969 |
{
|
| 970 |
"epoch": 0.64,
|
| 971 |
"learning_rate": 1.7307272061765738e-06,
|
| 972 |
-
"logits/chosen": -2.
|
| 973 |
-
"logits/rejected": -2.
|
| 974 |
-
"logps/chosen": -
|
| 975 |
-
"logps/rejected": -
|
| 976 |
-
"loss": 0.
|
| 977 |
-
"rewards/accuracies": 0.
|
| 978 |
-
"rewards/chosen": -
|
| 979 |
-
"rewards/margins":
|
| 980 |
-
"rewards/rejected": -
|
| 981 |
"step": 650
|
| 982 |
},
|
| 983 |
{
|
| 984 |
"epoch": 0.65,
|
| 985 |
"learning_rate": 1.649437174408685e-06,
|
| 986 |
-
"logits/chosen": -2.
|
| 987 |
-
"logits/rejected": -2.
|
| 988 |
-
"logps/chosen": -
|
| 989 |
-
"logps/rejected": -
|
| 990 |
-
"loss": 0.
|
| 991 |
"rewards/accuracies": 0.75,
|
| 992 |
-
"rewards/chosen": -
|
| 993 |
-
"rewards/margins":
|
| 994 |
-
"rewards/rejected": -
|
| 995 |
"step": 660
|
| 996 |
},
|
| 997 |
{
|
| 998 |
"epoch": 0.66,
|
| 999 |
"learning_rate": 1.569151922668422e-06,
|
| 1000 |
-
"logits/chosen": -2.
|
| 1001 |
-
"logits/rejected": -2.
|
| 1002 |
-
"logps/chosen": -
|
| 1003 |
-
"logps/rejected": -
|
| 1004 |
-
"loss": 0.
|
| 1005 |
-
"rewards/accuracies": 0.
|
| 1006 |
-
"rewards/chosen": -
|
| 1007 |
-
"rewards/margins":
|
| 1008 |
-
"rewards/rejected": -
|
| 1009 |
"step": 670
|
| 1010 |
},
|
| 1011 |
{
|
| 1012 |
"epoch": 0.67,
|
| 1013 |
"learning_rate": 1.4899662928828428e-06,
|
| 1014 |
-
"logits/chosen": -2.
|
| 1015 |
-
"logits/rejected": -2.
|
| 1016 |
-
"logps/chosen": -
|
| 1017 |
-
"logps/rejected": -
|
| 1018 |
-
"loss": 0.
|
| 1019 |
-
"rewards/accuracies": 0.
|
| 1020 |
-
"rewards/chosen": -
|
| 1021 |
-
"rewards/margins":
|
| 1022 |
-
"rewards/rejected": -
|
| 1023 |
"step": 680
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
"epoch": 0.68,
|
| 1027 |
"learning_rate": 1.4119738279824507e-06,
|
| 1028 |
-
"logits/chosen": -2.
|
| 1029 |
-
"logits/rejected": -2.
|
| 1030 |
-
"logps/chosen": -
|
| 1031 |
-
"logps/rejected": -
|
| 1032 |
-
"loss": 0.
|
| 1033 |
"rewards/accuracies": 0.7250000238418579,
|
| 1034 |
-
"rewards/chosen": -
|
| 1035 |
-
"rewards/margins": 0.
|
| 1036 |
-
"rewards/rejected": -
|
| 1037 |
"step": 690
|
| 1038 |
},
|
| 1039 |
{
|
| 1040 |
"epoch": 0.69,
|
| 1041 |
"learning_rate": 1.3352666613978152e-06,
|
| 1042 |
-
"logits/chosen": -2.
|
| 1043 |
-
"logits/rejected": -2.
|
| 1044 |
-
"logps/chosen": -
|
| 1045 |
-
"logps/rejected": -
|
| 1046 |
-
"loss": 0.
|
| 1047 |
-
"rewards/accuracies": 0.
|
| 1048 |
-
"rewards/chosen": -
|
| 1049 |
-
"rewards/margins": 0.
|
| 1050 |
-
"rewards/rejected": -
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1051 |
"step": 700
|
| 1052 |
},
|
| 1053 |
{
|
| 1054 |
"epoch": 0.7,
|
| 1055 |
"learning_rate": 1.2599354082212523e-06,
|
| 1056 |
-
"logits/chosen": -2.
|
| 1057 |
-
"logits/rejected": -2.
|
| 1058 |
-
"logps/chosen": -
|
| 1059 |
-
"logps/rejected": -
|
| 1060 |
-
"loss": 0.
|
| 1061 |
-
"rewards/accuracies": 0.
|
| 1062 |
-
"rewards/chosen": -
|
| 1063 |
-
"rewards/margins": 0.
|
| 1064 |
-
"rewards/rejected": -
|
| 1065 |
"step": 710
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
"epoch": 0.71,
|
| 1069 |
"learning_rate": 1.186069058162127e-06,
|
| 1070 |
-
"logits/chosen": -2.
|
| 1071 |
-
"logits/rejected": -2.
|
| 1072 |
-
"logps/chosen": -
|
| 1073 |
-
"logps/rejected": -
|
| 1074 |
-
"loss": 0.
|
| 1075 |
"rewards/accuracies": 0.737500011920929,
|
| 1076 |
-
"rewards/chosen": -
|
| 1077 |
-
"rewards/margins": 0.
|
| 1078 |
-
"rewards/rejected": -
|
| 1079 |
"step": 720
|
| 1080 |
},
|
| 1081 |
{
|
| 1082 |
"epoch": 0.72,
|
| 1083 |
"learning_rate": 1.113754870422254e-06,
|
| 1084 |
-
"logits/chosen": -2.
|
| 1085 |
-
"logits/rejected": -2.
|
| 1086 |
-
"logps/chosen": -
|
| 1087 |
-
"logps/rejected": -
|
| 1088 |
-
"loss": 0.
|
| 1089 |
-
"rewards/accuracies": 0.
|
| 1090 |
-
"rewards/chosen": -
|
| 1091 |
-
"rewards/margins": 0.
|
| 1092 |
-
"rewards/rejected": -
|
| 1093 |
"step": 730
|
| 1094 |
},
|
| 1095 |
{
|
| 1096 |
"epoch": 0.73,
|
| 1097 |
"learning_rate": 1.0430782706155545e-06,
|
| 1098 |
-
"logits/chosen": -2.
|
| 1099 |
-
"logits/rejected": -2.
|
| 1100 |
-
"logps/chosen": -
|
| 1101 |
-
"logps/rejected": -
|
| 1102 |
-
"loss": 0.
|
| 1103 |
-
"rewards/accuracies": 0.
|
| 1104 |
-
"rewards/chosen": -
|
| 1105 |
-
"rewards/margins":
|
| 1106 |
-
"rewards/rejected": -
|
| 1107 |
"step": 740
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
"epoch": 0.74,
|
| 1111 |
"learning_rate": 9.741227498537615e-07,
|
| 1112 |
-
"logits/chosen": -2.
|
| 1113 |
-
"logits/rejected": -2.
|
| 1114 |
-
"logps/chosen": -
|
| 1115 |
-
"logps/rejected": -
|
| 1116 |
-
"loss": 0.
|
| 1117 |
-
"rewards/accuracies": 0.
|
| 1118 |
-
"rewards/chosen": -
|
| 1119 |
-
"rewards/margins":
|
| 1120 |
-
"rewards/rejected": -
|
| 1121 |
"step": 750
|
| 1122 |
},
|
| 1123 |
{
|
| 1124 |
"epoch": 0.75,
|
| 1125 |
"learning_rate": 9.069697661173668e-07,
|
| 1126 |
-
"logits/chosen": -2.
|
| 1127 |
-
"logits/rejected": -2.
|
| 1128 |
-
"logps/chosen": -
|
| 1129 |
-
"logps/rejected": -
|
| 1130 |
-
"loss": 0.
|
| 1131 |
-
"rewards/accuracies": 0.
|
| 1132 |
-
"rewards/chosen": -
|
| 1133 |
-
"rewards/margins": 0.
|
| 1134 |
-
"rewards/rejected": -
|
| 1135 |
"step": 760
|
| 1136 |
},
|
| 1137 |
{
|
| 1138 |
"epoch": 0.76,
|
| 1139 |
"learning_rate": 8.416986480283434e-07,
|
| 1140 |
-
"logits/chosen": -2.
|
| 1141 |
-
"logits/rejected": -2.
|
| 1142 |
-
"logps/chosen": -
|
| 1143 |
-
"logps/rejected": -
|
| 1144 |
-
"loss": 0.
|
| 1145 |
-
"rewards/accuracies": 0.
|
| 1146 |
-
"rewards/chosen": -
|
| 1147 |
-
"rewards/margins":
|
| 1148 |
-
"rewards/rejected": -
|
| 1149 |
"step": 770
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
"epoch": 0.77,
|
| 1153 |
"learning_rate": 7.783865011382876e-07,
|
| 1154 |
-
"logits/chosen": -2.
|
| 1155 |
-
"logits/rejected": -2.
|
| 1156 |
-
"logps/chosen": -
|
| 1157 |
-
"logps/rejected": -
|
| 1158 |
-
"loss": 0.
|
| 1159 |
-
"rewards/accuracies": 0.
|
| 1160 |
-
"rewards/chosen": -
|
| 1161 |
-
"rewards/margins":
|
| 1162 |
-
"rewards/rejected": -
|
| 1163 |
"step": 780
|
| 1164 |
},
|
| 1165 |
{
|
| 1166 |
"epoch": 0.78,
|
| 1167 |
"learning_rate": 7.171081168427205e-07,
|
| 1168 |
-
"logits/chosen": -2.
|
| 1169 |
-
"logits/rejected": -2.
|
| 1170 |
-
"logps/chosen": -
|
| 1171 |
-
"logps/rejected": -
|
| 1172 |
-
"loss": 0.
|
| 1173 |
-
"rewards/accuracies": 0.
|
| 1174 |
-
"rewards/chosen": -
|
| 1175 |
-
"rewards/margins":
|
| 1176 |
-
"rewards/rejected": -
|
| 1177 |
"step": 790
|
| 1178 |
},
|
| 1179 |
{
|
| 1180 |
"epoch": 0.79,
|
| 1181 |
"learning_rate": 6.579358840291064e-07,
|
| 1182 |
-
"logits/chosen": -2.
|
| 1183 |
-
"logits/rejected": -2.
|
| 1184 |
-
"logps/chosen": -
|
| 1185 |
-
"logps/rejected": -
|
| 1186 |
-
"loss": 0.
|
| 1187 |
-
"rewards/accuracies": 0.
|
| 1188 |
-
"rewards/chosen": -
|
| 1189 |
-
"rewards/margins": 0.
|
| 1190 |
-
"rewards/rejected": -
|
| 1191 |
"step": 800
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
"epoch": 0.79,
|
| 1195 |
-
"eval_logits/chosen": -2.
|
| 1196 |
-
"eval_logits/rejected": -2.
|
| 1197 |
-
"eval_logps/chosen": -
|
| 1198 |
-
"eval_logps/rejected": -
|
| 1199 |
-
"eval_loss": 0.
|
| 1200 |
-
"eval_rewards/accuracies": 0.
|
| 1201 |
-
"eval_rewards/chosen": -
|
| 1202 |
-
"eval_rewards/margins": 0.
|
| 1203 |
-
"eval_rewards/rejected": -
|
| 1204 |
-
"eval_runtime":
|
| 1205 |
-
"eval_samples_per_second":
|
| 1206 |
-
"eval_steps_per_second": 0.
|
| 1207 |
"step": 800
|
| 1208 |
},
|
| 1209 |
{
|
| 1210 |
"epoch": 0.8,
|
| 1211 |
"learning_rate": 6.00939703563006e-07,
|
| 1212 |
-
"logits/chosen": -2.
|
| 1213 |
-
"logits/rejected": -2.
|
| 1214 |
-
"logps/chosen": -
|
| 1215 |
-
"logps/rejected": -
|
| 1216 |
-
"loss": 0.
|
| 1217 |
-
"rewards/accuracies": 0.
|
| 1218 |
-
"rewards/chosen": -
|
| 1219 |
-
"rewards/margins": 0.
|
| 1220 |
-
"rewards/rejected": -
|
| 1221 |
"step": 810
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
"epoch": 0.81,
|
| 1225 |
"learning_rate": 5.461869057133412e-07,
|
| 1226 |
-
"logits/chosen": -2.
|
| 1227 |
-
"logits/rejected": -2.
|
| 1228 |
-
"logps/chosen": -
|
| 1229 |
-
"logps/rejected": -
|
| 1230 |
-
"loss": 0.
|
| 1231 |
-
"rewards/accuracies": 0.
|
| 1232 |
-
"rewards/chosen": -
|
| 1233 |
-
"rewards/margins":
|
| 1234 |
-
"rewards/rejected": -
|
| 1235 |
"step": 820
|
| 1236 |
},
|
| 1237 |
{
|
| 1238 |
"epoch": 0.82,
|
| 1239 |
"learning_rate": 4.937421706143497e-07,
|
| 1240 |
-
"logits/chosen": -2.
|
| 1241 |
-
"logits/rejected": -2.
|
| 1242 |
-
"logps/chosen": -
|
| 1243 |
-
"logps/rejected": -
|
| 1244 |
-
"loss": 0.
|
| 1245 |
-
"rewards/accuracies": 0.
|
| 1246 |
-
"rewards/chosen": -
|
| 1247 |
-
"rewards/margins": 0.
|
| 1248 |
-
"rewards/rejected": -
|
| 1249 |
"step": 830
|
| 1250 |
},
|
| 1251 |
{
|
| 1252 |
"epoch": 0.83,
|
| 1253 |
"learning_rate": 4.43667451858166e-07,
|
| 1254 |
-
"logits/chosen": -2.
|
| 1255 |
-
"logits/rejected": -2.
|
| 1256 |
-
"logps/chosen": -
|
| 1257 |
-
"logps/rejected": -
|
| 1258 |
-
"loss": 0.
|
| 1259 |
-
"rewards/accuracies": 0.
|
| 1260 |
-
"rewards/chosen": -
|
| 1261 |
-
"rewards/margins":
|
| 1262 |
-
"rewards/rejected": -
|
| 1263 |
"step": 840
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
"epoch": 0.84,
|
| 1267 |
"learning_rate": 3.9602190330830484e-07,
|
| 1268 |
-
"logits/chosen": -2.
|
| 1269 |
-
"logits/rejected": -2.
|
| 1270 |
-
"logps/chosen": -
|
| 1271 |
-
"logps/rejected": -
|
| 1272 |
-
"loss": 0.
|
| 1273 |
-
"rewards/accuracies": 0.
|
| 1274 |
-
"rewards/chosen": -
|
| 1275 |
-
"rewards/margins":
|
| 1276 |
-
"rewards/rejected": -
|
| 1277 |
"step": 850
|
| 1278 |
},
|
| 1279 |
{
|
| 1280 |
"epoch": 0.85,
|
| 1281 |
"learning_rate": 3.5086180922049295e-07,
|
| 1282 |
-
"logits/chosen": -2.
|
| 1283 |
-
"logits/rejected": -2.
|
| 1284 |
-
"logps/chosen": -
|
| 1285 |
-
"logps/rejected": -
|
| 1286 |
-
"loss": 0.
|
| 1287 |
-
"rewards/accuracies": 0.
|
| 1288 |
-
"rewards/chosen": -
|
| 1289 |
-
"rewards/margins":
|
| 1290 |
-
"rewards/rejected": -
|
| 1291 |
"step": 860
|
| 1292 |
},
|
| 1293 |
{
|
| 1294 |
"epoch": 0.86,
|
| 1295 |
"learning_rate": 3.0824051775340895e-07,
|
| 1296 |
-
"logits/chosen": -2.
|
| 1297 |
-
"logits/rejected": -2.
|
| 1298 |
-
"logps/chosen": -
|
| 1299 |
-
"logps/rejected": -
|
| 1300 |
-
"loss": 0.
|
| 1301 |
-
"rewards/accuracies": 0.
|
| 1302 |
-
"rewards/chosen": -
|
| 1303 |
-
"rewards/margins": 0.
|
| 1304 |
-
"rewards/rejected": -
|
| 1305 |
"step": 870
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
"epoch": 0.87,
|
| 1309 |
"learning_rate": 2.6820837794786336e-07,
|
| 1310 |
-
"logits/chosen": -2.
|
| 1311 |
-
"logits/rejected": -2.
|
| 1312 |
-
"logps/chosen": -
|
| 1313 |
-
"logps/rejected": -
|
| 1314 |
-
"loss": 0.
|
| 1315 |
-
"rewards/accuracies": 0.
|
| 1316 |
-
"rewards/chosen": -
|
| 1317 |
-
"rewards/margins": 0.
|
| 1318 |
-
"rewards/rejected": -
|
| 1319 |
"step": 880
|
| 1320 |
},
|
| 1321 |
{
|
| 1322 |
"epoch": 0.88,
|
| 1323 |
"learning_rate": 2.3081268024887694e-07,
|
| 1324 |
-
"logits/chosen": -2.
|
| 1325 |
-
"logits/rejected": -2.
|
| 1326 |
-
"logps/chosen": -
|
| 1327 |
-
"logps/rejected": -
|
| 1328 |
-
"loss": 0.
|
| 1329 |
-
"rewards/accuracies": 0.
|
| 1330 |
-
"rewards/chosen": -
|
| 1331 |
-
"rewards/margins":
|
| 1332 |
-
"rewards/rejected": -
|
| 1333 |
"step": 890
|
| 1334 |
},
|
| 1335 |
{
|
| 1336 |
"epoch": 0.89,
|
| 1337 |
"learning_rate": 1.9609760064091044e-07,
|
| 1338 |
-
"logits/chosen": -2.
|
| 1339 |
-
"logits/rejected": -2.
|
| 1340 |
-
"logps/chosen": -
|
| 1341 |
-
"logps/rejected": -
|
| 1342 |
-
"loss": 0.
|
| 1343 |
-
"rewards/accuracies": 0.
|
| 1344 |
-
"rewards/chosen": -
|
| 1345 |
-
"rewards/margins": 0.
|
| 1346 |
-
"rewards/rejected": -
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1347 |
"step": 900
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
"epoch": 0.9,
|
| 1351 |
"learning_rate": 1.6410414846224992e-07,
|
| 1352 |
-
"logits/chosen": -2.
|
| 1353 |
-
"logits/rejected": -2.
|
| 1354 |
-
"logps/chosen": -
|
| 1355 |
-
"logps/rejected": -
|
| 1356 |
-
"loss": 0.
|
| 1357 |
"rewards/accuracies": 0.7875000238418579,
|
| 1358 |
-
"rewards/chosen": -
|
| 1359 |
-
"rewards/margins":
|
| 1360 |
-
"rewards/rejected": -
|
| 1361 |
"step": 910
|
| 1362 |
},
|
| 1363 |
{
|
| 1364 |
"epoch": 0.91,
|
| 1365 |
"learning_rate": 1.348701179601819e-07,
|
| 1366 |
-
"logits/chosen": -2.
|
| 1367 |
-
"logits/rejected": -2.
|
| 1368 |
-
"logps/chosen": -
|
| 1369 |
-
"logps/rejected": -
|
| 1370 |
-
"loss": 0.
|
| 1371 |
"rewards/accuracies": 0.7875000238418579,
|
| 1372 |
-
"rewards/chosen": -
|
| 1373 |
-
"rewards/margins":
|
| 1374 |
-
"rewards/rejected": -
|
| 1375 |
"step": 920
|
| 1376 |
},
|
| 1377 |
{
|
| 1378 |
"epoch": 0.91,
|
| 1379 |
"learning_rate": 1.0843004364420151e-07,
|
| 1380 |
-
"logits/chosen": -2.
|
| 1381 |
-
"logits/rejected": -2.
|
| 1382 |
-
"logps/chosen": -
|
| 1383 |
-
"logps/rejected": -
|
| 1384 |
-
"loss": 0.
|
| 1385 |
-
"rewards/accuracies": 0.
|
| 1386 |
-
"rewards/chosen": -
|
| 1387 |
-
"rewards/margins": 0.
|
| 1388 |
-
"rewards/rejected": -
|
| 1389 |
"step": 930
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
"epoch": 0.92,
|
| 1393 |
"learning_rate": 8.481515948997931e-08,
|
| 1394 |
-
"logits/chosen": -2.
|
| 1395 |
-
"logits/rejected": -2.
|
| 1396 |
-
"logps/chosen": -
|
| 1397 |
-
"logps/rejected": -
|
| 1398 |
-
"loss": 0.
|
| 1399 |
-
"rewards/accuracies": 0.
|
| 1400 |
-
"rewards/chosen": -
|
| 1401 |
-
"rewards/margins": 0.
|
| 1402 |
-
"rewards/rejected": -
|
| 1403 |
"step": 940
|
| 1404 |
},
|
| 1405 |
{
|
| 1406 |
"epoch": 0.93,
|
| 1407 |
"learning_rate": 6.4053362042297e-08,
|
| 1408 |
-
"logits/chosen": -2.
|
| 1409 |
-
"logits/rejected": -2.
|
| 1410 |
-
"logps/chosen": -
|
| 1411 |
-
"logps/rejected": -
|
| 1412 |
-
"loss": 0.
|
| 1413 |
-
"rewards/accuracies": 0.
|
| 1414 |
-
"rewards/chosen": -
|
| 1415 |
-
"rewards/margins":
|
| 1416 |
-
"rewards/rejected": -
|
| 1417 |
"step": 950
|
| 1418 |
},
|
| 1419 |
{
|
| 1420 |
"epoch": 0.94,
|
| 1421 |
"learning_rate": 4.616917746052163e-08,
|
| 1422 |
-
"logits/chosen": -2.
|
| 1423 |
-
"logits/rejected": -2.
|
| 1424 |
-
"logps/chosen": -
|
| 1425 |
-
"logps/rejected": -
|
| 1426 |
-
"loss": 0.
|
| 1427 |
-
"rewards/accuracies": 0.
|
| 1428 |
-
"rewards/chosen": -
|
| 1429 |
-
"rewards/margins": 0.
|
| 1430 |
-
"rewards/rejected": -
|
| 1431 |
"step": 960
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
"epoch": 0.95,
|
| 1435 |
"learning_rate": 3.118373254556412e-08,
|
| 1436 |
-
"logits/chosen": -2.
|
| 1437 |
-
"logits/rejected": -2.
|
| 1438 |
-
"logps/chosen": -
|
| 1439 |
-
"logps/rejected": -
|
| 1440 |
-
"loss": 0.
|
| 1441 |
-
"rewards/accuracies": 0.
|
| 1442 |
-
"rewards/chosen": -
|
| 1443 |
-
"rewards/margins":
|
| 1444 |
-
"rewards/rejected": -
|
| 1445 |
"step": 970
|
| 1446 |
},
|
| 1447 |
{
|
| 1448 |
"epoch": 0.96,
|
| 1449 |
"learning_rate": 1.9114729782535037e-08,
|
| 1450 |
-
"logits/chosen": -2.
|
| 1451 |
-
"logits/rejected": -2.
|
| 1452 |
-
"logps/chosen": -
|
| 1453 |
-
"logps/rejected": -
|
| 1454 |
-
"loss": 0.
|
| 1455 |
-
"rewards/accuracies": 0.
|
| 1456 |
-
"rewards/chosen": -
|
| 1457 |
-
"rewards/margins": 0.
|
| 1458 |
-
"rewards/rejected": -
|
| 1459 |
"step": 980
|
| 1460 |
},
|
| 1461 |
{
|
| 1462 |
"epoch": 0.97,
|
| 1463 |
"learning_rate": 9.97642642858815e-09,
|
| 1464 |
-
"logits/chosen": -2.
|
| 1465 |
-
"logits/rejected": -2.
|
| 1466 |
-
"logps/chosen": -
|
| 1467 |
-
"logps/rejected": -
|
| 1468 |
-
"loss": 0.
|
| 1469 |
-
"rewards/accuracies": 0.
|
| 1470 |
-
"rewards/chosen": -
|
| 1471 |
-
"rewards/margins": 0.
|
| 1472 |
-
"rewards/rejected": -
|
| 1473 |
"step": 990
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
"epoch": 0.98,
|
| 1477 |
"learning_rate": 3.779617670651436e-09,
|
| 1478 |
-
"logits/chosen": -2.
|
| 1479 |
-
"logits/rejected": -2.
|
| 1480 |
-
"logps/chosen": -
|
| 1481 |
-
"logps/rejected": -
|
| 1482 |
-
"loss": 0.
|
| 1483 |
-
"rewards/accuracies": 0.
|
| 1484 |
-
"rewards/chosen": -
|
| 1485 |
-
"rewards/margins": 0.
|
| 1486 |
-
"rewards/rejected": -
|
| 1487 |
"step": 1000
|
| 1488 |
},
|
| 1489 |
{
|
| 1490 |
"epoch": 0.98,
|
| 1491 |
-
"eval_logits/chosen": -2.
|
| 1492 |
-
"eval_logits/rejected": -2.
|
| 1493 |
-
"eval_logps/chosen": -
|
| 1494 |
-
"eval_logps/rejected": -
|
| 1495 |
-
"eval_loss": 0.
|
| 1496 |
-
"eval_rewards/accuracies": 0.
|
| 1497 |
-
"eval_rewards/chosen": -
|
| 1498 |
-
"eval_rewards/margins": 0.
|
| 1499 |
-
"eval_rewards/rejected": -
|
| 1500 |
-
"eval_runtime":
|
| 1501 |
-
"eval_samples_per_second":
|
| 1502 |
-
"eval_steps_per_second": 0.
|
| 1503 |
"step": 1000
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
"epoch": 0.99,
|
| 1507 |
"learning_rate": 5.316238729444201e-10,
|
| 1508 |
-
"logits/chosen": -2.
|
| 1509 |
-
"logits/rejected": -2.
|
| 1510 |
-
"logps/chosen": -
|
| 1511 |
-
"logps/rejected": -
|
| 1512 |
-
"loss": 0.
|
| 1513 |
-
"rewards/accuracies": 0.
|
| 1514 |
-
"rewards/chosen": -
|
| 1515 |
-
"rewards/margins":
|
| 1516 |
-
"rewards/rejected": -
|
| 1517 |
"step": 1010
|
| 1518 |
},
|
| 1519 |
{
|
| 1520 |
"epoch": 1.0,
|
| 1521 |
"step": 1016,
|
| 1522 |
"total_flos": 0.0,
|
| 1523 |
-
"train_loss": 0.
|
| 1524 |
-
"train_runtime":
|
| 1525 |
-
"train_samples_per_second": 0.
|
| 1526 |
-
"train_steps_per_second": 0.
|
| 1527 |
}
|
| 1528 |
],
|
| 1529 |
"logging_steps": 10,
|
|
|
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 0.999508116084604,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
"global_step": 1016,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
|
|
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
"learning_rate": 4.901960784313726e-08,
|
| 14 |
+
"logits/chosen": -2.493000030517578,
|
| 15 |
+
"logits/rejected": -2.5710701942443848,
|
| 16 |
+
"logps/chosen": -136.35806274414062,
|
| 17 |
+
"logps/rejected": -147.18719482421875,
|
| 18 |
"loss": 0.6931,
|
| 19 |
"rewards/accuracies": 0.0,
|
| 20 |
"rewards/chosen": 0.0,
|
|
|
|
| 25 |
{
|
| 26 |
"epoch": 0.01,
|
| 27 |
"learning_rate": 4.901960784313725e-07,
|
| 28 |
+
"logits/chosen": -2.5276925563812256,
|
| 29 |
+
"logits/rejected": -2.4688756465911865,
|
| 30 |
+
"logps/chosen": -160.81719970703125,
|
| 31 |
+
"logps/rejected": -124.57797241210938,
|
| 32 |
+
"loss": 0.6933,
|
| 33 |
+
"rewards/accuracies": 0.3888888955116272,
|
| 34 |
+
"rewards/chosen": -0.00038504067924804986,
|
| 35 |
+
"rewards/margins": -0.0003390010679140687,
|
| 36 |
+
"rewards/rejected": -4.6039687731536105e-05,
|
| 37 |
"step": 10
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.02,
|
| 41 |
"learning_rate": 9.80392156862745e-07,
|
| 42 |
+
"logits/chosen": -2.5096638202667236,
|
| 43 |
+
"logits/rejected": -2.4554123878479004,
|
| 44 |
+
"logps/chosen": -169.54324340820312,
|
| 45 |
+
"logps/rejected": -142.311767578125,
|
| 46 |
+
"loss": 0.6933,
|
| 47 |
+
"rewards/accuracies": 0.4375,
|
| 48 |
+
"rewards/chosen": -0.00011383053060853854,
|
| 49 |
+
"rewards/margins": -0.00031454654526896775,
|
| 50 |
+
"rewards/rejected": 0.0002007160073844716,
|
| 51 |
"step": 20
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.03,
|
| 55 |
"learning_rate": 1.4705882352941177e-06,
|
| 56 |
+
"logits/chosen": -2.5936665534973145,
|
| 57 |
+
"logits/rejected": -2.5427086353302,
|
| 58 |
+
"logps/chosen": -158.86428833007812,
|
| 59 |
+
"logps/rejected": -142.90164184570312,
|
| 60 |
+
"loss": 0.6932,
|
| 61 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 62 |
+
"rewards/chosen": 0.0005116682150401175,
|
| 63 |
+
"rewards/margins": -0.0001444850640837103,
|
| 64 |
+
"rewards/rejected": 0.0006561531918123364,
|
| 65 |
"step": 30
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.04,
|
| 69 |
"learning_rate": 1.96078431372549e-06,
|
| 70 |
+
"logits/chosen": -2.5783634185791016,
|
| 71 |
+
"logits/rejected": -2.514772415161133,
|
| 72 |
+
"logps/chosen": -163.1798858642578,
|
| 73 |
+
"logps/rejected": -140.1341094970703,
|
| 74 |
+
"loss": 0.6928,
|
| 75 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 76 |
+
"rewards/chosen": 0.0009201603243127465,
|
| 77 |
+
"rewards/margins": 0.0006632919539697468,
|
| 78 |
+
"rewards/rejected": 0.0002568683703429997,
|
| 79 |
"step": 40
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.05,
|
| 83 |
"learning_rate": 2.450980392156863e-06,
|
| 84 |
+
"logits/chosen": -2.593667984008789,
|
| 85 |
+
"logits/rejected": -2.540072202682495,
|
| 86 |
+
"logps/chosen": -162.23658752441406,
|
| 87 |
+
"logps/rejected": -134.82162475585938,
|
| 88 |
+
"loss": 0.6925,
|
| 89 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 90 |
+
"rewards/chosen": 0.0019953851588070393,
|
| 91 |
+
"rewards/margins": 0.0013176979264244437,
|
| 92 |
+
"rewards/rejected": 0.0006776872323825955,
|
| 93 |
"step": 50
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.06,
|
| 97 |
"learning_rate": 2.9411764705882355e-06,
|
| 98 |
+
"logits/chosen": -2.5982372760772705,
|
| 99 |
+
"logits/rejected": -2.565171718597412,
|
| 100 |
+
"logps/chosen": -167.44467163085938,
|
| 101 |
+
"logps/rejected": -147.24746704101562,
|
| 102 |
+
"loss": 0.6927,
|
| 103 |
+
"rewards/accuracies": 0.612500011920929,
|
| 104 |
+
"rewards/chosen": 0.0029606518801301718,
|
| 105 |
+
"rewards/margins": 0.0008354202145710588,
|
| 106 |
+
"rewards/rejected": 0.002125231781974435,
|
| 107 |
"step": 60
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.07,
|
| 111 |
"learning_rate": 3.431372549019608e-06,
|
| 112 |
+
"logits/chosen": -2.6009891033172607,
|
| 113 |
+
"logits/rejected": -2.564056873321533,
|
| 114 |
+
"logps/chosen": -155.53407287597656,
|
| 115 |
+
"logps/rejected": -126.87095642089844,
|
| 116 |
+
"loss": 0.6924,
|
| 117 |
+
"rewards/accuracies": 0.625,
|
| 118 |
+
"rewards/chosen": 0.003734055906534195,
|
| 119 |
+
"rewards/margins": 0.0015770606696605682,
|
| 120 |
+
"rewards/rejected": 0.0021569952368736267,
|
| 121 |
"step": 70
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.08,
|
| 125 |
"learning_rate": 3.92156862745098e-06,
|
| 126 |
+
"logits/chosen": -2.549417018890381,
|
| 127 |
+
"logits/rejected": -2.4943110942840576,
|
| 128 |
+
"logps/chosen": -169.3980712890625,
|
| 129 |
+
"logps/rejected": -143.16995239257812,
|
| 130 |
+
"loss": 0.6915,
|
| 131 |
"rewards/accuracies": 0.637499988079071,
|
| 132 |
+
"rewards/chosen": 0.008090810850262642,
|
| 133 |
+
"rewards/margins": 0.003366248682141304,
|
| 134 |
+
"rewards/rejected": 0.004724562633782625,
|
| 135 |
"step": 80
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.09,
|
| 139 |
"learning_rate": 4.411764705882353e-06,
|
| 140 |
+
"logits/chosen": -2.497745990753174,
|
| 141 |
+
"logits/rejected": -2.445770740509033,
|
| 142 |
+
"logps/chosen": -143.51312255859375,
|
| 143 |
+
"logps/rejected": -118.7002944946289,
|
| 144 |
+
"loss": 0.6909,
|
| 145 |
+
"rewards/accuracies": 0.5625,
|
| 146 |
+
"rewards/chosen": 0.009854977950453758,
|
| 147 |
+
"rewards/margins": 0.004432962276041508,
|
| 148 |
+
"rewards/rejected": 0.0054220156744122505,
|
| 149 |
"step": 90
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.1,
|
| 153 |
"learning_rate": 4.901960784313726e-06,
|
| 154 |
+
"logits/chosen": -2.501483917236328,
|
| 155 |
+
"logits/rejected": -2.4084057807922363,
|
| 156 |
+
"logps/chosen": -180.1342010498047,
|
| 157 |
+
"logps/rejected": -141.8975372314453,
|
| 158 |
+
"loss": 0.6874,
|
| 159 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 160 |
+
"rewards/chosen": 0.021252702921628952,
|
| 161 |
+
"rewards/margins": 0.01164860837161541,
|
| 162 |
+
"rewards/rejected": 0.009604096412658691,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1,
|
| 167 |
+
"eval_logits/chosen": -2.484314203262329,
|
| 168 |
+
"eval_logits/rejected": -2.4462974071502686,
|
| 169 |
+
"eval_logps/chosen": -157.93946838378906,
|
| 170 |
+
"eval_logps/rejected": -132.1924285888672,
|
| 171 |
+
"eval_loss": 0.6892239451408386,
|
| 172 |
+
"eval_rewards/accuracies": 0.6698113083839417,
|
| 173 |
+
"eval_rewards/chosen": 0.021303629502654076,
|
| 174 |
+
"eval_rewards/margins": 0.008022623136639595,
|
| 175 |
+
"eval_rewards/rejected": 0.013281003572046757,
|
| 176 |
+
"eval_runtime": 285.1988,
|
| 177 |
+
"eval_samples_per_second": 1.466,
|
| 178 |
+
"eval_steps_per_second": 0.186,
|
| 179 |
"step": 100
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"epoch": 0.11,
|
| 183 |
"learning_rate": 4.9990549169459415e-06,
|
| 184 |
+
"logits/chosen": -2.605736494064331,
|
| 185 |
+
"logits/rejected": -2.551860809326172,
|
| 186 |
+
"logps/chosen": -160.9684600830078,
|
| 187 |
+
"logps/rejected": -134.45443725585938,
|
| 188 |
+
"loss": 0.6862,
|
| 189 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 190 |
+
"rewards/chosen": 0.029262090101838112,
|
| 191 |
+
"rewards/margins": 0.014033245854079723,
|
| 192 |
+
"rewards/rejected": 0.015228847041726112,
|
| 193 |
"step": 110
|
| 194 |
},
|
| 195 |
{
|
| 196 |
"epoch": 0.12,
|
| 197 |
"learning_rate": 4.995216741642263e-06,
|
| 198 |
+
"logits/chosen": -2.6044211387634277,
|
| 199 |
+
"logits/rejected": -2.5780460834503174,
|
| 200 |
+
"logps/chosen": -152.76980590820312,
|
| 201 |
+
"logps/rejected": -132.57339477539062,
|
| 202 |
+
"loss": 0.6886,
|
| 203 |
+
"rewards/accuracies": 0.550000011920929,
|
| 204 |
+
"rewards/chosen": 0.0384935587644577,
|
| 205 |
+
"rewards/margins": 0.00926611851900816,
|
| 206 |
+
"rewards/rejected": 0.02922743931412697,
|
| 207 |
"step": 120
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 0.13,
|
| 211 |
"learning_rate": 4.988430936991089e-06,
|
| 212 |
+
"logits/chosen": -2.594087600708008,
|
| 213 |
+
"logits/rejected": -2.5282979011535645,
|
| 214 |
+
"logps/chosen": -162.58688354492188,
|
| 215 |
+
"logps/rejected": -133.3597869873047,
|
| 216 |
+
"loss": 0.6804,
|
| 217 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 218 |
+
"rewards/chosen": 0.06138850376009941,
|
| 219 |
+
"rewards/margins": 0.02613685093820095,
|
| 220 |
+
"rewards/rejected": 0.03525165468454361,
|
| 221 |
"step": 130
|
| 222 |
},
|
| 223 |
{
|
| 224 |
"epoch": 0.14,
|
| 225 |
"learning_rate": 4.978705519144525e-06,
|
| 226 |
+
"logits/chosen": -2.5525660514831543,
|
| 227 |
+
"logits/rejected": -2.4592275619506836,
|
| 228 |
+
"logps/chosen": -174.51707458496094,
|
| 229 |
+
"logps/rejected": -135.41912841796875,
|
| 230 |
+
"loss": 0.6759,
|
| 231 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 232 |
+
"rewards/chosen": 0.06431619822978973,
|
| 233 |
+
"rewards/margins": 0.03535788506269455,
|
| 234 |
+
"rewards/rejected": 0.028958311304450035,
|
| 235 |
"step": 140
|
| 236 |
},
|
| 237 |
{
|
| 238 |
"epoch": 0.15,
|
| 239 |
"learning_rate": 4.966051976854862e-06,
|
| 240 |
+
"logits/chosen": -2.6359076499938965,
|
| 241 |
+
"logits/rejected": -2.5874571800231934,
|
| 242 |
+
"logps/chosen": -144.5331268310547,
|
| 243 |
+
"logps/rejected": -122.17633056640625,
|
| 244 |
+
"loss": 0.6729,
|
| 245 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 246 |
+
"rewards/chosen": 0.08950608223676682,
|
| 247 |
+
"rewards/margins": 0.042167335748672485,
|
| 248 |
+
"rewards/rejected": 0.04733874648809433,
|
| 249 |
"step": 150
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.16,
|
| 253 |
"learning_rate": 4.950485257902782e-06,
|
| 254 |
+
"logits/chosen": -2.555276393890381,
|
| 255 |
+
"logits/rejected": -2.523134708404541,
|
| 256 |
+
"logps/chosen": -152.33584594726562,
|
| 257 |
+
"logps/rejected": -128.93971252441406,
|
| 258 |
+
"loss": 0.671,
|
| 259 |
+
"rewards/accuracies": 0.762499988079071,
|
| 260 |
+
"rewards/chosen": 0.10823699086904526,
|
| 261 |
+
"rewards/margins": 0.04584677889943123,
|
| 262 |
+
"rewards/rejected": 0.06239022687077522,
|
| 263 |
"step": 160
|
| 264 |
},
|
| 265 |
{
|
| 266 |
"epoch": 0.17,
|
| 267 |
"learning_rate": 4.932023751439358e-06,
|
| 268 |
+
"logits/chosen": -2.6216721534729004,
|
| 269 |
+
"logits/rejected": -2.5455801486968994,
|
| 270 |
+
"logps/chosen": -162.0784454345703,
|
| 271 |
+
"logps/rejected": -133.4407501220703,
|
| 272 |
+
"loss": 0.6717,
|
| 273 |
+
"rewards/accuracies": 0.675000011920929,
|
| 274 |
+
"rewards/chosen": 0.11643824726343155,
|
| 275 |
+
"rewards/margins": 0.0462535098195076,
|
| 276 |
+
"rewards/rejected": 0.07018472254276276,
|
| 277 |
"step": 170
|
| 278 |
},
|
| 279 |
{
|
| 280 |
"epoch": 0.18,
|
| 281 |
"learning_rate": 4.9106892662627395e-06,
|
| 282 |
+
"logits/chosen": -2.6522557735443115,
|
| 283 |
+
"logits/rejected": -2.5888028144836426,
|
| 284 |
+
"logps/chosen": -158.68748474121094,
|
| 285 |
+
"logps/rejected": -135.9111785888672,
|
| 286 |
+
"loss": 0.6686,
|
| 287 |
+
"rewards/accuracies": 0.6875,
|
| 288 |
+
"rewards/chosen": 0.11335412412881851,
|
| 289 |
+
"rewards/margins": 0.05232863873243332,
|
| 290 |
+
"rewards/rejected": 0.06102548912167549,
|
| 291 |
"step": 180
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 0.19,
|
| 295 |
"learning_rate": 4.886507005055149e-06,
|
| 296 |
+
"logits/chosen": -2.640076160430908,
|
| 297 |
+
"logits/rejected": -2.545991897583008,
|
| 298 |
+
"logps/chosen": -167.3929443359375,
|
| 299 |
+
"logps/rejected": -137.95436096191406,
|
| 300 |
+
"loss": 0.6509,
|
| 301 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 302 |
+
"rewards/chosen": 0.10093629360198975,
|
| 303 |
+
"rewards/margins": 0.09009988605976105,
|
| 304 |
+
"rewards/rejected": 0.010836412198841572,
|
| 305 |
"step": 190
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 0.2,
|
| 309 |
"learning_rate": 4.859505534610658e-06,
|
| 310 |
+
"logits/chosen": -2.614877700805664,
|
| 311 |
+
"logits/rejected": -2.576866865158081,
|
| 312 |
+
"logps/chosen": -154.33758544921875,
|
| 313 |
+
"logps/rejected": -138.316650390625,
|
| 314 |
+
"loss": 0.6592,
|
| 315 |
+
"rewards/accuracies": 0.625,
|
| 316 |
+
"rewards/chosen": 0.017874887213110924,
|
| 317 |
+
"rewards/margins": 0.0757044330239296,
|
| 318 |
+
"rewards/rejected": -0.05782955139875412,
|
| 319 |
"step": 200
|
| 320 |
},
|
| 321 |
{
|
| 322 |
"epoch": 0.2,
|
| 323 |
+
"eval_logits/chosen": -2.5370168685913086,
|
| 324 |
+
"eval_logits/rejected": -2.4922032356262207,
|
| 325 |
+
"eval_logps/chosen": -159.51800537109375,
|
| 326 |
+
"eval_logps/rejected": -140.55880737304688,
|
| 327 |
+
"eval_loss": 0.6593995690345764,
|
| 328 |
+
"eval_rewards/accuracies": 0.6698113083839417,
|
| 329 |
+
"eval_rewards/chosen": 0.00551834749057889,
|
| 330 |
+
"eval_rewards/margins": 0.07590115815401077,
|
| 331 |
+
"eval_rewards/rejected": -0.07038279622793198,
|
| 332 |
+
"eval_runtime": 284.6148,
|
| 333 |
+
"eval_samples_per_second": 1.469,
|
| 334 |
+
"eval_steps_per_second": 0.186,
|
| 335 |
"step": 200
|
| 336 |
},
|
| 337 |
{
|
| 338 |
"epoch": 0.21,
|
| 339 |
"learning_rate": 4.829716752088893e-06,
|
| 340 |
+
"logits/chosen": -2.55920147895813,
|
| 341 |
+
"logits/rejected": -2.5236589908599854,
|
| 342 |
+
"logps/chosen": -148.85671997070312,
|
| 343 |
+
"logps/rejected": -144.40414428710938,
|
| 344 |
+
"loss": 0.6669,
|
| 345 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 346 |
+
"rewards/chosen": -0.021094869822263718,
|
| 347 |
+
"rewards/margins": 0.05866061523556709,
|
| 348 |
+
"rewards/rejected": -0.07975548505783081,
|
| 349 |
"step": 210
|
| 350 |
},
|
| 351 |
{
|
| 352 |
"epoch": 0.22,
|
| 353 |
"learning_rate": 4.797175847334535e-06,
|
| 354 |
+
"logits/chosen": -2.614175319671631,
|
| 355 |
+
"logits/rejected": -2.5591626167297363,
|
| 356 |
+
"logps/chosen": -173.6731414794922,
|
| 357 |
+
"logps/rejected": -153.340087890625,
|
| 358 |
+
"loss": 0.6663,
|
| 359 |
+
"rewards/accuracies": 0.637499988079071,
|
| 360 |
+
"rewards/chosen": -0.06277570873498917,
|
| 361 |
+
"rewards/margins": 0.06162955239415169,
|
| 362 |
+
"rewards/rejected": -0.12440524995326996,
|
| 363 |
"step": 220
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 0.23,
|
| 367 |
"learning_rate": 4.761921261307143e-06,
|
| 368 |
+
"logits/chosen": -2.6242852210998535,
|
| 369 |
+
"logits/rejected": -2.5607199668884277,
|
| 370 |
+
"logps/chosen": -167.7920684814453,
|
| 371 |
+
"logps/rejected": -153.7227020263672,
|
| 372 |
+
"loss": 0.6412,
|
| 373 |
+
"rewards/accuracies": 0.699999988079071,
|
| 374 |
+
"rewards/chosen": -0.058848392218351364,
|
| 375 |
+
"rewards/margins": 0.12100015580654144,
|
| 376 |
+
"rewards/rejected": -0.1798485666513443,
|
| 377 |
"step": 230
|
| 378 |
},
|
| 379 |
{
|
| 380 |
"epoch": 0.24,
|
| 381 |
"learning_rate": 4.723994640670377e-06,
|
| 382 |
+
"logits/chosen": -2.6291072368621826,
|
| 383 |
+
"logits/rejected": -2.528074026107788,
|
| 384 |
+
"logps/chosen": -179.40382385253906,
|
| 385 |
+
"logps/rejected": -153.91192626953125,
|
| 386 |
+
"loss": 0.6197,
|
| 387 |
+
"rewards/accuracies": 0.824999988079071,
|
| 388 |
+
"rewards/chosen": -0.051289211958646774,
|
| 389 |
+
"rewards/margins": 0.1633162945508957,
|
| 390 |
+
"rewards/rejected": -0.21460552513599396,
|
| 391 |
"step": 240
|
| 392 |
},
|
| 393 |
{
|
| 394 |
"epoch": 0.25,
|
| 395 |
"learning_rate": 4.68344078859431e-06,
|
| 396 |
+
"logits/chosen": -2.5995261669158936,
|
| 397 |
+
"logits/rejected": -2.5724520683288574,
|
| 398 |
+
"logps/chosen": -163.8385009765625,
|
| 399 |
+
"logps/rejected": -167.84129333496094,
|
| 400 |
+
"loss": 0.6333,
|
| 401 |
+
"rewards/accuracies": 0.675000011920929,
|
| 402 |
+
"rewards/chosen": -0.11021089553833008,
|
| 403 |
+
"rewards/margins": 0.13913698494434357,
|
| 404 |
+
"rewards/rejected": -0.24934788048267365,
|
| 405 |
"step": 250
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 0.26,
|
| 409 |
"learning_rate": 4.6403076118289006e-06,
|
| 410 |
+
"logits/chosen": -2.5930025577545166,
|
| 411 |
+
"logits/rejected": -2.5077426433563232,
|
| 412 |
+
"logps/chosen": -182.69297790527344,
|
| 413 |
+
"logps/rejected": -169.84852600097656,
|
| 414 |
+
"loss": 0.6071,
|
| 415 |
+
"rewards/accuracies": 0.8125,
|
| 416 |
+
"rewards/chosen": -0.1480914056301117,
|
| 417 |
+
"rewards/margins": 0.20134063065052032,
|
| 418 |
+
"rewards/rejected": -0.3494320213794708,
|
| 419 |
"step": 260
|
| 420 |
},
|
| 421 |
{
|
| 422 |
"epoch": 0.27,
|
| 423 |
"learning_rate": 4.5946460641111776e-06,
|
| 424 |
+
"logits/chosen": -2.7093770503997803,
|
| 425 |
+
"logits/rejected": -2.577432632446289,
|
| 426 |
+
"logps/chosen": -188.6643524169922,
|
| 427 |
+
"logps/rejected": -187.13156127929688,
|
| 428 |
+
"loss": 0.5669,
|
| 429 |
+
"rewards/accuracies": 0.8125,
|
| 430 |
+
"rewards/chosen": -0.23154917359352112,
|
| 431 |
+
"rewards/margins": 0.31018704175949097,
|
| 432 |
+
"rewards/rejected": -0.5417362451553345,
|
| 433 |
"step": 270
|
| 434 |
},
|
| 435 |
{
|
| 436 |
"epoch": 0.28,
|
| 437 |
"learning_rate": 4.546510085972983e-06,
|
| 438 |
+
"logits/chosen": -2.679447650909424,
|
| 439 |
+
"logits/rejected": -2.6179795265197754,
|
| 440 |
+
"logps/chosen": -211.7890625,
|
| 441 |
+
"logps/rejected": -199.5372772216797,
|
| 442 |
+
"loss": 0.6167,
|
| 443 |
+
"rewards/accuracies": 0.699999988079071,
|
| 444 |
+
"rewards/chosen": -0.38407862186431885,
|
| 445 |
+
"rewards/margins": 0.20160210132598877,
|
| 446 |
+
"rewards/rejected": -0.5856807231903076,
|
| 447 |
"step": 280
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 0.29,
|
| 451 |
"learning_rate": 4.495956541020376e-06,
|
| 452 |
+
"logits/chosen": -2.7100636959075928,
|
| 453 |
+
"logits/rejected": -2.6252739429473877,
|
| 454 |
+
"logps/chosen": -222.813720703125,
|
| 455 |
+
"logps/rejected": -226.9723663330078,
|
| 456 |
+
"loss": 0.5781,
|
| 457 |
+
"rewards/accuracies": 0.75,
|
| 458 |
+
"rewards/chosen": -0.45635899901390076,
|
| 459 |
+
"rewards/margins": 0.3029539883136749,
|
| 460 |
+
"rewards/rejected": -0.7593129873275757,
|
| 461 |
"step": 290
|
| 462 |
},
|
| 463 |
{
|
| 464 |
"epoch": 0.3,
|
| 465 |
"learning_rate": 4.443045148759978e-06,
|
| 466 |
+
"logits/chosen": -2.6889595985412598,
|
| 467 |
+
"logits/rejected": -2.5976881980895996,
|
| 468 |
+
"logps/chosen": -219.4752960205078,
|
| 469 |
+
"logps/rejected": -215.431396484375,
|
| 470 |
+
"loss": 0.5451,
|
| 471 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 472 |
+
"rewards/chosen": -0.3557833135128021,
|
| 473 |
+
"rewards/margins": 0.4194447100162506,
|
| 474 |
+
"rewards/rejected": -0.7752280831336975,
|
| 475 |
+
"step": 300
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.3,
|
| 479 |
+
"eval_logits/chosen": -2.561957359313965,
|
| 480 |
+
"eval_logits/rejected": -2.512842893600464,
|
| 481 |
+
"eval_logps/chosen": -204.9713134765625,
|
| 482 |
+
"eval_logps/rejected": -209.39376831054688,
|
| 483 |
+
"eval_loss": 0.5867109298706055,
|
| 484 |
+
"eval_rewards/accuracies": 0.6863207817077637,
|
| 485 |
+
"eval_rewards/chosen": -0.44901466369628906,
|
| 486 |
+
"eval_rewards/margins": 0.30971792340278625,
|
| 487 |
+
"eval_rewards/rejected": -0.7587326169013977,
|
| 488 |
+
"eval_runtime": 286.3596,
|
| 489 |
+
"eval_samples_per_second": 1.46,
|
| 490 |
+
"eval_steps_per_second": 0.185,
|
| 491 |
"step": 300
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 0.3,
|
| 495 |
"learning_rate": 4.3878384140516025e-06,
|
| 496 |
+
"logits/chosen": -2.6861400604248047,
|
| 497 |
+
"logits/rejected": -2.6122899055480957,
|
| 498 |
+
"logps/chosen": -209.1049346923828,
|
| 499 |
+
"logps/rejected": -223.94308471679688,
|
| 500 |
+
"loss": 0.5685,
|
| 501 |
+
"rewards/accuracies": 0.675000011920929,
|
| 502 |
+
"rewards/chosen": -0.4111746847629547,
|
| 503 |
+
"rewards/margins": 0.355561226606369,
|
| 504 |
+
"rewards/rejected": -0.7667359113693237,
|
| 505 |
"step": 310
|
| 506 |
},
|
| 507 |
{
|
| 508 |
"epoch": 0.31,
|
| 509 |
"learning_rate": 4.330401553270522e-06,
|
| 510 |
+
"logits/chosen": -2.649998188018799,
|
| 511 |
+
"logits/rejected": -2.590202569961548,
|
| 512 |
+
"logps/chosen": -201.97621154785156,
|
| 513 |
+
"logps/rejected": -211.2521209716797,
|
| 514 |
+
"loss": 0.5949,
|
| 515 |
+
"rewards/accuracies": 0.6875,
|
| 516 |
+
"rewards/chosen": -0.3732680678367615,
|
| 517 |
+
"rewards/margins": 0.3101999759674072,
|
| 518 |
+
"rewards/rejected": -0.6834679841995239,
|
| 519 |
"step": 320
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 0.32,
|
| 523 |
"learning_rate": 4.2708024172665795e-06,
|
| 524 |
+
"logits/chosen": -2.7253522872924805,
|
| 525 |
+
"logits/rejected": -2.6092331409454346,
|
| 526 |
+
"logps/chosen": -201.93325805664062,
|
| 527 |
+
"logps/rejected": -203.0877685546875,
|
| 528 |
+
"loss": 0.5527,
|
| 529 |
+
"rewards/accuracies": 0.737500011920929,
|
| 530 |
+
"rewards/chosen": -0.35112464427948,
|
| 531 |
+
"rewards/margins": 0.431334912776947,
|
| 532 |
+
"rewards/rejected": -0.7824596166610718,
|
| 533 |
"step": 330
|
| 534 |
},
|
| 535 |
{
|
| 536 |
"epoch": 0.33,
|
| 537 |
"learning_rate": 4.209111411211174e-06,
|
| 538 |
+
"logits/chosen": -2.7140471935272217,
|
| 539 |
+
"logits/rejected": -2.611151933670044,
|
| 540 |
+
"logps/chosen": -198.9755096435547,
|
| 541 |
+
"logps/rejected": -215.7385711669922,
|
| 542 |
+
"loss": 0.5351,
|
| 543 |
"rewards/accuracies": 0.75,
|
| 544 |
+
"rewards/chosen": -0.41874709725379944,
|
| 545 |
+
"rewards/margins": 0.46593284606933594,
|
| 546 |
+
"rewards/rejected": -0.8846799731254578,
|
| 547 |
"step": 340
|
| 548 |
},
|
| 549 |
{
|
| 550 |
"epoch": 0.34,
|
| 551 |
"learning_rate": 4.145401411426788e-06,
|
| 552 |
+
"logits/chosen": -2.7166478633880615,
|
| 553 |
+
"logits/rejected": -2.650458574295044,
|
| 554 |
+
"logps/chosen": -201.68899536132812,
|
| 555 |
+
"logps/rejected": -224.9377899169922,
|
| 556 |
+
"loss": 0.5406,
|
| 557 |
+
"rewards/accuracies": 0.762499988079071,
|
| 558 |
+
"rewards/chosen": -0.44746604561805725,
|
| 559 |
+
"rewards/margins": 0.4966231882572174,
|
| 560 |
+
"rewards/rejected": -0.9440892934799194,
|
| 561 |
"step": 350
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 0.35,
|
| 565 |
"learning_rate": 4.079747679297314e-06,
|
| 566 |
+
"logits/chosen": -2.631408214569092,
|
| 567 |
+
"logits/rejected": -2.56895112991333,
|
| 568 |
+
"logps/chosen": -220.35415649414062,
|
| 569 |
+
"logps/rejected": -229.13998413085938,
|
| 570 |
+
"loss": 0.5639,
|
| 571 |
+
"rewards/accuracies": 0.6875,
|
| 572 |
+
"rewards/chosen": -0.5187793970108032,
|
| 573 |
+
"rewards/margins": 0.4147488474845886,
|
| 574 |
+
"rewards/rejected": -0.9335282444953918,
|
| 575 |
"step": 360
|
| 576 |
},
|
| 577 |
{
|
| 578 |
"epoch": 0.36,
|
| 579 |
"learning_rate": 4.012227772360889e-06,
|
| 580 |
+
"logits/chosen": -2.64164662361145,
|
| 581 |
+
"logits/rejected": -2.554852247238159,
|
| 582 |
+
"logps/chosen": -224.2294158935547,
|
| 583 |
+
"logps/rejected": -246.3839111328125,
|
| 584 |
+
"loss": 0.5202,
|
| 585 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 586 |
+
"rewards/chosen": -0.502095103263855,
|
| 587 |
+
"rewards/margins": 0.5576826930046082,
|
| 588 |
+
"rewards/rejected": -1.0597779750823975,
|
| 589 |
"step": 370
|
| 590 |
},
|
| 591 |
{
|
| 592 |
"epoch": 0.37,
|
| 593 |
"learning_rate": 3.942921452690245e-06,
|
| 594 |
+
"logits/chosen": -2.6596052646636963,
|
| 595 |
+
"logits/rejected": -2.597478151321411,
|
| 596 |
+
"logps/chosen": -226.8369140625,
|
| 597 |
+
"logps/rejected": -239.9619140625,
|
| 598 |
+
"loss": 0.6339,
|
| 599 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 600 |
+
"rewards/chosen": -0.5941166877746582,
|
| 601 |
+
"rewards/margins": 0.3102978467941284,
|
| 602 |
+
"rewards/rejected": -0.9044146537780762,
|
| 603 |
"step": 380
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 0.38,
|
| 607 |
"learning_rate": 3.871910592668817e-06,
|
| 608 |
+
"logits/chosen": -2.70902156829834,
|
| 609 |
+
"logits/rejected": -2.634817123413086,
|
| 610 |
+
"logps/chosen": -222.9804229736328,
|
| 611 |
+
"logps/rejected": -253.72097778320312,
|
| 612 |
+
"loss": 0.5281,
|
| 613 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 614 |
+
"rewards/chosen": -0.6194333434104919,
|
| 615 |
+
"rewards/margins": 0.5652093887329102,
|
| 616 |
+
"rewards/rejected": -1.1846425533294678,
|
| 617 |
"step": 390
|
| 618 |
},
|
| 619 |
{
|
| 620 |
"epoch": 0.39,
|
| 621 |
"learning_rate": 3.799279078273921e-06,
|
| 622 |
+
"logits/chosen": -2.6576032638549805,
|
| 623 |
+
"logits/rejected": -2.552799940109253,
|
| 624 |
+
"logps/chosen": -214.78018188476562,
|
| 625 |
+
"logps/rejected": -235.85055541992188,
|
| 626 |
+
"loss": 0.4933,
|
| 627 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 628 |
+
"rewards/chosen": -0.48761433362960815,
|
| 629 |
+
"rewards/margins": 0.6489143371582031,
|
| 630 |
+
"rewards/rejected": -1.136528730392456,
|
| 631 |
"step": 400
|
| 632 |
},
|
| 633 |
{
|
| 634 |
"epoch": 0.39,
|
| 635 |
+
"eval_logits/chosen": -2.5385656356811523,
|
| 636 |
+
"eval_logits/rejected": -2.486776113510132,
|
| 637 |
+
"eval_logps/chosen": -220.6713409423828,
|
| 638 |
+
"eval_logps/rejected": -243.80616760253906,
|
| 639 |
+
"eval_loss": 0.5590807795524597,
|
| 640 |
+
"eval_rewards/accuracies": 0.7146226167678833,
|
| 641 |
+
"eval_rewards/chosen": -0.6060149669647217,
|
| 642 |
+
"eval_rewards/margins": 0.4968414306640625,
|
| 643 |
+
"eval_rewards/rejected": -1.1028562784194946,
|
| 644 |
+
"eval_runtime": 285.6623,
|
| 645 |
+
"eval_samples_per_second": 1.463,
|
| 646 |
+
"eval_steps_per_second": 0.186,
|
| 647 |
"step": 400
|
| 648 |
},
|
| 649 |
{
|
| 650 |
"epoch": 0.4,
|
| 651 |
"learning_rate": 3.725112709981249e-06,
|
| 652 |
+
"logits/chosen": -2.5486836433410645,
|
| 653 |
+
"logits/rejected": -2.4727280139923096,
|
| 654 |
+
"logps/chosen": -231.33547973632812,
|
| 655 |
+
"logps/rejected": -252.45864868164062,
|
| 656 |
+
"loss": 0.5563,
|
| 657 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 658 |
+
"rewards/chosen": -0.5974920988082886,
|
| 659 |
+
"rewards/margins": 0.4878315329551697,
|
| 660 |
+
"rewards/rejected": -1.0853235721588135,
|
| 661 |
"step": 410
|
| 662 |
},
|
| 663 |
{
|
| 664 |
"epoch": 0.41,
|
| 665 |
"learning_rate": 3.649499101407737e-06,
|
| 666 |
+
"logits/chosen": -2.651151180267334,
|
| 667 |
+
"logits/rejected": -2.556567430496216,
|
| 668 |
+
"logps/chosen": -241.06167602539062,
|
| 669 |
+
"logps/rejected": -253.01986694335938,
|
| 670 |
+
"loss": 0.6047,
|
| 671 |
"rewards/accuracies": 0.6875,
|
| 672 |
+
"rewards/chosen": -0.6971057057380676,
|
| 673 |
+
"rewards/margins": 0.3935045599937439,
|
| 674 |
+
"rewards/rejected": -1.0906102657318115,
|
| 675 |
"step": 420
|
| 676 |
},
|
| 677 |
{
|
| 678 |
"epoch": 0.42,
|
| 679 |
"learning_rate": 3.5725275758125564e-06,
|
| 680 |
+
"logits/chosen": -2.616089105606079,
|
| 681 |
+
"logits/rejected": -2.516278028488159,
|
| 682 |
+
"logps/chosen": -213.630859375,
|
| 683 |
+
"logps/rejected": -243.5413818359375,
|
| 684 |
+
"loss": 0.5115,
|
| 685 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 686 |
+
"rewards/chosen": -0.6288307905197144,
|
| 687 |
+
"rewards/margins": 0.6640511751174927,
|
| 688 |
+
"rewards/rejected": -1.2928818464279175,
|
| 689 |
"step": 430
|
| 690 |
},
|
| 691 |
{
|
| 692 |
"epoch": 0.43,
|
| 693 |
"learning_rate": 3.494289060578478e-06,
|
| 694 |
+
"logits/chosen": -2.6611900329589844,
|
| 695 |
+
"logits/rejected": -2.609663486480713,
|
| 696 |
+
"logps/chosen": -239.8604736328125,
|
| 697 |
+
"logps/rejected": -275.2757873535156,
|
| 698 |
+
"loss": 0.5247,
|
| 699 |
+
"rewards/accuracies": 0.75,
|
| 700 |
+
"rewards/chosen": -0.8715305328369141,
|
| 701 |
+
"rewards/margins": 0.5755683779716492,
|
| 702 |
+
"rewards/rejected": -1.447098970413208,
|
| 703 |
"step": 440
|
| 704 |
},
|
| 705 |
{
|
| 706 |
"epoch": 0.44,
|
| 707 |
"learning_rate": 3.414875979798272e-06,
|
| 708 |
+
"logits/chosen": -2.640732526779175,
|
| 709 |
+
"logits/rejected": -2.5787458419799805,
|
| 710 |
+
"logps/chosen": -237.64920043945312,
|
| 711 |
+
"logps/rejected": -276.0288391113281,
|
| 712 |
+
"loss": 0.4854,
|
| 713 |
+
"rewards/accuracies": 0.699999988079071,
|
| 714 |
+
"rewards/chosen": -0.5312285423278809,
|
| 715 |
+
"rewards/margins": 0.7248728275299072,
|
| 716 |
+
"rewards/rejected": -1.256101369857788,
|
| 717 |
"step": 450
|
| 718 |
},
|
| 719 |
{
|
| 720 |
"epoch": 0.45,
|
| 721 |
"learning_rate": 3.3343821450930196e-06,
|
| 722 |
+
"logits/chosen": -2.7021923065185547,
|
| 723 |
+
"logits/rejected": -2.6052348613739014,
|
| 724 |
+
"logps/chosen": -231.91165161132812,
|
| 725 |
+
"logps/rejected": -256.6305847167969,
|
| 726 |
+
"loss": 0.5014,
|
| 727 |
+
"rewards/accuracies": 0.762499988079071,
|
| 728 |
+
"rewards/chosen": -0.6199955940246582,
|
| 729 |
+
"rewards/margins": 0.6785587072372437,
|
| 730 |
+
"rewards/rejected": -1.2985544204711914,
|
| 731 |
"step": 460
|
| 732 |
},
|
| 733 |
{
|
| 734 |
"epoch": 0.46,
|
| 735 |
"learning_rate": 3.252902644791325e-06,
|
| 736 |
+
"logits/chosen": -2.5665388107299805,
|
| 737 |
+
"logits/rejected": -2.487617015838623,
|
| 738 |
+
"logps/chosen": -232.90200805664062,
|
| 739 |
+
"logps/rejected": -276.9921569824219,
|
| 740 |
+
"loss": 0.4943,
|
| 741 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 742 |
+
"rewards/chosen": -0.6983159780502319,
|
| 743 |
+
"rewards/margins": 0.7069972157478333,
|
| 744 |
+
"rewards/rejected": -1.40531325340271,
|
| 745 |
"step": 470
|
| 746 |
},
|
| 747 |
{
|
| 748 |
"epoch": 0.47,
|
| 749 |
"learning_rate": 3.170533731600339e-06,
|
| 750 |
+
"logits/chosen": -2.618633985519409,
|
| 751 |
+
"logits/rejected": -2.5317630767822266,
|
| 752 |
+
"logps/chosen": -239.92919921875,
|
| 753 |
+
"logps/rejected": -261.753173828125,
|
| 754 |
+
"loss": 0.543,
|
| 755 |
+
"rewards/accuracies": 0.699999988079071,
|
| 756 |
+
"rewards/chosen": -0.7530533075332642,
|
| 757 |
+
"rewards/margins": 0.5510448217391968,
|
| 758 |
+
"rewards/rejected": -1.304098129272461,
|
| 759 |
"step": 480
|
| 760 |
},
|
| 761 |
{
|
| 762 |
"epoch": 0.48,
|
| 763 |
"learning_rate": 3.0873727089012816e-06,
|
| 764 |
+
"logits/chosen": -2.7126545906066895,
|
| 765 |
+
"logits/rejected": -2.6558868885040283,
|
| 766 |
+
"logps/chosen": -242.5832977294922,
|
| 767 |
+
"logps/rejected": -258.93365478515625,
|
| 768 |
+
"loss": 0.5719,
|
| 769 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 770 |
+
"rewards/chosen": -0.7276023626327515,
|
| 771 |
+
"rewards/margins": 0.48720917105674744,
|
| 772 |
+
"rewards/rejected": -1.2148113250732422,
|
| 773 |
"step": 490
|
| 774 |
},
|
| 775 |
{
|
| 776 |
"epoch": 0.49,
|
| 777 |
"learning_rate": 3.0035178158038026e-06,
|
| 778 |
+
"logits/chosen": -2.594712018966675,
|
| 779 |
+
"logits/rejected": -2.4889943599700928,
|
| 780 |
+
"logps/chosen": -238.0956268310547,
|
| 781 |
+
"logps/rejected": -263.0235290527344,
|
| 782 |
+
"loss": 0.5271,
|
| 783 |
+
"rewards/accuracies": 0.75,
|
| 784 |
+
"rewards/chosen": -0.6214576959609985,
|
| 785 |
+
"rewards/margins": 0.6590222716331482,
|
| 786 |
+
"rewards/rejected": -1.280479907989502,
|
| 787 |
+
"step": 500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.49,
|
| 791 |
+
"eval_logits/chosen": -2.531158924102783,
|
| 792 |
+
"eval_logits/rejected": -2.4784035682678223,
|
| 793 |
+
"eval_logps/chosen": -227.1888885498047,
|
| 794 |
+
"eval_logps/rejected": -260.8957824707031,
|
| 795 |
+
"eval_loss": 0.5488241314888,
|
| 796 |
+
"eval_rewards/accuracies": 0.7193396091461182,
|
| 797 |
+
"eval_rewards/chosen": -0.6711904406547546,
|
| 798 |
+
"eval_rewards/margins": 0.6025623083114624,
|
| 799 |
+
"eval_rewards/rejected": -1.2737528085708618,
|
| 800 |
+
"eval_runtime": 286.5537,
|
| 801 |
+
"eval_samples_per_second": 1.459,
|
| 802 |
+
"eval_steps_per_second": 0.185,
|
| 803 |
"step": 500
|
| 804 |
},
|
| 805 |
{
|
| 806 |
"epoch": 0.5,
|
| 807 |
"learning_rate": 2.919068111094937e-06,
|
| 808 |
+
"logits/chosen": -2.6373202800750732,
|
| 809 |
+
"logits/rejected": -2.5553860664367676,
|
| 810 |
+
"logps/chosen": -229.17578125,
|
| 811 |
+
"logps/rejected": -268.92559814453125,
|
| 812 |
+
"loss": 0.5683,
|
| 813 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 814 |
+
"rewards/chosen": -0.7968514561653137,
|
| 815 |
+
"rewards/margins": 0.5847809314727783,
|
| 816 |
+
"rewards/rejected": -1.3816323280334473,
|
| 817 |
"step": 510
|
| 818 |
},
|
| 819 |
{
|
| 820 |
"epoch": 0.51,
|
| 821 |
"learning_rate": 2.8341233562197895e-06,
|
| 822 |
+
"logits/chosen": -2.64227032661438,
|
| 823 |
+
"logits/rejected": -2.599839687347412,
|
| 824 |
+
"logps/chosen": -215.4423828125,
|
| 825 |
+
"logps/rejected": -253.928955078125,
|
| 826 |
+
"loss": 0.5432,
|
| 827 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 828 |
+
"rewards/chosen": -0.6491907835006714,
|
| 829 |
+
"rewards/margins": 0.5866181254386902,
|
| 830 |
+
"rewards/rejected": -1.235809087753296,
|
| 831 |
"step": 520
|
| 832 |
},
|
| 833 |
{
|
| 834 |
"epoch": 0.52,
|
| 835 |
"learning_rate": 2.7487838974321352e-06,
|
| 836 |
+
"logits/chosen": -2.613891124725342,
|
| 837 |
+
"logits/rejected": -2.5329458713531494,
|
| 838 |
+
"logps/chosen": -212.8946533203125,
|
| 839 |
+
"logps/rejected": -254.96609497070312,
|
| 840 |
+
"loss": 0.5181,
|
| 841 |
+
"rewards/accuracies": 0.824999988079071,
|
| 842 |
+
"rewards/chosen": -0.5764448046684265,
|
| 843 |
+
"rewards/margins": 0.6870908737182617,
|
| 844 |
+
"rewards/rejected": -1.263535737991333,
|
| 845 |
"step": 530
|
| 846 |
},
|
| 847 |
{
|
| 848 |
"epoch": 0.53,
|
| 849 |
"learning_rate": 2.6631505472541997e-06,
|
| 850 |
+
"logits/chosen": -2.60998272895813,
|
| 851 |
+
"logits/rejected": -2.516542911529541,
|
| 852 |
+
"logps/chosen": -222.1146697998047,
|
| 853 |
+
"logps/rejected": -263.4886779785156,
|
| 854 |
+
"loss": 0.492,
|
| 855 |
+
"rewards/accuracies": 0.800000011920929,
|
| 856 |
+
"rewards/chosen": -0.549847424030304,
|
| 857 |
+
"rewards/margins": 0.7145514488220215,
|
| 858 |
+
"rewards/rejected": -1.2643989324569702,
|
| 859 |
"step": 540
|
| 860 |
},
|
| 861 |
{
|
| 862 |
"epoch": 0.54,
|
| 863 |
"learning_rate": 2.5773244653856173e-06,
|
| 864 |
+
"logits/chosen": -2.579040050506592,
|
| 865 |
+
"logits/rejected": -2.4950509071350098,
|
| 866 |
+
"logps/chosen": -228.0793914794922,
|
| 867 |
+
"logps/rejected": -273.71649169921875,
|
| 868 |
+
"loss": 0.4673,
|
| 869 |
+
"rewards/accuracies": 0.75,
|
| 870 |
+
"rewards/chosen": -0.5778298377990723,
|
| 871 |
+
"rewards/margins": 0.7735764980316162,
|
| 872 |
+
"rewards/rejected": -1.3514063358306885,
|
| 873 |
"step": 550
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 0.55,
|
| 877 |
"learning_rate": 2.4914070392022717e-06,
|
| 878 |
+
"logits/chosen": -2.612720251083374,
|
| 879 |
+
"logits/rejected": -2.5363128185272217,
|
| 880 |
+
"logps/chosen": -247.00271606445312,
|
| 881 |
+
"logps/rejected": -277.1578369140625,
|
| 882 |
+
"loss": 0.5677,
|
| 883 |
+
"rewards/accuracies": 0.737500011920929,
|
| 884 |
+
"rewards/chosen": -0.7081204652786255,
|
| 885 |
+
"rewards/margins": 0.5781086087226868,
|
| 886 |
+
"rewards/rejected": -1.286229133605957,
|
| 887 |
"step": 560
|
| 888 |
},
|
| 889 |
{
|
| 890 |
"epoch": 0.56,
|
| 891 |
"learning_rate": 2.4054997639861778e-06,
|
| 892 |
+
"logits/chosen": -2.5522828102111816,
|
| 893 |
+
"logits/rejected": -2.4502105712890625,
|
| 894 |
+
"logps/chosen": -244.18655395507812,
|
| 895 |
+
"logps/rejected": -291.7619934082031,
|
| 896 |
+
"loss": 0.4671,
|
| 897 |
+
"rewards/accuracies": 0.824999988079071,
|
| 898 |
+
"rewards/chosen": -0.685632050037384,
|
| 899 |
+
"rewards/margins": 0.871982753276825,
|
| 900 |
+
"rewards/rejected": -1.557614803314209,
|
| 901 |
"step": 570
|
| 902 |
},
|
| 903 |
{
|
| 904 |
"epoch": 0.57,
|
| 905 |
"learning_rate": 2.3197041230278905e-06,
|
| 906 |
+
"logits/chosen": -2.607151508331299,
|
| 907 |
+
"logits/rejected": -2.5367910861968994,
|
| 908 |
+
"logps/chosen": -242.89242553710938,
|
| 909 |
+
"logps/rejected": -303.93426513671875,
|
| 910 |
+
"loss": 0.4696,
|
| 911 |
+
"rewards/accuracies": 0.800000011920929,
|
| 912 |
+
"rewards/chosen": -0.7322829961776733,
|
| 913 |
+
"rewards/margins": 0.8701601028442383,
|
| 914 |
+
"rewards/rejected": -1.6024430990219116,
|
| 915 |
"step": 580
|
| 916 |
},
|
| 917 |
{
|
| 918 |
"epoch": 0.58,
|
| 919 |
"learning_rate": 2.234121467743082e-06,
|
| 920 |
+
"logits/chosen": -2.630315065383911,
|
| 921 |
+
"logits/rejected": -2.5930495262145996,
|
| 922 |
+
"logps/chosen": -234.6758270263672,
|
| 923 |
+
"logps/rejected": -278.23175048828125,
|
| 924 |
+
"loss": 0.5646,
|
| 925 |
+
"rewards/accuracies": 0.75,
|
| 926 |
+
"rewards/chosen": -0.7757617235183716,
|
| 927 |
+
"rewards/margins": 0.6909924149513245,
|
| 928 |
+
"rewards/rejected": -1.4667540788650513,
|
| 929 |
"step": 590
|
| 930 |
},
|
| 931 |
{
|
| 932 |
"epoch": 0.59,
|
| 933 |
"learning_rate": 2.148852897944905e-06,
|
| 934 |
+
"logits/chosen": -2.6809496879577637,
|
| 935 |
+
"logits/rejected": -2.5784506797790527,
|
| 936 |
+
"logps/chosen": -226.6197052001953,
|
| 937 |
+
"logps/rejected": -279.7991027832031,
|
| 938 |
+
"loss": 0.4594,
|
| 939 |
+
"rewards/accuracies": 0.737500011920929,
|
| 940 |
+
"rewards/chosen": -0.6655336022377014,
|
| 941 |
+
"rewards/margins": 0.8101358413696289,
|
| 942 |
+
"rewards/rejected": -1.475669503211975,
|
| 943 |
"step": 600
|
| 944 |
},
|
| 945 |
{
|
| 946 |
"epoch": 0.59,
|
| 947 |
+
"eval_logits/chosen": -2.519953966140747,
|
| 948 |
+
"eval_logits/rejected": -2.4671545028686523,
|
| 949 |
+
"eval_logps/chosen": -239.843017578125,
|
| 950 |
+
"eval_logps/rejected": -280.2419738769531,
|
| 951 |
+
"eval_loss": 0.5418068170547485,
|
| 952 |
+
"eval_rewards/accuracies": 0.7311320900917053,
|
| 953 |
+
"eval_rewards/chosen": -0.7977315783500671,
|
| 954 |
+
"eval_rewards/margins": 0.6694828271865845,
|
| 955 |
+
"eval_rewards/rejected": -1.4672144651412964,
|
| 956 |
+
"eval_runtime": 286.7576,
|
| 957 |
+
"eval_samples_per_second": 1.458,
|
| 958 |
+
"eval_steps_per_second": 0.185,
|
| 959 |
"step": 600
|
| 960 |
},
|
| 961 |
{
|
| 962 |
"epoch": 0.6,
|
| 963 |
"learning_rate": 2.063999142413574e-06,
|
| 964 |
+
"logits/chosen": -2.5949954986572266,
|
| 965 |
+
"logits/rejected": -2.512058734893799,
|
| 966 |
+
"logps/chosen": -240.7419891357422,
|
| 967 |
+
"logps/rejected": -261.2515869140625,
|
| 968 |
+
"loss": 0.5227,
|
| 969 |
"rewards/accuracies": 0.762499988079071,
|
| 970 |
+
"rewards/chosen": -0.7232692241668701,
|
| 971 |
+
"rewards/margins": 0.6376546621322632,
|
| 972 |
+
"rewards/rejected": -1.3609240055084229,
|
| 973 |
"step": 610
|
| 974 |
},
|
| 975 |
{
|
| 976 |
"epoch": 0.61,
|
| 977 |
"learning_rate": 1.9796604399042547e-06,
|
| 978 |
+
"logits/chosen": -2.6472525596618652,
|
| 979 |
+
"logits/rejected": -2.5546352863311768,
|
| 980 |
+
"logps/chosen": -257.9134521484375,
|
| 981 |
+
"logps/rejected": -305.9988098144531,
|
| 982 |
+
"loss": 0.4877,
|
| 983 |
+
"rewards/accuracies": 0.737500011920929,
|
| 984 |
+
"rewards/chosen": -0.944113552570343,
|
| 985 |
+
"rewards/margins": 0.8468207120895386,
|
| 986 |
+
"rewards/rejected": -1.7909342050552368,
|
| 987 |
"step": 620
|
| 988 |
},
|
| 989 |
{
|
| 990 |
"epoch": 0.62,
|
| 991 |
"learning_rate": 1.8959364207338216e-06,
|
| 992 |
+
"logits/chosen": -2.6289591789245605,
|
| 993 |
+
"logits/rejected": -2.536094903945923,
|
| 994 |
+
"logps/chosen": -235.07626342773438,
|
| 995 |
+
"logps/rejected": -293.07818603515625,
|
| 996 |
+
"loss": 0.499,
|
| 997 |
+
"rewards/accuracies": 0.75,
|
| 998 |
+
"rewards/chosen": -0.8861194849014282,
|
| 999 |
+
"rewards/margins": 0.8537158966064453,
|
| 1000 |
+
"rewards/rejected": -1.7398353815078735,
|
| 1001 |
"step": 630
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
"epoch": 0.63,
|
| 1005 |
"learning_rate": 1.8129259890863825e-06,
|
| 1006 |
+
"logits/chosen": -2.599257469177246,
|
| 1007 |
+
"logits/rejected": -2.542430877685547,
|
| 1008 |
+
"logps/chosen": -262.864013671875,
|
| 1009 |
+
"logps/rejected": -305.956787109375,
|
| 1010 |
+
"loss": 0.5611,
|
| 1011 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1012 |
+
"rewards/chosen": -0.9360750317573547,
|
| 1013 |
+
"rewards/margins": 0.6403096318244934,
|
| 1014 |
+
"rewards/rejected": -1.5763845443725586,
|
| 1015 |
"step": 640
|
| 1016 |
},
|
| 1017 |
{
|
| 1018 |
"epoch": 0.64,
|
| 1019 |
"learning_rate": 1.7307272061765738e-06,
|
| 1020 |
+
"logits/chosen": -2.641481876373291,
|
| 1021 |
+
"logits/rejected": -2.58241605758667,
|
| 1022 |
+
"logps/chosen": -251.8070068359375,
|
| 1023 |
+
"logps/rejected": -303.80462646484375,
|
| 1024 |
+
"loss": 0.5001,
|
| 1025 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1026 |
+
"rewards/chosen": -0.8562856912612915,
|
| 1027 |
+
"rewards/margins": 0.783263087272644,
|
| 1028 |
+
"rewards/rejected": -1.639548659324646,
|
| 1029 |
"step": 650
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
"epoch": 0.65,
|
| 1033 |
"learning_rate": 1.649437174408685e-06,
|
| 1034 |
+
"logits/chosen": -2.597597599029541,
|
| 1035 |
+
"logits/rejected": -2.541797637939453,
|
| 1036 |
+
"logps/chosen": -235.8343505859375,
|
| 1037 |
+
"logps/rejected": -286.0321350097656,
|
| 1038 |
+
"loss": 0.5432,
|
| 1039 |
"rewards/accuracies": 0.75,
|
| 1040 |
+
"rewards/chosen": -0.6598041653633118,
|
| 1041 |
+
"rewards/margins": 0.7441025972366333,
|
| 1042 |
+
"rewards/rejected": -1.4039068222045898,
|
| 1043 |
"step": 660
|
| 1044 |
},
|
| 1045 |
{
|
| 1046 |
"epoch": 0.66,
|
| 1047 |
"learning_rate": 1.569151922668422e-06,
|
| 1048 |
+
"logits/chosen": -2.638943910598755,
|
| 1049 |
+
"logits/rejected": -2.5677754878997803,
|
| 1050 |
+
"logps/chosen": -246.6320037841797,
|
| 1051 |
+
"logps/rejected": -297.98944091796875,
|
| 1052 |
+
"loss": 0.5573,
|
| 1053 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1054 |
+
"rewards/chosen": -0.9078775644302368,
|
| 1055 |
+
"rewards/margins": 0.736061692237854,
|
| 1056 |
+
"rewards/rejected": -1.6439392566680908,
|
| 1057 |
"step": 670
|
| 1058 |
},
|
| 1059 |
{
|
| 1060 |
"epoch": 0.67,
|
| 1061 |
"learning_rate": 1.4899662928828428e-06,
|
| 1062 |
+
"logits/chosen": -2.6092512607574463,
|
| 1063 |
+
"logits/rejected": -2.5140151977539062,
|
| 1064 |
+
"logps/chosen": -229.247802734375,
|
| 1065 |
+
"logps/rejected": -278.5384521484375,
|
| 1066 |
+
"loss": 0.5307,
|
| 1067 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1068 |
+
"rewards/chosen": -0.7807116508483887,
|
| 1069 |
+
"rewards/margins": 0.7786010503768921,
|
| 1070 |
+
"rewards/rejected": -1.5593125820159912,
|
| 1071 |
"step": 680
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
"epoch": 0.68,
|
| 1075 |
"learning_rate": 1.4119738279824507e-06,
|
| 1076 |
+
"logits/chosen": -2.5588324069976807,
|
| 1077 |
+
"logits/rejected": -2.4750542640686035,
|
| 1078 |
+
"logps/chosen": -229.9861602783203,
|
| 1079 |
+
"logps/rejected": -279.2628479003906,
|
| 1080 |
+
"loss": 0.5406,
|
| 1081 |
"rewards/accuracies": 0.7250000238418579,
|
| 1082 |
+
"rewards/chosen": -0.7474340200424194,
|
| 1083 |
+
"rewards/margins": 0.7327946424484253,
|
| 1084 |
+
"rewards/rejected": -1.4802285432815552,
|
| 1085 |
"step": 690
|
| 1086 |
},
|
| 1087 |
{
|
| 1088 |
"epoch": 0.69,
|
| 1089 |
"learning_rate": 1.3352666613978152e-06,
|
| 1090 |
+
"logits/chosen": -2.522113084793091,
|
| 1091 |
+
"logits/rejected": -2.478847026824951,
|
| 1092 |
+
"logps/chosen": -247.2416229248047,
|
| 1093 |
+
"logps/rejected": -284.3504943847656,
|
| 1094 |
+
"loss": 0.5444,
|
| 1095 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1096 |
+
"rewards/chosen": -0.8833193778991699,
|
| 1097 |
+
"rewards/margins": 0.644477367401123,
|
| 1098 |
+
"rewards/rejected": -1.5277965068817139,
|
| 1099 |
+
"step": 700
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.69,
|
| 1103 |
+
"eval_logits/chosen": -2.5126736164093018,
|
| 1104 |
+
"eval_logits/rejected": -2.459439277648926,
|
| 1105 |
+
"eval_logps/chosen": -236.95314025878906,
|
| 1106 |
+
"eval_logps/rejected": -278.8013610839844,
|
| 1107 |
+
"eval_loss": 0.5358262062072754,
|
| 1108 |
+
"eval_rewards/accuracies": 0.7334905862808228,
|
| 1109 |
+
"eval_rewards/chosen": -0.7688332200050354,
|
| 1110 |
+
"eval_rewards/margins": 0.683975100517273,
|
| 1111 |
+
"eval_rewards/rejected": -1.4528082609176636,
|
| 1112 |
+
"eval_runtime": 286.9906,
|
| 1113 |
+
"eval_samples_per_second": 1.456,
|
| 1114 |
+
"eval_steps_per_second": 0.185,
|
| 1115 |
"step": 700
|
| 1116 |
},
|
| 1117 |
{
|
| 1118 |
"epoch": 0.7,
|
| 1119 |
"learning_rate": 1.2599354082212523e-06,
|
| 1120 |
+
"logits/chosen": -2.5635337829589844,
|
| 1121 |
+
"logits/rejected": -2.475813388824463,
|
| 1122 |
+
"logps/chosen": -212.53189086914062,
|
| 1123 |
+
"logps/rejected": -260.5958557128906,
|
| 1124 |
+
"loss": 0.5358,
|
| 1125 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1126 |
+
"rewards/chosen": -0.5574060678482056,
|
| 1127 |
+
"rewards/margins": 0.7405737042427063,
|
| 1128 |
+
"rewards/rejected": -1.297979712486267,
|
| 1129 |
"step": 710
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.71,
|
| 1133 |
"learning_rate": 1.186069058162127e-06,
|
| 1134 |
+
"logits/chosen": -2.6438024044036865,
|
| 1135 |
+
"logits/rejected": -2.565239906311035,
|
| 1136 |
+
"logps/chosen": -234.189697265625,
|
| 1137 |
+
"logps/rejected": -277.8279724121094,
|
| 1138 |
+
"loss": 0.5251,
|
| 1139 |
"rewards/accuracies": 0.737500011920929,
|
| 1140 |
+
"rewards/chosen": -0.8120694160461426,
|
| 1141 |
+
"rewards/margins": 0.6847533583641052,
|
| 1142 |
+
"rewards/rejected": -1.496822714805603,
|
| 1143 |
"step": 720
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.72,
|
| 1147 |
"learning_rate": 1.113754870422254e-06,
|
| 1148 |
+
"logits/chosen": -2.6735897064208984,
|
| 1149 |
+
"logits/rejected": -2.599342107772827,
|
| 1150 |
+
"logps/chosen": -255.75906372070312,
|
| 1151 |
+
"logps/rejected": -297.49774169921875,
|
| 1152 |
+
"loss": 0.5672,
|
| 1153 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1154 |
+
"rewards/chosen": -0.9736843109130859,
|
| 1155 |
+
"rewards/margins": 0.5445233583450317,
|
| 1156 |
+
"rewards/rejected": -1.5182076692581177,
|
| 1157 |
"step": 730
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
"epoch": 0.73,
|
| 1161 |
"learning_rate": 1.0430782706155545e-06,
|
| 1162 |
+
"logits/chosen": -2.6966583728790283,
|
| 1163 |
+
"logits/rejected": -2.6031241416931152,
|
| 1164 |
+
"logps/chosen": -255.4834747314453,
|
| 1165 |
+
"logps/rejected": -305.7486267089844,
|
| 1166 |
+
"loss": 0.4757,
|
| 1167 |
+
"rewards/accuracies": 0.8125,
|
| 1168 |
+
"rewards/chosen": -0.8519601821899414,
|
| 1169 |
+
"rewards/margins": 0.8620556592941284,
|
| 1170 |
+
"rewards/rejected": -1.7140159606933594,
|
| 1171 |
"step": 740
|
| 1172 |
},
|
| 1173 |
{
|
| 1174 |
"epoch": 0.74,
|
| 1175 |
"learning_rate": 9.741227498537615e-07,
|
| 1176 |
+
"logits/chosen": -2.7119975090026855,
|
| 1177 |
+
"logits/rejected": -2.626258373260498,
|
| 1178 |
+
"logps/chosen": -239.32894897460938,
|
| 1179 |
+
"logps/rejected": -292.5682678222656,
|
| 1180 |
+
"loss": 0.4648,
|
| 1181 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1182 |
+
"rewards/chosen": -0.9102465510368347,
|
| 1183 |
+
"rewards/margins": 0.7873125076293945,
|
| 1184 |
+
"rewards/rejected": -1.697559118270874,
|
| 1185 |
"step": 750
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.75,
|
| 1189 |
"learning_rate": 9.069697661173668e-07,
|
| 1190 |
+
"logits/chosen": -2.6146960258483887,
|
| 1191 |
+
"logits/rejected": -2.5526528358459473,
|
| 1192 |
+
"logps/chosen": -253.11514282226562,
|
| 1193 |
+
"logps/rejected": -305.19647216796875,
|
| 1194 |
+
"loss": 0.5725,
|
| 1195 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1196 |
+
"rewards/chosen": -0.9893601536750793,
|
| 1197 |
+
"rewards/margins": 0.6918299794197083,
|
| 1198 |
+
"rewards/rejected": -1.6811901330947876,
|
| 1199 |
"step": 760
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
"epoch": 0.76,
|
| 1203 |
"learning_rate": 8.416986480283434e-07,
|
| 1204 |
+
"logits/chosen": -2.593777894973755,
|
| 1205 |
+
"logits/rejected": -2.497063636779785,
|
| 1206 |
+
"logps/chosen": -259.0460510253906,
|
| 1207 |
+
"logps/rejected": -318.4884033203125,
|
| 1208 |
+
"loss": 0.5069,
|
| 1209 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1210 |
+
"rewards/chosen": -1.0463413000106812,
|
| 1211 |
+
"rewards/margins": 0.8268297910690308,
|
| 1212 |
+
"rewards/rejected": -1.8731712102890015,
|
| 1213 |
"step": 770
|
| 1214 |
},
|
| 1215 |
{
|
| 1216 |
"epoch": 0.77,
|
| 1217 |
"learning_rate": 7.783865011382876e-07,
|
| 1218 |
+
"logits/chosen": -2.6105198860168457,
|
| 1219 |
+
"logits/rejected": -2.5155270099639893,
|
| 1220 |
+
"logps/chosen": -267.23785400390625,
|
| 1221 |
+
"logps/rejected": -316.72027587890625,
|
| 1222 |
+
"loss": 0.4906,
|
| 1223 |
+
"rewards/accuracies": 0.75,
|
| 1224 |
+
"rewards/chosen": -0.9409424662590027,
|
| 1225 |
+
"rewards/margins": 0.8792687654495239,
|
| 1226 |
+
"rewards/rejected": -1.8202111721038818,
|
| 1227 |
"step": 780
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.78,
|
| 1231 |
"learning_rate": 7.171081168427205e-07,
|
| 1232 |
+
"logits/chosen": -2.642396926879883,
|
| 1233 |
+
"logits/rejected": -2.5414986610412598,
|
| 1234 |
+
"logps/chosen": -278.53076171875,
|
| 1235 |
+
"logps/rejected": -327.7699279785156,
|
| 1236 |
+
"loss": 0.4906,
|
| 1237 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1238 |
+
"rewards/chosen": -0.944370448589325,
|
| 1239 |
+
"rewards/margins": 0.8692710995674133,
|
| 1240 |
+
"rewards/rejected": -1.8136417865753174,
|
| 1241 |
"step": 790
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.79,
|
| 1245 |
"learning_rate": 6.579358840291064e-07,
|
| 1246 |
+
"logits/chosen": -2.619293451309204,
|
| 1247 |
+
"logits/rejected": -2.5392627716064453,
|
| 1248 |
+
"logps/chosen": -275.8484802246094,
|
| 1249 |
+
"logps/rejected": -315.7076721191406,
|
| 1250 |
+
"loss": 0.5755,
|
| 1251 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1252 |
+
"rewards/chosen": -1.142333745956421,
|
| 1253 |
+
"rewards/margins": 0.5214550495147705,
|
| 1254 |
+
"rewards/rejected": -1.6637885570526123,
|
| 1255 |
"step": 800
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 0.79,
|
| 1259 |
+
"eval_logits/chosen": -2.511751890182495,
|
| 1260 |
+
"eval_logits/rejected": -2.458522081375122,
|
| 1261 |
+
"eval_logps/chosen": -266.7905578613281,
|
| 1262 |
+
"eval_logps/rejected": -309.8293151855469,
|
| 1263 |
+
"eval_loss": 0.5405105352401733,
|
| 1264 |
+
"eval_rewards/accuracies": 0.7311320900917053,
|
| 1265 |
+
"eval_rewards/chosen": -1.0672070980072021,
|
| 1266 |
+
"eval_rewards/margins": 0.695880651473999,
|
| 1267 |
+
"eval_rewards/rejected": -1.7630878686904907,
|
| 1268 |
+
"eval_runtime": 287.1767,
|
| 1269 |
+
"eval_samples_per_second": 1.456,
|
| 1270 |
+
"eval_steps_per_second": 0.185,
|
| 1271 |
"step": 800
|
| 1272 |
},
|
| 1273 |
{
|
| 1274 |
"epoch": 0.8,
|
| 1275 |
"learning_rate": 6.00939703563006e-07,
|
| 1276 |
+
"logits/chosen": -2.647855281829834,
|
| 1277 |
+
"logits/rejected": -2.6014881134033203,
|
| 1278 |
+
"logps/chosen": -251.8603057861328,
|
| 1279 |
+
"logps/rejected": -300.19500732421875,
|
| 1280 |
+
"loss": 0.5697,
|
| 1281 |
+
"rewards/accuracies": 0.6875,
|
| 1282 |
+
"rewards/chosen": -1.0764825344085693,
|
| 1283 |
+
"rewards/margins": 0.6206298470497131,
|
| 1284 |
+
"rewards/rejected": -1.6971123218536377,
|
| 1285 |
"step": 810
|
| 1286 |
},
|
| 1287 |
{
|
| 1288 |
"epoch": 0.81,
|
| 1289 |
"learning_rate": 5.461869057133412e-07,
|
| 1290 |
+
"logits/chosen": -2.6618542671203613,
|
| 1291 |
+
"logits/rejected": -2.6055150032043457,
|
| 1292 |
+
"logps/chosen": -266.5983581542969,
|
| 1293 |
+
"logps/rejected": -328.2469177246094,
|
| 1294 |
+
"loss": 0.498,
|
| 1295 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1296 |
+
"rewards/chosen": -1.0761178731918335,
|
| 1297 |
+
"rewards/margins": 0.7884665727615356,
|
| 1298 |
+
"rewards/rejected": -1.8645845651626587,
|
| 1299 |
"step": 820
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
"epoch": 0.82,
|
| 1303 |
"learning_rate": 4.937421706143497e-07,
|
| 1304 |
+
"logits/chosen": -2.5800223350524902,
|
| 1305 |
+
"logits/rejected": -2.5038163661956787,
|
| 1306 |
+
"logps/chosen": -276.6379699707031,
|
| 1307 |
+
"logps/rejected": -326.5852355957031,
|
| 1308 |
+
"loss": 0.5034,
|
| 1309 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1310 |
+
"rewards/chosen": -1.1193444728851318,
|
| 1311 |
+
"rewards/margins": 0.8236180543899536,
|
| 1312 |
+
"rewards/rejected": -1.9429620504379272,
|
| 1313 |
"step": 830
|
| 1314 |
},
|
| 1315 |
{
|
| 1316 |
"epoch": 0.83,
|
| 1317 |
"learning_rate": 4.43667451858166e-07,
|
| 1318 |
+
"logits/chosen": -2.5895838737487793,
|
| 1319 |
+
"logits/rejected": -2.4980013370513916,
|
| 1320 |
+
"logps/chosen": -245.9558563232422,
|
| 1321 |
+
"logps/rejected": -303.1318054199219,
|
| 1322 |
+
"loss": 0.4828,
|
| 1323 |
+
"rewards/accuracies": 0.8125,
|
| 1324 |
+
"rewards/chosen": -0.9489553570747375,
|
| 1325 |
+
"rewards/margins": 0.8449093103408813,
|
| 1326 |
+
"rewards/rejected": -1.7938644886016846,
|
| 1327 |
"step": 840
|
| 1328 |
},
|
| 1329 |
{
|
| 1330 |
"epoch": 0.84,
|
| 1331 |
"learning_rate": 3.9602190330830484e-07,
|
| 1332 |
+
"logits/chosen": -2.5641655921936035,
|
| 1333 |
+
"logits/rejected": -2.5129027366638184,
|
| 1334 |
+
"logps/chosen": -265.5952453613281,
|
| 1335 |
+
"logps/rejected": -319.8917236328125,
|
| 1336 |
+
"loss": 0.5089,
|
| 1337 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1338 |
+
"rewards/chosen": -1.1247035264968872,
|
| 1339 |
+
"rewards/margins": 0.7079113721847534,
|
| 1340 |
+
"rewards/rejected": -1.8326148986816406,
|
| 1341 |
"step": 850
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
"epoch": 0.85,
|
| 1345 |
"learning_rate": 3.5086180922049295e-07,
|
| 1346 |
+
"logits/chosen": -2.653895854949951,
|
| 1347 |
+
"logits/rejected": -2.570190906524658,
|
| 1348 |
+
"logps/chosen": -277.4553527832031,
|
| 1349 |
+
"logps/rejected": -311.83050537109375,
|
| 1350 |
+
"loss": 0.4957,
|
| 1351 |
+
"rewards/accuracies": 0.75,
|
| 1352 |
+
"rewards/chosen": -0.9871084094047546,
|
| 1353 |
+
"rewards/margins": 0.7766774892807007,
|
| 1354 |
+
"rewards/rejected": -1.7637859582901,
|
| 1355 |
"step": 860
|
| 1356 |
},
|
| 1357 |
{
|
| 1358 |
"epoch": 0.86,
|
| 1359 |
"learning_rate": 3.0824051775340895e-07,
|
| 1360 |
+
"logits/chosen": -2.6221535205841064,
|
| 1361 |
+
"logits/rejected": -2.560570001602173,
|
| 1362 |
+
"logps/chosen": -223.2664337158203,
|
| 1363 |
+
"logps/rejected": -278.30316162109375,
|
| 1364 |
+
"loss": 0.5322,
|
| 1365 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1366 |
+
"rewards/chosen": -0.8692310452461243,
|
| 1367 |
+
"rewards/margins": 0.6867448091506958,
|
| 1368 |
+
"rewards/rejected": -1.5559759140014648,
|
| 1369 |
"step": 870
|
| 1370 |
},
|
| 1371 |
{
|
| 1372 |
"epoch": 0.87,
|
| 1373 |
"learning_rate": 2.6820837794786336e-07,
|
| 1374 |
+
"logits/chosen": -2.5683445930480957,
|
| 1375 |
+
"logits/rejected": -2.54264497756958,
|
| 1376 |
+
"logps/chosen": -264.15106201171875,
|
| 1377 |
+
"logps/rejected": -287.0775451660156,
|
| 1378 |
+
"loss": 0.6567,
|
| 1379 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1380 |
+
"rewards/chosen": -1.0795668363571167,
|
| 1381 |
+
"rewards/margins": 0.3045983612537384,
|
| 1382 |
+
"rewards/rejected": -1.3841651678085327,
|
| 1383 |
"step": 880
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
"epoch": 0.88,
|
| 1387 |
"learning_rate": 2.3081268024887694e-07,
|
| 1388 |
+
"logits/chosen": -2.5630886554718018,
|
| 1389 |
+
"logits/rejected": -2.45955228805542,
|
| 1390 |
+
"logps/chosen": -250.3737335205078,
|
| 1391 |
+
"logps/rejected": -303.13177490234375,
|
| 1392 |
+
"loss": 0.4701,
|
| 1393 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1394 |
+
"rewards/chosen": -0.9046028852462769,
|
| 1395 |
+
"rewards/margins": 0.8787680864334106,
|
| 1396 |
+
"rewards/rejected": -1.7833709716796875,
|
| 1397 |
"step": 890
|
| 1398 |
},
|
| 1399 |
{
|
| 1400 |
"epoch": 0.89,
|
| 1401 |
"learning_rate": 1.9609760064091044e-07,
|
| 1402 |
+
"logits/chosen": -2.6273365020751953,
|
| 1403 |
+
"logits/rejected": -2.597182512283325,
|
| 1404 |
+
"logps/chosen": -255.65463256835938,
|
| 1405 |
+
"logps/rejected": -284.71759033203125,
|
| 1406 |
+
"loss": 0.5495,
|
| 1407 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1408 |
+
"rewards/chosen": -0.941402792930603,
|
| 1409 |
+
"rewards/margins": 0.628917396068573,
|
| 1410 |
+
"rewards/rejected": -1.5703203678131104,
|
| 1411 |
+
"step": 900
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 0.89,
|
| 1415 |
+
"eval_logits/chosen": -2.509645938873291,
|
| 1416 |
+
"eval_logits/rejected": -2.4557979106903076,
|
| 1417 |
+
"eval_logps/chosen": -253.28036499023438,
|
| 1418 |
+
"eval_logps/rejected": -298.0242004394531,
|
| 1419 |
+
"eval_loss": 0.5370938777923584,
|
| 1420 |
+
"eval_rewards/accuracies": 0.7287735939025879,
|
| 1421 |
+
"eval_rewards/chosen": -0.932105302810669,
|
| 1422 |
+
"eval_rewards/margins": 0.7129315733909607,
|
| 1423 |
+
"eval_rewards/rejected": -1.645037055015564,
|
| 1424 |
+
"eval_runtime": 286.9177,
|
| 1425 |
+
"eval_samples_per_second": 1.457,
|
| 1426 |
+
"eval_steps_per_second": 0.185,
|
| 1427 |
"step": 900
|
| 1428 |
},
|
| 1429 |
{
|
| 1430 |
"epoch": 0.9,
|
| 1431 |
"learning_rate": 1.6410414846224992e-07,
|
| 1432 |
+
"logits/chosen": -2.5519988536834717,
|
| 1433 |
+
"logits/rejected": -2.4820520877838135,
|
| 1434 |
+
"logps/chosen": -240.66830444335938,
|
| 1435 |
+
"logps/rejected": -293.9994812011719,
|
| 1436 |
+
"loss": 0.4861,
|
| 1437 |
"rewards/accuracies": 0.7875000238418579,
|
| 1438 |
+
"rewards/chosen": -0.8613694906234741,
|
| 1439 |
+
"rewards/margins": 0.8020180463790894,
|
| 1440 |
+
"rewards/rejected": -1.6633875370025635,
|
| 1441 |
"step": 910
|
| 1442 |
},
|
| 1443 |
{
|
| 1444 |
"epoch": 0.91,
|
| 1445 |
"learning_rate": 1.348701179601819e-07,
|
| 1446 |
+
"logits/chosen": -2.6739838123321533,
|
| 1447 |
+
"logits/rejected": -2.5836825370788574,
|
| 1448 |
+
"logps/chosen": -261.11077880859375,
|
| 1449 |
+
"logps/rejected": -305.97698974609375,
|
| 1450 |
+
"loss": 0.5087,
|
| 1451 |
"rewards/accuracies": 0.7875000238418579,
|
| 1452 |
+
"rewards/chosen": -0.8714740872383118,
|
| 1453 |
+
"rewards/margins": 0.7631353139877319,
|
| 1454 |
+
"rewards/rejected": -1.6346092224121094,
|
| 1455 |
"step": 920
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
"epoch": 0.91,
|
| 1459 |
"learning_rate": 1.0843004364420151e-07,
|
| 1460 |
+
"logits/chosen": -2.5310449600219727,
|
| 1461 |
+
"logits/rejected": -2.465944766998291,
|
| 1462 |
+
"logps/chosen": -262.35516357421875,
|
| 1463 |
+
"logps/rejected": -307.0565185546875,
|
| 1464 |
+
"loss": 0.5616,
|
| 1465 |
+
"rewards/accuracies": 0.75,
|
| 1466 |
+
"rewards/chosen": -0.8902053833007812,
|
| 1467 |
+
"rewards/margins": 0.6049326658248901,
|
| 1468 |
+
"rewards/rejected": -1.4951379299163818,
|
| 1469 |
"step": 930
|
| 1470 |
},
|
| 1471 |
{
|
| 1472 |
"epoch": 0.92,
|
| 1473 |
"learning_rate": 8.481515948997931e-08,
|
| 1474 |
+
"logits/chosen": -2.6589980125427246,
|
| 1475 |
+
"logits/rejected": -2.611510753631592,
|
| 1476 |
+
"logps/chosen": -270.62567138671875,
|
| 1477 |
+
"logps/rejected": -293.7890319824219,
|
| 1478 |
+
"loss": 0.6245,
|
| 1479 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1480 |
+
"rewards/chosen": -0.9786608815193176,
|
| 1481 |
+
"rewards/margins": 0.5622497797012329,
|
| 1482 |
+
"rewards/rejected": -1.5409104824066162,
|
| 1483 |
"step": 940
|
| 1484 |
},
|
| 1485 |
{
|
| 1486 |
"epoch": 0.93,
|
| 1487 |
"learning_rate": 6.4053362042297e-08,
|
| 1488 |
+
"logits/chosen": -2.5867984294891357,
|
| 1489 |
+
"logits/rejected": -2.4980480670928955,
|
| 1490 |
+
"logps/chosen": -251.9007568359375,
|
| 1491 |
+
"logps/rejected": -304.0977478027344,
|
| 1492 |
+
"loss": 0.4937,
|
| 1493 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1494 |
+
"rewards/chosen": -0.8755489587783813,
|
| 1495 |
+
"rewards/margins": 0.829258918762207,
|
| 1496 |
+
"rewards/rejected": -1.7048076391220093,
|
| 1497 |
"step": 950
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
"epoch": 0.94,
|
| 1501 |
"learning_rate": 4.616917746052163e-08,
|
| 1502 |
+
"logits/chosen": -2.6321659088134766,
|
| 1503 |
+
"logits/rejected": -2.556123971939087,
|
| 1504 |
+
"logps/chosen": -261.781982421875,
|
| 1505 |
+
"logps/rejected": -302.46527099609375,
|
| 1506 |
+
"loss": 0.5602,
|
| 1507 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1508 |
+
"rewards/chosen": -1.0545774698257446,
|
| 1509 |
+
"rewards/margins": 0.6820313334465027,
|
| 1510 |
+
"rewards/rejected": -1.7366087436676025,
|
| 1511 |
"step": 960
|
| 1512 |
},
|
| 1513 |
{
|
| 1514 |
"epoch": 0.95,
|
| 1515 |
"learning_rate": 3.118373254556412e-08,
|
| 1516 |
+
"logits/chosen": -2.625166416168213,
|
| 1517 |
+
"logits/rejected": -2.547811269760132,
|
| 1518 |
+
"logps/chosen": -235.4457550048828,
|
| 1519 |
+
"logps/rejected": -285.9036560058594,
|
| 1520 |
+
"loss": 0.4703,
|
| 1521 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1522 |
+
"rewards/chosen": -0.8371899724006653,
|
| 1523 |
+
"rewards/margins": 0.8404737710952759,
|
| 1524 |
+
"rewards/rejected": -1.677663803100586,
|
| 1525 |
"step": 970
|
| 1526 |
},
|
| 1527 |
{
|
| 1528 |
"epoch": 0.96,
|
| 1529 |
"learning_rate": 1.9114729782535037e-08,
|
| 1530 |
+
"logits/chosen": -2.6821470260620117,
|
| 1531 |
+
"logits/rejected": -2.605581045150757,
|
| 1532 |
+
"logps/chosen": -240.4625701904297,
|
| 1533 |
+
"logps/rejected": -286.89263916015625,
|
| 1534 |
+
"loss": 0.5037,
|
| 1535 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1536 |
+
"rewards/chosen": -0.86335289478302,
|
| 1537 |
+
"rewards/margins": 0.7098163366317749,
|
| 1538 |
+
"rewards/rejected": -1.5731691122055054,
|
| 1539 |
"step": 980
|
| 1540 |
},
|
| 1541 |
{
|
| 1542 |
"epoch": 0.97,
|
| 1543 |
"learning_rate": 9.97642642858815e-09,
|
| 1544 |
+
"logits/chosen": -2.5950284004211426,
|
| 1545 |
+
"logits/rejected": -2.50787615776062,
|
| 1546 |
+
"logps/chosen": -258.5409851074219,
|
| 1547 |
+
"logps/rejected": -290.9730224609375,
|
| 1548 |
+
"loss": 0.5398,
|
| 1549 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1550 |
+
"rewards/chosen": -0.9514690637588501,
|
| 1551 |
+
"rewards/margins": 0.6696909666061401,
|
| 1552 |
+
"rewards/rejected": -1.6211597919464111,
|
| 1553 |
"step": 990
|
| 1554 |
},
|
| 1555 |
{
|
| 1556 |
"epoch": 0.98,
|
| 1557 |
"learning_rate": 3.779617670651436e-09,
|
| 1558 |
+
"logits/chosen": -2.56034517288208,
|
| 1559 |
+
"logits/rejected": -2.5073485374450684,
|
| 1560 |
+
"logps/chosen": -267.876953125,
|
| 1561 |
+
"logps/rejected": -305.6854248046875,
|
| 1562 |
+
"loss": 0.5948,
|
| 1563 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1564 |
+
"rewards/chosen": -1.0524874925613403,
|
| 1565 |
+
"rewards/margins": 0.6285669207572937,
|
| 1566 |
+
"rewards/rejected": -1.6810543537139893,
|
| 1567 |
"step": 1000
|
| 1568 |
},
|
| 1569 |
{
|
| 1570 |
"epoch": 0.98,
|
| 1571 |
+
"eval_logits/chosen": -2.5093116760253906,
|
| 1572 |
+
"eval_logits/rejected": -2.4554219245910645,
|
| 1573 |
+
"eval_logps/chosen": -253.41494750976562,
|
| 1574 |
+
"eval_logps/rejected": -298.073486328125,
|
| 1575 |
+
"eval_loss": 0.5370525121688843,
|
| 1576 |
+
"eval_rewards/accuracies": 0.7264150977134705,
|
| 1577 |
+
"eval_rewards/chosen": -0.9334509372711182,
|
| 1578 |
+
"eval_rewards/margins": 0.7120789289474487,
|
| 1579 |
+
"eval_rewards/rejected": -1.6455297470092773,
|
| 1580 |
+
"eval_runtime": 287.936,
|
| 1581 |
+
"eval_samples_per_second": 1.452,
|
| 1582 |
+
"eval_steps_per_second": 0.184,
|
| 1583 |
"step": 1000
|
| 1584 |
},
|
| 1585 |
{
|
| 1586 |
"epoch": 0.99,
|
| 1587 |
"learning_rate": 5.316238729444201e-10,
|
| 1588 |
+
"logits/chosen": -2.580871820449829,
|
| 1589 |
+
"logits/rejected": -2.4882521629333496,
|
| 1590 |
+
"logps/chosen": -263.0270690917969,
|
| 1591 |
+
"logps/rejected": -309.5653076171875,
|
| 1592 |
+
"loss": 0.4902,
|
| 1593 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1594 |
+
"rewards/chosen": -1.025237798690796,
|
| 1595 |
+
"rewards/margins": 0.790520191192627,
|
| 1596 |
+
"rewards/rejected": -1.8157583475112915,
|
| 1597 |
"step": 1010
|
| 1598 |
},
|
| 1599 |
{
|
| 1600 |
"epoch": 1.0,
|
| 1601 |
"step": 1016,
|
| 1602 |
"total_flos": 0.0,
|
| 1603 |
+
"train_loss": 0.5672297686573089,
|
| 1604 |
+
"train_runtime": 12946.9794,
|
| 1605 |
+
"train_samples_per_second": 0.628,
|
| 1606 |
+
"train_steps_per_second": 0.078
|
| 1607 |
}
|
| 1608 |
],
|
| 1609 |
"logging_steps": 10,
|