Model save
Browse files- README.md +42 -13
- all_results.json +5 -5
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- train_results.json +5 -5
- trainer_state.json +1756 -362
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -15,15 +15,15 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.3](https://huggingface.co/mistralai/Mistral-7B-v0.3) on an unknown dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
-
- Loss:
|
| 19 |
-
- Rewards/real:
|
| 20 |
-
- Rewards/generated:
|
| 21 |
-
- Rewards/accuracies: 0.
|
| 22 |
-
- Rewards/margins: 0.
|
| 23 |
-
- Logps/generated: -
|
| 24 |
-
- Logps/real: -
|
| 25 |
-
- Logits/generated: -2.
|
| 26 |
-
- Logits/real: -2.
|
| 27 |
|
| 28 |
## Model description
|
| 29 |
|
|
@@ -54,13 +54,42 @@ The following hyperparameters were used during training:
|
|
| 54 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 55 |
- lr_scheduler_type: linear
|
| 56 |
- lr_scheduler_warmup_ratio: 0.1
|
| 57 |
-
- num_epochs:
|
| 58 |
|
| 59 |
### Training results
|
| 60 |
|
| 61 |
-
| Training Loss | Epoch
|
| 62 |
-
|
| 63 |
-
| 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
|
| 65 |
|
| 66 |
### Framework versions
|
|
|
|
| 15 |
|
| 16 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.3](https://huggingface.co/mistralai/Mistral-7B-v0.3) on an unknown dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 1.4922
|
| 19 |
+
- Rewards/real: -4.5129
|
| 20 |
+
- Rewards/generated: -4.6699
|
| 21 |
+
- Rewards/accuracies: 0.4423
|
| 22 |
+
- Rewards/margins: 0.1570
|
| 23 |
+
- Logps/generated: -155.2124
|
| 24 |
+
- Logps/real: -181.5346
|
| 25 |
+
- Logits/generated: -2.1203
|
| 26 |
+
- Logits/real: -2.3164
|
| 27 |
|
| 28 |
## Model description
|
| 29 |
|
|
|
|
| 54 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 55 |
- lr_scheduler_type: linear
|
| 56 |
- lr_scheduler_warmup_ratio: 0.1
|
| 57 |
+
- num_epochs: 3
|
| 58 |
|
| 59 |
### Training results
|
| 60 |
|
| 61 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/real | Rewards/generated | Rewards/accuracies | Rewards/margins | Logps/generated | Logps/real | Logits/generated | Logits/real |
|
| 62 |
+
|:-------------:|:------:|:----:|:---------------:|:------------:|:-----------------:|:------------------:|:---------------:|:---------------:|:----------:|:----------------:|:-----------:|
|
| 63 |
+
| 0.7877 | 0.0992 | 31 | 0.7717 | 0.6905 | 0.4761 | 0.7308 | 0.2144 | -103.7523 | -129.5006 | -2.5223 | -2.6981 |
|
| 64 |
+
| 0.6309 | 0.1984 | 62 | 0.7320 | 2.1415 | 1.7428 | 0.7115 | 0.3987 | -91.0858 | -114.9906 | -2.5564 | -2.7396 |
|
| 65 |
+
| 0.5309 | 0.2976 | 93 | 0.7175 | 1.5709 | 0.9016 | 0.6538 | 0.6692 | -99.4969 | -120.6967 | -2.4703 | -2.6171 |
|
| 66 |
+
| 0.4323 | 0.3968 | 124 | 0.7714 | 1.9586 | 1.4739 | 0.6923 | 0.4847 | -93.7744 | -116.8195 | -2.6808 | -2.8349 |
|
| 67 |
+
| 0.297 | 0.496 | 155 | 0.7161 | 2.2903 | 1.7549 | 0.8077 | 0.5355 | -90.9648 | -113.5018 | -2.6256 | -2.7696 |
|
| 68 |
+
| 0.2144 | 0.5952 | 186 | 0.8257 | 1.6038 | 1.0213 | 0.7115 | 0.5825 | -98.3000 | -120.3671 | -2.8900 | -3.0602 |
|
| 69 |
+
| 0.2497 | 0.6944 | 217 | 0.6849 | 2.4543 | 1.7831 | 0.8077 | 0.6712 | -90.6823 | -111.8619 | -2.6469 | -2.8201 |
|
| 70 |
+
| 0.1112 | 0.7936 | 248 | 0.6993 | 2.2831 | 1.4322 | 0.7885 | 0.8508 | -94.1910 | -113.5747 | -2.7020 | -2.8645 |
|
| 71 |
+
| 0.176 | 0.8928 | 279 | 0.6700 | 2.7841 | 2.2447 | 0.7692 | 0.5394 | -86.0663 | -108.5641 | -2.8051 | -2.9280 |
|
| 72 |
+
| 0.1135 | 0.992 | 310 | 0.6956 | 2.3849 | 1.8198 | 0.7885 | 0.5651 | -90.3149 | -112.5561 | -2.8024 | -2.9203 |
|
| 73 |
+
| 0.1221 | 1.0912 | 341 | 0.7314 | 2.2046 | 1.6143 | 0.7308 | 0.5903 | -92.3708 | -114.3593 | -2.5886 | -2.7365 |
|
| 74 |
+
| 0.0864 | 1.1904 | 372 | 0.7718 | 2.3206 | 1.9459 | 0.6346 | 0.3747 | -89.0543 | -113.1994 | -2.5355 | -2.7014 |
|
| 75 |
+
| 0.0871 | 1.2896 | 403 | 0.8231 | 1.9873 | 1.7063 | 0.5962 | 0.2810 | -91.4506 | -116.5322 | -2.5240 | -2.6833 |
|
| 76 |
+
| 0.1454 | 1.3888 | 434 | 0.7980 | 1.7358 | 1.2782 | 0.6731 | 0.4576 | -95.7309 | -119.0471 | -2.4325 | -2.6120 |
|
| 77 |
+
| 0.0747 | 1.488 | 465 | 0.8086 | 1.9033 | 1.4938 | 0.6538 | 0.4094 | -93.5750 | -117.3725 | -2.3557 | -2.5683 |
|
| 78 |
+
| 0.0882 | 1.5872 | 496 | 0.9281 | 0.8252 | 0.4834 | 0.5192 | 0.3418 | -103.6798 | -128.1537 | -2.2722 | -2.4783 |
|
| 79 |
+
| 0.0693 | 1.6864 | 527 | 0.8954 | 0.5032 | -0.0439 | 0.6154 | 0.5471 | -108.9523 | -131.3737 | -2.1399 | -2.3681 |
|
| 80 |
+
| 0.0982 | 1.7856 | 558 | 0.8777 | 1.0122 | 0.5411 | 0.6538 | 0.4711 | -103.1028 | -126.2834 | -2.3326 | -2.5183 |
|
| 81 |
+
| 0.0674 | 1.8848 | 589 | 0.9360 | -0.0587 | -0.5311 | 0.5962 | 0.4724 | -113.8238 | -136.9920 | -2.3026 | -2.4848 |
|
| 82 |
+
| 0.0424 | 1.984 | 620 | 0.9421 | -0.2586 | -0.6968 | 0.5769 | 0.4382 | -115.4816 | -138.9915 | -2.2955 | -2.4846 |
|
| 83 |
+
| 0.0235 | 2.0832 | 651 | 1.0939 | -1.6766 | -2.0193 | 0.5 | 0.3428 | -128.7065 | -153.1709 | -2.2115 | -2.3974 |
|
| 84 |
+
| 0.024 | 2.1824 | 682 | 1.1491 | -2.1565 | -2.5396 | 0.5 | 0.3831 | -133.9093 | -157.9701 | -2.2049 | -2.3936 |
|
| 85 |
+
| 0.0469 | 2.2816 | 713 | 1.1324 | -2.0618 | -2.4801 | 0.5 | 0.4183 | -133.3140 | -157.0232 | -2.2161 | -2.4094 |
|
| 86 |
+
| 0.0328 | 2.3808 | 744 | 1.1837 | -2.4534 | -2.7702 | 0.4808 | 0.3168 | -136.2151 | -160.9390 | -2.2080 | -2.3952 |
|
| 87 |
+
| 0.0367 | 2.48 | 775 | 1.1779 | -2.6139 | -2.9724 | 0.4808 | 0.3585 | -138.2376 | -162.5442 | -2.1815 | -2.3777 |
|
| 88 |
+
| 0.0596 | 2.5792 | 806 | 1.2847 | -3.3490 | -3.6206 | 0.4231 | 0.2716 | -144.7193 | -169.8953 | -2.1523 | -2.3458 |
|
| 89 |
+
| 0.0395 | 2.6784 | 837 | 1.3358 | -3.6588 | -3.9010 | 0.4231 | 0.2422 | -147.5237 | -172.9937 | -2.1399 | -2.3346 |
|
| 90 |
+
| 0.0302 | 2.7776 | 868 | 1.3725 | -3.7911 | -4.0386 | 0.4231 | 0.2474 | -148.8990 | -174.3167 | -2.1529 | -2.3475 |
|
| 91 |
+
| 0.0132 | 2.8768 | 899 | 1.4969 | -4.4629 | -4.6237 | 0.4423 | 0.1607 | -154.7499 | -181.0344 | -2.1227 | -2.3178 |
|
| 92 |
+
| 0.034 | 2.976 | 930 | 1.4922 | -4.5129 | -4.6699 | 0.4423 | 0.1570 | -155.2124 | -181.5346 | -2.1203 | -2.3164 |
|
| 93 |
|
| 94 |
|
| 95 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"total_flos": 0.0,
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 9995,
|
| 7 |
-
"train_samples_per_second":
|
| 8 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.9952,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.16162516827830392,
|
| 5 |
+
"train_runtime": 16659.2082,
|
| 6 |
"train_samples": 9995,
|
| 7 |
+
"train_samples_per_second": 1.8,
|
| 8 |
+
"train_steps_per_second": 0.056
|
| 9 |
}
|
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4949453792
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c755a9ddd5397f31a35ae0b7b058e44d821260a840794d592c47c75cce3ed264
|
| 3 |
size 4949453792
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dec74a10ef2bc20f4dd23b9d3702a7b2824030863aee416fada31f261fdd4c26
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4546807800
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6886523874fff7b19f42a5f7cc04b7f16a549108ad3dd468ef266ee193d1a4c9
|
| 3 |
size 4546807800
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"total_flos": 0.0,
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 9995,
|
| 7 |
-
"train_samples_per_second":
|
| 8 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.9952,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.16162516827830392,
|
| 5 |
+
"train_runtime": 16659.2082,
|
| 6 |
"train_samples": 9995,
|
| 7 |
+
"train_samples_per_second": 1.8,
|
| 8 |
+
"train_steps_per_second": 0.056
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -1,17 +1,17 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"eval_steps":
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0032,
|
| 13 |
-
"grad_norm": 104.
|
| 14 |
-
"learning_rate":
|
| 15 |
"logits/generated": -2.7060699462890625,
|
| 16 |
"logits/real": -2.864065647125244,
|
| 17 |
"logps/generated": -135.05746459960938,
|
|
@@ -25,500 +25,1894 @@
|
|
| 25 |
},
|
| 26 |
{
|
| 27 |
"epoch": 0.032,
|
| 28 |
-
"grad_norm":
|
| 29 |
-
"learning_rate":
|
| 30 |
-
"logits/generated": -2.
|
| 31 |
-
"logits/real": -2.
|
| 32 |
-
"logps/generated": -114.
|
| 33 |
-
"logps/real": -
|
| 34 |
-
"loss": 0.
|
| 35 |
-
"rewards/accuracies": 0.
|
| 36 |
-
"rewards/generated": 0.
|
| 37 |
-
"rewards/margins": 0.
|
| 38 |
-
"rewards/real": 0.
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.064,
|
| 43 |
-
"grad_norm":
|
| 44 |
-
"learning_rate":
|
| 45 |
-
"logits/generated": -2.
|
| 46 |
-
"logits/real": -2.
|
| 47 |
-
"logps/generated": -
|
| 48 |
-
"logps/real": -
|
| 49 |
-
"loss": 0.
|
| 50 |
-
"rewards/accuracies": 0.
|
| 51 |
-
"rewards/generated": 0.
|
| 52 |
-
"rewards/margins": 0.
|
| 53 |
-
"rewards/real": 0.
|
| 54 |
"step": 20
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"epoch": 0.096,
|
| 58 |
-
"grad_norm":
|
| 59 |
-
"learning_rate":
|
| 60 |
-
"logits/generated": -2.
|
| 61 |
-
"logits/real": -2.
|
| 62 |
-
"logps/generated": -
|
| 63 |
-
"logps/real": -
|
| 64 |
-
"loss": 0.
|
| 65 |
-
"rewards/accuracies": 0.
|
| 66 |
-
"rewards/generated": 0.
|
| 67 |
-
"rewards/margins": 0.
|
| 68 |
-
"rewards/real":
|
| 69 |
"step": 30
|
| 70 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 71 |
{
|
| 72 |
"epoch": 0.128,
|
| 73 |
-
"grad_norm":
|
| 74 |
-
"learning_rate":
|
| 75 |
-
"logits/generated": -2.
|
| 76 |
-
"logits/real": -2.
|
| 77 |
-
"logps/generated": -
|
| 78 |
-
"logps/real": -
|
| 79 |
-
"loss": 0.
|
| 80 |
-
"rewards/accuracies": 0.
|
| 81 |
-
"rewards/generated":
|
| 82 |
-
"rewards/margins":
|
| 83 |
-
"rewards/real":
|
| 84 |
"step": 40
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 0.16,
|
| 88 |
-
"grad_norm":
|
| 89 |
-
"learning_rate":
|
| 90 |
-
"logits/generated": -2.
|
| 91 |
-
"logits/real": -2.
|
| 92 |
-
"logps/generated": -
|
| 93 |
-
"logps/real": -
|
| 94 |
-
"loss": 0.
|
| 95 |
-
"rewards/accuracies": 0.
|
| 96 |
-
"rewards/generated": 0.
|
| 97 |
-
"rewards/margins":
|
| 98 |
-
"rewards/real":
|
| 99 |
"step": 50
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.192,
|
| 103 |
-
"grad_norm":
|
| 104 |
-
"learning_rate":
|
| 105 |
-
"logits/generated": -2.
|
| 106 |
-
"logits/real": -2.
|
| 107 |
-
"logps/generated": -118.
|
| 108 |
-
"logps/real": -
|
| 109 |
-
"loss": 0.
|
| 110 |
-
"rewards/accuracies": 0.
|
| 111 |
-
"rewards/generated": 1.
|
| 112 |
-
"rewards/margins":
|
| 113 |
-
"rewards/real":
|
| 114 |
"step": 60
|
| 115 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 116 |
{
|
| 117 |
"epoch": 0.224,
|
| 118 |
-
"grad_norm":
|
| 119 |
-
"learning_rate":
|
| 120 |
-
"logits/generated": -2.
|
| 121 |
-
"logits/real": -2.
|
| 122 |
-
"logps/generated": -
|
| 123 |
-
"logps/real": -
|
| 124 |
-
"loss": 0.
|
| 125 |
-
"rewards/accuracies": 0.
|
| 126 |
-
"rewards/generated":
|
| 127 |
-
"rewards/margins":
|
| 128 |
-
"rewards/real": 2.
|
| 129 |
"step": 70
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.256,
|
| 133 |
-
"grad_norm":
|
| 134 |
-
"learning_rate": 4.
|
| 135 |
-
"logits/generated": -2.
|
| 136 |
-
"logits/real": -2.
|
| 137 |
-
"logps/generated": -
|
| 138 |
-
"logps/real": -
|
| 139 |
-
"loss": 0.
|
| 140 |
-
"rewards/accuracies": 0.
|
| 141 |
-
"rewards/generated": 1.
|
| 142 |
-
"rewards/margins":
|
| 143 |
-
"rewards/real":
|
| 144 |
"step": 80
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"epoch": 0.288,
|
| 148 |
-
"grad_norm":
|
| 149 |
-
"learning_rate":
|
| 150 |
-
"logits/generated": -2.
|
| 151 |
-
"logits/real": -2.
|
| 152 |
-
"logps/generated": -
|
| 153 |
-
"logps/real": -
|
| 154 |
-
"loss": 0.
|
| 155 |
-
"rewards/accuracies": 0.
|
| 156 |
-
"rewards/generated":
|
| 157 |
-
"rewards/margins":
|
| 158 |
-
"rewards/real":
|
| 159 |
"step": 90
|
| 160 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 161 |
{
|
| 162 |
"epoch": 0.32,
|
| 163 |
-
"grad_norm":
|
| 164 |
-
"learning_rate":
|
| 165 |
-
"logits/generated": -2.
|
| 166 |
-
"logits/real": -2.
|
| 167 |
-
"logps/generated": -
|
| 168 |
-
"logps/real": -
|
| 169 |
-
"loss": 0.
|
| 170 |
-
"rewards/accuracies": 0.
|
| 171 |
-
"rewards/generated": 0.
|
| 172 |
-
"rewards/margins":
|
| 173 |
-
"rewards/real":
|
| 174 |
"step": 100
|
| 175 |
},
|
| 176 |
{
|
| 177 |
"epoch": 0.352,
|
| 178 |
-
"grad_norm":
|
| 179 |
-
"learning_rate":
|
| 180 |
-
"logits/generated": -2.
|
| 181 |
-
"logits/real": -2.
|
| 182 |
-
"logps/generated": -
|
| 183 |
-
"logps/real": -
|
| 184 |
-
"loss": 0.
|
| 185 |
-
"rewards/accuracies": 0.
|
| 186 |
-
"rewards/generated": 0.
|
| 187 |
-
"rewards/margins":
|
| 188 |
-
"rewards/real":
|
| 189 |
"step": 110
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.384,
|
| 193 |
-
"grad_norm":
|
| 194 |
-
"learning_rate":
|
| 195 |
-
"logits/generated": -2.
|
| 196 |
-
"logits/real": -2.
|
| 197 |
-
"logps/generated": -137.
|
| 198 |
-
"logps/real": -
|
| 199 |
-
"loss": 0.
|
| 200 |
"rewards/accuracies": 0.8500000238418579,
|
| 201 |
-
"rewards/generated": 0.
|
| 202 |
-
"rewards/margins":
|
| 203 |
-
"rewards/real": 3.
|
| 204 |
"step": 120
|
| 205 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 206 |
{
|
| 207 |
"epoch": 0.416,
|
| 208 |
-
"grad_norm":
|
| 209 |
-
"learning_rate":
|
| 210 |
-
"logits/generated": -2.
|
| 211 |
-
"logits/real": -2.
|
| 212 |
-
"logps/generated": -
|
| 213 |
-
"logps/real": -
|
| 214 |
-
"loss": 0.
|
| 215 |
-
"rewards/accuracies": 0.
|
| 216 |
-
"rewards/generated": 0.
|
| 217 |
-
"rewards/margins": 3.
|
| 218 |
-
"rewards/real": 3.
|
| 219 |
"step": 130
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.448,
|
| 223 |
-
"grad_norm":
|
| 224 |
-
"learning_rate":
|
| 225 |
-
"logits/generated": -2.
|
| 226 |
-
"logits/real": -2.
|
| 227 |
-
"logps/generated": -
|
| 228 |
-
"logps/real": -
|
| 229 |
-
"loss": 0.
|
| 230 |
"rewards/accuracies": 0.9125000238418579,
|
| 231 |
-
"rewards/generated":
|
| 232 |
-
"rewards/margins":
|
| 233 |
-
"rewards/real":
|
| 234 |
"step": 140
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 0.48,
|
| 238 |
-
"grad_norm": 73.
|
| 239 |
-
"learning_rate":
|
| 240 |
-
"logits/generated": -2.
|
| 241 |
-
"logits/real": -2.
|
| 242 |
-
"logps/generated": -
|
| 243 |
-
"logps/real": -
|
| 244 |
-
"loss": 0.
|
| 245 |
-
"rewards/accuracies": 0.
|
| 246 |
-
"rewards/generated": -0.
|
| 247 |
-
"rewards/margins": 4.
|
| 248 |
-
"rewards/real":
|
| 249 |
"step": 150
|
| 250 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 251 |
{
|
| 252 |
"epoch": 0.512,
|
| 253 |
-
"grad_norm":
|
| 254 |
-
"learning_rate":
|
| 255 |
-
"logits/generated": -2.
|
| 256 |
-
"logits/real": -2.
|
| 257 |
-
"logps/generated": -
|
| 258 |
-
"logps/real": -
|
| 259 |
-
"loss": 0.
|
| 260 |
"rewards/accuracies": 0.9375,
|
| 261 |
-
"rewards/generated": -0.
|
| 262 |
-
"rewards/margins": 4.
|
| 263 |
-
"rewards/real":
|
| 264 |
"step": 160
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 0.544,
|
| 268 |
-
"grad_norm":
|
| 269 |
-
"learning_rate":
|
| 270 |
-
"logits/generated": -2.
|
| 271 |
-
"logits/real": -2.
|
| 272 |
-
"logps/generated": -
|
| 273 |
-
"logps/real": -
|
| 274 |
-
"loss": 0.
|
| 275 |
-
"rewards/accuracies": 0.
|
| 276 |
-
"rewards/generated": 0.
|
| 277 |
-
"rewards/margins": 3.
|
| 278 |
-
"rewards/real":
|
| 279 |
"step": 170
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.576,
|
| 283 |
-
"grad_norm":
|
| 284 |
-
"learning_rate":
|
| 285 |
-
"logits/generated": -2.
|
| 286 |
-
"logits/real": -2.
|
| 287 |
-
"logps/generated": -135.
|
| 288 |
-
"logps/real": -
|
| 289 |
-
"loss": 0.
|
| 290 |
"rewards/accuracies": 0.9750000238418579,
|
| 291 |
-
"rewards/generated": -1.
|
| 292 |
-
"rewards/margins":
|
| 293 |
-
"rewards/real": 3.
|
| 294 |
"step": 180
|
| 295 |
},
|
| 296 |
{
|
| 297 |
-
"epoch": 0.
|
| 298 |
-
"
|
| 299 |
-
"
|
| 300 |
-
"
|
| 301 |
-
"
|
| 302 |
-
"
|
| 303 |
-
"
|
| 304 |
-
"
|
| 305 |
-
"
|
| 306 |
-
"
|
| 307 |
-
"
|
| 308 |
-
"
|
| 309 |
-
"
|
|
|
|
| 310 |
},
|
| 311 |
{
|
| 312 |
-
"epoch": 0.
|
| 313 |
-
"grad_norm":
|
| 314 |
-
"learning_rate":
|
| 315 |
-
"logits/generated": -2.
|
| 316 |
-
"logits/real": -
|
| 317 |
-
"logps/generated": -
|
| 318 |
-
"logps/real": -
|
| 319 |
-
"loss": 0.
|
| 320 |
-
"rewards/accuracies":
|
| 321 |
-
"rewards/generated": -
|
| 322 |
-
"rewards/margins": 5.
|
| 323 |
-
"rewards/real":
|
| 324 |
-
"step":
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.64,
|
| 328 |
-
"
|
| 329 |
-
"
|
| 330 |
-
"
|
| 331 |
-
"
|
| 332 |
-
"
|
| 333 |
-
"
|
| 334 |
-
"
|
| 335 |
-
"
|
| 336 |
-
"
|
| 337 |
-
"
|
| 338 |
-
"
|
| 339 |
-
"eval_steps_per_second": 0.416,
|
| 340 |
"step": 200
|
| 341 |
},
|
| 342 |
{
|
| 343 |
"epoch": 0.672,
|
| 344 |
-
"grad_norm":
|
| 345 |
-
"learning_rate":
|
| 346 |
-
"logits/generated": -2.
|
| 347 |
-
"logits/real": -2.
|
| 348 |
-
"logps/generated": -139.
|
| 349 |
-
"logps/real": -
|
| 350 |
-
"loss": 0.
|
| 351 |
"rewards/accuracies": 0.9375,
|
| 352 |
-
"rewards/generated": -1.
|
| 353 |
-
"rewards/margins": 5.
|
| 354 |
-
"rewards/real": 4.
|
| 355 |
"step": 210
|
| 356 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 357 |
{
|
| 358 |
"epoch": 0.704,
|
| 359 |
-
"grad_norm":
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"logits/generated": -2.
|
| 362 |
-
"logits/real": -2.
|
| 363 |
-
"logps/generated": -
|
| 364 |
-
"logps/real": -
|
| 365 |
-
"loss": 0.
|
| 366 |
-
"rewards/accuracies": 1.0,
|
| 367 |
-
"rewards/generated": -1.
|
| 368 |
-
"rewards/margins": 6.
|
| 369 |
-
"rewards/real":
|
| 370 |
"step": 220
|
| 371 |
},
|
| 372 |
{
|
| 373 |
"epoch": 0.736,
|
| 374 |
-
"grad_norm":
|
| 375 |
-
"learning_rate":
|
| 376 |
-
"logits/generated": -2.
|
| 377 |
-
"logits/real": -2.
|
| 378 |
-
"logps/generated": -
|
| 379 |
-
"logps/real": -
|
| 380 |
-
"loss": 0.
|
| 381 |
-
"rewards/accuracies": 1.0,
|
| 382 |
-
"rewards/generated": -1.
|
| 383 |
-
"rewards/margins":
|
| 384 |
-
"rewards/real": 4.
|
| 385 |
"step": 230
|
| 386 |
},
|
| 387 |
{
|
| 388 |
"epoch": 0.768,
|
| 389 |
-
"grad_norm":
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"logits/generated": -2.
|
| 392 |
-
"logits/real": -2.
|
| 393 |
-
"logps/generated": -
|
| 394 |
-
"logps/real": -89.
|
| 395 |
-
"loss": 0.
|
| 396 |
-
"rewards/accuracies": 0
|
| 397 |
-
"rewards/generated": -
|
| 398 |
-
"rewards/margins": 6.
|
| 399 |
-
"rewards/real": 4.
|
| 400 |
"step": 240
|
| 401 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 402 |
{
|
| 403 |
"epoch": 0.8,
|
| 404 |
-
"grad_norm":
|
| 405 |
-
"learning_rate":
|
| 406 |
-
"logits/generated": -2.
|
| 407 |
-
"logits/real": -2.
|
| 408 |
-
"logps/generated": -
|
| 409 |
-
"logps/real": -84.
|
| 410 |
-
"loss": 0.
|
| 411 |
"rewards/accuracies": 0.9624999761581421,
|
| 412 |
-
"rewards/generated": -1.
|
| 413 |
-
"rewards/margins": 6.
|
| 414 |
-
"rewards/real": 4.
|
| 415 |
"step": 250
|
| 416 |
},
|
| 417 |
{
|
| 418 |
"epoch": 0.832,
|
| 419 |
-
"grad_norm":
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"logits/generated": -2.
|
| 422 |
-
"logits/real": -2.
|
| 423 |
-
"logps/generated": -
|
| 424 |
-
"logps/real": -
|
| 425 |
-
"loss": 0.
|
| 426 |
"rewards/accuracies": 0.987500011920929,
|
| 427 |
-
"rewards/generated": -
|
| 428 |
-
"rewards/margins":
|
| 429 |
-
"rewards/real":
|
| 430 |
"step": 260
|
| 431 |
},
|
| 432 |
{
|
| 433 |
"epoch": 0.864,
|
| 434 |
-
"grad_norm":
|
| 435 |
-
"learning_rate":
|
| 436 |
-
"logits/generated": -2.
|
| 437 |
-
"logits/real": -2.
|
| 438 |
-
"logps/generated": -
|
| 439 |
-
"logps/real": -
|
| 440 |
-
"loss": 0.
|
| 441 |
"rewards/accuracies": 0.987500011920929,
|
| 442 |
-
"rewards/generated": -
|
| 443 |
-
"rewards/margins":
|
| 444 |
-
"rewards/real":
|
| 445 |
"step": 270
|
| 446 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 447 |
{
|
| 448 |
"epoch": 0.896,
|
| 449 |
-
"grad_norm":
|
| 450 |
-
"learning_rate":
|
| 451 |
-
"logits/generated": -2.
|
| 452 |
-
"logits/real": -2.
|
| 453 |
-
"logps/generated": -
|
| 454 |
-
"logps/real": -
|
| 455 |
-
"loss": 0.
|
| 456 |
-
"rewards/accuracies": 0.
|
| 457 |
-
"rewards/generated": -1.
|
| 458 |
-
"rewards/margins": 6.
|
| 459 |
-
"rewards/real":
|
| 460 |
"step": 280
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 0.928,
|
| 464 |
-
"grad_norm":
|
| 465 |
-
"learning_rate": 3.
|
| 466 |
-
"logits/generated": -2.
|
| 467 |
-
"logits/real": -2.
|
| 468 |
-
"logps/generated": -
|
| 469 |
-
"logps/real": -
|
| 470 |
-
"loss": 0.
|
| 471 |
"rewards/accuracies": 0.9750000238418579,
|
| 472 |
-
"rewards/generated": -
|
| 473 |
-
"rewards/margins":
|
| 474 |
-
"rewards/real": 5.
|
| 475 |
"step": 290
|
| 476 |
},
|
| 477 |
{
|
| 478 |
"epoch": 0.96,
|
| 479 |
-
"grad_norm":
|
| 480 |
-
"learning_rate":
|
| 481 |
-
"logits/generated": -2.
|
| 482 |
-
"logits/real": -2.
|
| 483 |
-
"logps/generated": -
|
| 484 |
-
"logps/real": -
|
| 485 |
-
"loss": 0.
|
| 486 |
-
"rewards/accuracies":
|
| 487 |
-
"rewards/generated": -1.
|
| 488 |
-
"rewards/margins":
|
| 489 |
-
"rewards/real":
|
| 490 |
"step": 300
|
| 491 |
},
|
| 492 |
{
|
| 493 |
"epoch": 0.992,
|
| 494 |
-
"grad_norm":
|
| 495 |
-
"learning_rate": 3.
|
| 496 |
-
"logits/generated": -2.
|
| 497 |
-
"logits/real": -2.
|
| 498 |
-
"logps/generated": -
|
| 499 |
-
"logps/real": -
|
| 500 |
-
"loss": 0.
|
| 501 |
-
"rewards/accuracies": 1.0,
|
| 502 |
-
"rewards/generated": -
|
| 503 |
-
"rewards/margins":
|
| 504 |
-
"rewards/real": 5.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 505 |
"step": 310
|
| 506 |
},
|
| 507 |
{
|
| 508 |
-
"epoch":
|
| 509 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 510 |
"total_flos": 0.0,
|
| 511 |
-
"train_loss": 0.
|
| 512 |
-
"train_runtime":
|
| 513 |
-
"train_samples_per_second":
|
| 514 |
-
"train_steps_per_second": 0.
|
| 515 |
}
|
| 516 |
],
|
| 517 |
"logging_steps": 10,
|
| 518 |
-
"max_steps":
|
| 519 |
"num_input_tokens_seen": 0,
|
| 520 |
-
"num_train_epochs":
|
| 521 |
-
"save_steps":
|
| 522 |
"stateful_callbacks": {
|
| 523 |
"TrainerControl": {
|
| 524 |
"args": {
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9952,
|
| 5 |
+
"eval_steps": 31,
|
| 6 |
+
"global_step": 936,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0032,
|
| 13 |
+
"grad_norm": 104.32759944491974,
|
| 14 |
+
"learning_rate": 5.3191489361702125e-09,
|
| 15 |
"logits/generated": -2.7060699462890625,
|
| 16 |
"logits/real": -2.864065647125244,
|
| 17 |
"logps/generated": -135.05746459960938,
|
|
|
|
| 25 |
},
|
| 26 |
{
|
| 27 |
"epoch": 0.032,
|
| 28 |
+
"grad_norm": 108.69554011112191,
|
| 29 |
+
"learning_rate": 5.3191489361702123e-08,
|
| 30 |
+
"logits/generated": -2.6056575775146484,
|
| 31 |
+
"logits/real": -2.554622173309326,
|
| 32 |
+
"logps/generated": -114.74867248535156,
|
| 33 |
+
"logps/real": -128.5237274169922,
|
| 34 |
+
"loss": 0.8329,
|
| 35 |
+
"rewards/accuracies": 0.5,
|
| 36 |
+
"rewards/generated": 0.0076134600676596165,
|
| 37 |
+
"rewards/margins": -0.0009522127802483737,
|
| 38 |
+
"rewards/real": 0.006661247927695513,
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.064,
|
| 43 |
+
"grad_norm": 104.94374849378612,
|
| 44 |
+
"learning_rate": 1.0638297872340425e-07,
|
| 45 |
+
"logits/generated": -2.67850923538208,
|
| 46 |
+
"logits/real": -2.651012897491455,
|
| 47 |
+
"logps/generated": -120.8056640625,
|
| 48 |
+
"logps/real": -133.23611450195312,
|
| 49 |
+
"loss": 0.8169,
|
| 50 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 51 |
+
"rewards/generated": 0.1307533234357834,
|
| 52 |
+
"rewards/margins": 0.002721247496083379,
|
| 53 |
+
"rewards/real": 0.13347458839416504,
|
| 54 |
"step": 20
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"epoch": 0.096,
|
| 58 |
+
"grad_norm": 96.02197035555018,
|
| 59 |
+
"learning_rate": 1.5957446808510638e-07,
|
| 60 |
+
"logits/generated": -2.677885055541992,
|
| 61 |
+
"logits/real": -2.7564849853515625,
|
| 62 |
+
"logps/generated": -118.22355651855469,
|
| 63 |
+
"logps/real": -132.5462646484375,
|
| 64 |
+
"loss": 0.7877,
|
| 65 |
+
"rewards/accuracies": 0.574999988079071,
|
| 66 |
+
"rewards/generated": 0.34465548396110535,
|
| 67 |
+
"rewards/margins": 0.08713299036026001,
|
| 68 |
+
"rewards/real": 0.43178844451904297,
|
| 69 |
"step": 30
|
| 70 |
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.0992,
|
| 73 |
+
"eval_logits/generated": -2.522256374359131,
|
| 74 |
+
"eval_logits/real": -2.6980574131011963,
|
| 75 |
+
"eval_logps/generated": -103.75227355957031,
|
| 76 |
+
"eval_logps/real": -129.50059509277344,
|
| 77 |
+
"eval_loss": 0.7716670036315918,
|
| 78 |
+
"eval_rewards/accuracies": 0.7307692170143127,
|
| 79 |
+
"eval_rewards/generated": 0.47610312700271606,
|
| 80 |
+
"eval_rewards/margins": 0.2143612504005432,
|
| 81 |
+
"eval_rewards/real": 0.6904643177986145,
|
| 82 |
+
"eval_runtime": 27.37,
|
| 83 |
+
"eval_samples_per_second": 7.307,
|
| 84 |
+
"eval_steps_per_second": 0.475,
|
| 85 |
+
"step": 31
|
| 86 |
+
},
|
| 87 |
{
|
| 88 |
"epoch": 0.128,
|
| 89 |
+
"grad_norm": 95.3097031879678,
|
| 90 |
+
"learning_rate": 2.127659574468085e-07,
|
| 91 |
+
"logits/generated": -2.6211893558502197,
|
| 92 |
+
"logits/real": -2.675510883331299,
|
| 93 |
+
"logps/generated": -126.9448471069336,
|
| 94 |
+
"logps/real": -137.4635467529297,
|
| 95 |
+
"loss": 0.7133,
|
| 96 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 97 |
+
"rewards/generated": 0.4792709946632385,
|
| 98 |
+
"rewards/margins": 0.2818766236305237,
|
| 99 |
+
"rewards/real": 0.7611476182937622,
|
| 100 |
"step": 40
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.16,
|
| 104 |
+
"grad_norm": 106.05520318577042,
|
| 105 |
+
"learning_rate": 2.659574468085106e-07,
|
| 106 |
+
"logits/generated": -2.4675774574279785,
|
| 107 |
+
"logits/real": -2.49176287651062,
|
| 108 |
+
"logps/generated": -124.37459564208984,
|
| 109 |
+
"logps/real": -129.06646728515625,
|
| 110 |
+
"loss": 0.677,
|
| 111 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 112 |
+
"rewards/generated": 0.4745818078517914,
|
| 113 |
+
"rewards/margins": 0.48313817381858826,
|
| 114 |
+
"rewards/real": 0.9577198028564453,
|
| 115 |
"step": 50
|
| 116 |
},
|
| 117 |
{
|
| 118 |
"epoch": 0.192,
|
| 119 |
+
"grad_norm": 84.95894852367661,
|
| 120 |
+
"learning_rate": 3.1914893617021275e-07,
|
| 121 |
+
"logits/generated": -2.628202438354492,
|
| 122 |
+
"logits/real": -2.581429958343506,
|
| 123 |
+
"logps/generated": -118.05242919921875,
|
| 124 |
+
"logps/real": -122.88636779785156,
|
| 125 |
+
"loss": 0.6309,
|
| 126 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 127 |
+
"rewards/generated": 1.081741452217102,
|
| 128 |
+
"rewards/margins": 0.6737821102142334,
|
| 129 |
+
"rewards/real": 1.7555233240127563,
|
| 130 |
"step": 60
|
| 131 |
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.1984,
|
| 134 |
+
"eval_logits/generated": -2.5564322471618652,
|
| 135 |
+
"eval_logits/real": -2.7395834922790527,
|
| 136 |
+
"eval_logps/generated": -91.08578491210938,
|
| 137 |
+
"eval_logps/real": -114.99056243896484,
|
| 138 |
+
"eval_loss": 0.7320010662078857,
|
| 139 |
+
"eval_rewards/accuracies": 0.7115384340286255,
|
| 140 |
+
"eval_rewards/generated": 1.742753505706787,
|
| 141 |
+
"eval_rewards/margins": 0.3987158238887787,
|
| 142 |
+
"eval_rewards/real": 2.141469717025757,
|
| 143 |
+
"eval_runtime": 24.6101,
|
| 144 |
+
"eval_samples_per_second": 8.127,
|
| 145 |
+
"eval_steps_per_second": 0.528,
|
| 146 |
+
"step": 62
|
| 147 |
+
},
|
| 148 |
{
|
| 149 |
"epoch": 0.224,
|
| 150 |
+
"grad_norm": 96.85757291577832,
|
| 151 |
+
"learning_rate": 3.7234042553191484e-07,
|
| 152 |
+
"logits/generated": -2.5130114555358887,
|
| 153 |
+
"logits/real": -2.5616209506988525,
|
| 154 |
+
"logps/generated": -138.7924346923828,
|
| 155 |
+
"logps/real": -111.9786148071289,
|
| 156 |
+
"loss": 0.5514,
|
| 157 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 158 |
+
"rewards/generated": 1.2395274639129639,
|
| 159 |
+
"rewards/margins": 1.0533320903778076,
|
| 160 |
+
"rewards/real": 2.2928595542907715,
|
| 161 |
"step": 70
|
| 162 |
},
|
| 163 |
{
|
| 164 |
"epoch": 0.256,
|
| 165 |
+
"grad_norm": 69.52077411879176,
|
| 166 |
+
"learning_rate": 4.25531914893617e-07,
|
| 167 |
+
"logits/generated": -2.4999165534973145,
|
| 168 |
+
"logits/real": -2.664947509765625,
|
| 169 |
+
"logps/generated": -105.97293853759766,
|
| 170 |
+
"logps/real": -114.73320007324219,
|
| 171 |
+
"loss": 0.5366,
|
| 172 |
+
"rewards/accuracies": 0.8125,
|
| 173 |
+
"rewards/generated": 1.6926729679107666,
|
| 174 |
+
"rewards/margins": 1.063332200050354,
|
| 175 |
+
"rewards/real": 2.75600528717041,
|
| 176 |
"step": 80
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.288,
|
| 180 |
+
"grad_norm": 82.22178145539344,
|
| 181 |
+
"learning_rate": 4.787234042553192e-07,
|
| 182 |
+
"logits/generated": -2.6474456787109375,
|
| 183 |
+
"logits/real": -2.5698373317718506,
|
| 184 |
+
"logps/generated": -135.85374450683594,
|
| 185 |
+
"logps/real": -129.9317626953125,
|
| 186 |
+
"loss": 0.5309,
|
| 187 |
+
"rewards/accuracies": 0.800000011920929,
|
| 188 |
+
"rewards/generated": 1.0643891096115112,
|
| 189 |
+
"rewards/margins": 1.536333441734314,
|
| 190 |
+
"rewards/real": 2.6007227897644043,
|
| 191 |
"step": 90
|
| 192 |
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2976,
|
| 195 |
+
"eval_logits/generated": -2.4702792167663574,
|
| 196 |
+
"eval_logits/real": -2.6170670986175537,
|
| 197 |
+
"eval_logps/generated": -99.49687957763672,
|
| 198 |
+
"eval_logps/real": -120.69674682617188,
|
| 199 |
+
"eval_loss": 0.7175261974334717,
|
| 200 |
+
"eval_rewards/accuracies": 0.6538461446762085,
|
| 201 |
+
"eval_rewards/generated": 0.9016433358192444,
|
| 202 |
+
"eval_rewards/margins": 0.6692073345184326,
|
| 203 |
+
"eval_rewards/real": 1.5708506107330322,
|
| 204 |
+
"eval_runtime": 24.5423,
|
| 205 |
+
"eval_samples_per_second": 8.149,
|
| 206 |
+
"eval_steps_per_second": 0.53,
|
| 207 |
+
"step": 93
|
| 208 |
+
},
|
| 209 |
{
|
| 210 |
"epoch": 0.32,
|
| 211 |
+
"grad_norm": 67.99154678016933,
|
| 212 |
+
"learning_rate": 4.96437054631829e-07,
|
| 213 |
+
"logits/generated": -2.6401119232177734,
|
| 214 |
+
"logits/real": -2.613041400909424,
|
| 215 |
+
"logps/generated": -123.557861328125,
|
| 216 |
+
"logps/real": -129.47727966308594,
|
| 217 |
+
"loss": 0.4506,
|
| 218 |
+
"rewards/accuracies": 0.800000011920929,
|
| 219 |
+
"rewards/generated": 0.6703820824623108,
|
| 220 |
+
"rewards/margins": 1.8516438007354736,
|
| 221 |
+
"rewards/real": 2.5220255851745605,
|
| 222 |
"step": 100
|
| 223 |
},
|
| 224 |
{
|
| 225 |
"epoch": 0.352,
|
| 226 |
+
"grad_norm": 74.51996731564344,
|
| 227 |
+
"learning_rate": 4.904988123515439e-07,
|
| 228 |
+
"logits/generated": -2.6040279865264893,
|
| 229 |
+
"logits/real": -2.6699700355529785,
|
| 230 |
+
"logps/generated": -155.3791961669922,
|
| 231 |
+
"logps/real": -119.61136627197266,
|
| 232 |
+
"loss": 0.3797,
|
| 233 |
+
"rewards/accuracies": 0.887499988079071,
|
| 234 |
+
"rewards/generated": -0.05242754891514778,
|
| 235 |
+
"rewards/margins": 2.9086718559265137,
|
| 236 |
+
"rewards/real": 2.8562443256378174,
|
| 237 |
"step": 110
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.384,
|
| 241 |
+
"grad_norm": 90.56758343390324,
|
| 242 |
+
"learning_rate": 4.845605700712589e-07,
|
| 243 |
+
"logits/generated": -2.8406381607055664,
|
| 244 |
+
"logits/real": -2.7864930629730225,
|
| 245 |
+
"logps/generated": -137.2823028564453,
|
| 246 |
+
"logps/real": -111.62813568115234,
|
| 247 |
+
"loss": 0.4323,
|
| 248 |
"rewards/accuracies": 0.8500000238418579,
|
| 249 |
+
"rewards/generated": 0.4588840901851654,
|
| 250 |
+
"rewards/margins": 2.8661108016967773,
|
| 251 |
+
"rewards/real": 3.3249950408935547,
|
| 252 |
"step": 120
|
| 253 |
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 0.3968,
|
| 256 |
+
"eval_logits/generated": -2.6808340549468994,
|
| 257 |
+
"eval_logits/real": -2.834907293319702,
|
| 258 |
+
"eval_logps/generated": -93.77437591552734,
|
| 259 |
+
"eval_logps/real": -116.81954193115234,
|
| 260 |
+
"eval_loss": 0.7713803648948669,
|
| 261 |
+
"eval_rewards/accuracies": 0.692307710647583,
|
| 262 |
+
"eval_rewards/generated": 1.4738935232162476,
|
| 263 |
+
"eval_rewards/margins": 0.4846768081188202,
|
| 264 |
+
"eval_rewards/real": 1.9585703611373901,
|
| 265 |
+
"eval_runtime": 25.6959,
|
| 266 |
+
"eval_samples_per_second": 7.783,
|
| 267 |
+
"eval_steps_per_second": 0.506,
|
| 268 |
+
"step": 124
|
| 269 |
+
},
|
| 270 |
{
|
| 271 |
"epoch": 0.416,
|
| 272 |
+
"grad_norm": 41.82001394961611,
|
| 273 |
+
"learning_rate": 4.786223277909738e-07,
|
| 274 |
+
"logits/generated": -2.6804909706115723,
|
| 275 |
+
"logits/real": -2.8820011615753174,
|
| 276 |
+
"logps/generated": -121.76007080078125,
|
| 277 |
+
"logps/real": -104.3374252319336,
|
| 278 |
+
"loss": 0.3483,
|
| 279 |
+
"rewards/accuracies": 0.925000011920929,
|
| 280 |
+
"rewards/generated": 0.10464005172252655,
|
| 281 |
+
"rewards/margins": 3.1740787029266357,
|
| 282 |
+
"rewards/real": 3.278719425201416,
|
| 283 |
"step": 130
|
| 284 |
},
|
| 285 |
{
|
| 286 |
"epoch": 0.448,
|
| 287 |
+
"grad_norm": 55.13818556944925,
|
| 288 |
+
"learning_rate": 4.7268408551068883e-07,
|
| 289 |
+
"logits/generated": -2.637974977493286,
|
| 290 |
+
"logits/real": -2.62568998336792,
|
| 291 |
+
"logps/generated": -138.10079956054688,
|
| 292 |
+
"logps/real": -127.15523529052734,
|
| 293 |
+
"loss": 0.3623,
|
| 294 |
"rewards/accuracies": 0.9125000238418579,
|
| 295 |
+
"rewards/generated": 0.21112258732318878,
|
| 296 |
+
"rewards/margins": 3.419811725616455,
|
| 297 |
+
"rewards/real": 3.630934238433838,
|
| 298 |
"step": 140
|
| 299 |
},
|
| 300 |
{
|
| 301 |
"epoch": 0.48,
|
| 302 |
+
"grad_norm": 73.76709240088495,
|
| 303 |
+
"learning_rate": 4.667458432304038e-07,
|
| 304 |
+
"logits/generated": -2.753176212310791,
|
| 305 |
+
"logits/real": -2.7446675300598145,
|
| 306 |
+
"logps/generated": -147.35614013671875,
|
| 307 |
+
"logps/real": -124.41835021972656,
|
| 308 |
+
"loss": 0.297,
|
| 309 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 310 |
+
"rewards/generated": -0.5855969190597534,
|
| 311 |
+
"rewards/margins": 4.5017008781433105,
|
| 312 |
+
"rewards/real": 3.9161038398742676,
|
| 313 |
"step": 150
|
| 314 |
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.496,
|
| 317 |
+
"eval_logits/generated": -2.625582456588745,
|
| 318 |
+
"eval_logits/real": -2.76962947845459,
|
| 319 |
+
"eval_logps/generated": -90.96475219726562,
|
| 320 |
+
"eval_logps/real": -113.50178527832031,
|
| 321 |
+
"eval_loss": 0.7160908579826355,
|
| 322 |
+
"eval_rewards/accuracies": 0.807692289352417,
|
| 323 |
+
"eval_rewards/generated": 1.7548563480377197,
|
| 324 |
+
"eval_rewards/margins": 0.5354906916618347,
|
| 325 |
+
"eval_rewards/real": 2.290347099304199,
|
| 326 |
+
"eval_runtime": 25.5334,
|
| 327 |
+
"eval_samples_per_second": 7.833,
|
| 328 |
+
"eval_steps_per_second": 0.509,
|
| 329 |
+
"step": 155
|
| 330 |
+
},
|
| 331 |
{
|
| 332 |
"epoch": 0.512,
|
| 333 |
+
"grad_norm": 124.72244829757992,
|
| 334 |
+
"learning_rate": 4.6080760095011875e-07,
|
| 335 |
+
"logits/generated": -2.7324862480163574,
|
| 336 |
+
"logits/real": -2.8079638481140137,
|
| 337 |
+
"logps/generated": -128.80862426757812,
|
| 338 |
+
"logps/real": -107.00062561035156,
|
| 339 |
+
"loss": 0.2956,
|
| 340 |
"rewards/accuracies": 0.9375,
|
| 341 |
+
"rewards/generated": -0.16917666792869568,
|
| 342 |
+
"rewards/margins": 4.109906196594238,
|
| 343 |
+
"rewards/real": 3.940729856491089,
|
| 344 |
"step": 160
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.544,
|
| 348 |
+
"grad_norm": 23.575152887764823,
|
| 349 |
+
"learning_rate": 4.548693586698337e-07,
|
| 350 |
+
"logits/generated": -2.5645294189453125,
|
| 351 |
+
"logits/real": -2.688098430633545,
|
| 352 |
+
"logps/generated": -122.01734924316406,
|
| 353 |
+
"logps/real": -102.08761596679688,
|
| 354 |
+
"loss": 0.2795,
|
| 355 |
+
"rewards/accuracies": 0.949999988079071,
|
| 356 |
+
"rewards/generated": -0.07551076263189316,
|
| 357 |
+
"rewards/margins": 3.8361549377441406,
|
| 358 |
+
"rewards/real": 3.7606441974639893,
|
| 359 |
"step": 170
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 0.576,
|
| 363 |
+
"grad_norm": 32.24179696117588,
|
| 364 |
+
"learning_rate": 4.4893111638954866e-07,
|
| 365 |
+
"logits/generated": -2.7335903644561768,
|
| 366 |
+
"logits/real": -2.827608823776245,
|
| 367 |
+
"logps/generated": -135.81558227539062,
|
| 368 |
+
"logps/real": -96.14201354980469,
|
| 369 |
+
"loss": 0.2144,
|
| 370 |
"rewards/accuracies": 0.9750000238418579,
|
| 371 |
+
"rewards/generated": -1.2115757465362549,
|
| 372 |
+
"rewards/margins": 4.802846908569336,
|
| 373 |
+
"rewards/real": 3.591270923614502,
|
| 374 |
"step": 180
|
| 375 |
},
|
| 376 |
{
|
| 377 |
+
"epoch": 0.5952,
|
| 378 |
+
"eval_logits/generated": -2.8900225162506104,
|
| 379 |
+
"eval_logits/real": -3.0601844787597656,
|
| 380 |
+
"eval_logps/generated": -98.30004119873047,
|
| 381 |
+
"eval_logps/real": -120.36709594726562,
|
| 382 |
+
"eval_loss": 0.8256979584693909,
|
| 383 |
+
"eval_rewards/accuracies": 0.7115384340286255,
|
| 384 |
+
"eval_rewards/generated": 1.0213286876678467,
|
| 385 |
+
"eval_rewards/margins": 0.5824873447418213,
|
| 386 |
+
"eval_rewards/real": 1.6038157939910889,
|
| 387 |
+
"eval_runtime": 24.9055,
|
| 388 |
+
"eval_samples_per_second": 8.03,
|
| 389 |
+
"eval_steps_per_second": 0.522,
|
| 390 |
+
"step": 186
|
| 391 |
},
|
| 392 |
{
|
| 393 |
+
"epoch": 0.608,
|
| 394 |
+
"grad_norm": 41.48311970139271,
|
| 395 |
+
"learning_rate": 4.429928741092636e-07,
|
| 396 |
+
"logits/generated": -2.9662532806396484,
|
| 397 |
+
"logits/real": -3.030586004257202,
|
| 398 |
+
"logps/generated": -152.9399871826172,
|
| 399 |
+
"logps/real": -105.7096176147461,
|
| 400 |
+
"loss": 0.1739,
|
| 401 |
+
"rewards/accuracies": 0.987500011920929,
|
| 402 |
+
"rewards/generated": -2.151095390319824,
|
| 403 |
+
"rewards/margins": 5.8033766746521,
|
| 404 |
+
"rewards/real": 3.6522812843322754,
|
| 405 |
+
"step": 190
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 0.64,
|
| 409 |
+
"grad_norm": 54.1414198416636,
|
| 410 |
+
"learning_rate": 4.3705463182897863e-07,
|
| 411 |
+
"logits/generated": -2.9348952770233154,
|
| 412 |
+
"logits/real": -3.019658327102661,
|
| 413 |
+
"logps/generated": -130.03958129882812,
|
| 414 |
+
"logps/real": -91.68030548095703,
|
| 415 |
+
"loss": 0.2408,
|
| 416 |
+
"rewards/accuracies": 0.987500011920929,
|
| 417 |
+
"rewards/generated": -1.0477750301361084,
|
| 418 |
+
"rewards/margins": 5.414088249206543,
|
| 419 |
+
"rewards/real": 4.36631441116333,
|
|
|
|
| 420 |
"step": 200
|
| 421 |
},
|
| 422 |
{
|
| 423 |
"epoch": 0.672,
|
| 424 |
+
"grad_norm": 46.728112406251626,
|
| 425 |
+
"learning_rate": 4.311163895486936e-07,
|
| 426 |
+
"logits/generated": -2.939383029937744,
|
| 427 |
+
"logits/real": -2.944849967956543,
|
| 428 |
+
"logps/generated": -139.65420532226562,
|
| 429 |
+
"logps/real": -83.42400360107422,
|
| 430 |
+
"loss": 0.2497,
|
| 431 |
"rewards/accuracies": 0.9375,
|
| 432 |
+
"rewards/generated": -1.1235296726226807,
|
| 433 |
+
"rewards/margins": 5.62716007232666,
|
| 434 |
+
"rewards/real": 4.5036301612854,
|
| 435 |
"step": 210
|
| 436 |
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.6944,
|
| 439 |
+
"eval_logits/generated": -2.6469037532806396,
|
| 440 |
+
"eval_logits/real": -2.8200910091400146,
|
| 441 |
+
"eval_logps/generated": -90.68226623535156,
|
| 442 |
+
"eval_logps/real": -111.86186981201172,
|
| 443 |
+
"eval_loss": 0.6848626732826233,
|
| 444 |
+
"eval_rewards/accuracies": 0.807692289352417,
|
| 445 |
+
"eval_rewards/generated": 1.7831059694290161,
|
| 446 |
+
"eval_rewards/margins": 0.6712326407432556,
|
| 447 |
+
"eval_rewards/real": 2.454339027404785,
|
| 448 |
+
"eval_runtime": 24.5749,
|
| 449 |
+
"eval_samples_per_second": 8.138,
|
| 450 |
+
"eval_steps_per_second": 0.529,
|
| 451 |
+
"step": 217
|
| 452 |
+
},
|
| 453 |
{
|
| 454 |
"epoch": 0.704,
|
| 455 |
+
"grad_norm": 20.434906367353744,
|
| 456 |
+
"learning_rate": 4.251781472684085e-07,
|
| 457 |
+
"logits/generated": -2.7813591957092285,
|
| 458 |
+
"logits/real": -2.8630871772766113,
|
| 459 |
+
"logps/generated": -147.4626007080078,
|
| 460 |
+
"logps/real": -120.3201675415039,
|
| 461 |
+
"loss": 0.1744,
|
| 462 |
+
"rewards/accuracies": 1.0,
|
| 463 |
+
"rewards/generated": -1.358475923538208,
|
| 464 |
+
"rewards/margins": 6.355828285217285,
|
| 465 |
+
"rewards/real": 4.9973530769348145,
|
| 466 |
"step": 220
|
| 467 |
},
|
| 468 |
{
|
| 469 |
"epoch": 0.736,
|
| 470 |
+
"grad_norm": 17.208585033734995,
|
| 471 |
+
"learning_rate": 4.192399049881235e-07,
|
| 472 |
+
"logits/generated": -2.753448247909546,
|
| 473 |
+
"logits/real": -2.8509767055511475,
|
| 474 |
+
"logps/generated": -139.57362365722656,
|
| 475 |
+
"logps/real": -95.3445053100586,
|
| 476 |
+
"loss": 0.1283,
|
| 477 |
+
"rewards/accuracies": 1.0,
|
| 478 |
+
"rewards/generated": -1.6594841480255127,
|
| 479 |
+
"rewards/margins": 6.16472864151001,
|
| 480 |
+
"rewards/real": 4.505244255065918,
|
| 481 |
"step": 230
|
| 482 |
},
|
| 483 |
{
|
| 484 |
"epoch": 0.768,
|
| 485 |
+
"grad_norm": 13.206387425612215,
|
| 486 |
+
"learning_rate": 4.1330166270783846e-07,
|
| 487 |
+
"logits/generated": -2.7692253589630127,
|
| 488 |
+
"logits/real": -2.7172296047210693,
|
| 489 |
+
"logps/generated": -154.8363800048828,
|
| 490 |
+
"logps/real": -89.86678314208984,
|
| 491 |
+
"loss": 0.1112,
|
| 492 |
+
"rewards/accuracies": 1.0,
|
| 493 |
+
"rewards/generated": -2.302570343017578,
|
| 494 |
+
"rewards/margins": 6.421963691711426,
|
| 495 |
+
"rewards/real": 4.1193928718566895,
|
| 496 |
"step": 240
|
| 497 |
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 0.7936,
|
| 500 |
+
"eval_logits/generated": -2.7020058631896973,
|
| 501 |
+
"eval_logits/real": -2.864475727081299,
|
| 502 |
+
"eval_logps/generated": -94.19103240966797,
|
| 503 |
+
"eval_logps/real": -113.57474517822266,
|
| 504 |
+
"eval_loss": 0.6993240118026733,
|
| 505 |
+
"eval_rewards/accuracies": 0.7884615659713745,
|
| 506 |
+
"eval_rewards/generated": 1.4322292804718018,
|
| 507 |
+
"eval_rewards/margins": 0.8508216142654419,
|
| 508 |
+
"eval_rewards/real": 2.283051013946533,
|
| 509 |
+
"eval_runtime": 25.5028,
|
| 510 |
+
"eval_samples_per_second": 7.842,
|
| 511 |
+
"eval_steps_per_second": 0.51,
|
| 512 |
+
"step": 248
|
| 513 |
+
},
|
| 514 |
{
|
| 515 |
"epoch": 0.8,
|
| 516 |
+
"grad_norm": 9.621166462092326,
|
| 517 |
+
"learning_rate": 4.0736342042755347e-07,
|
| 518 |
+
"logits/generated": -2.7592086791992188,
|
| 519 |
+
"logits/real": -2.8254053592681885,
|
| 520 |
+
"logps/generated": -137.39413452148438,
|
| 521 |
+
"logps/real": -84.954833984375,
|
| 522 |
+
"loss": 0.1385,
|
| 523 |
"rewards/accuracies": 0.9624999761581421,
|
| 524 |
+
"rewards/generated": -1.9928325414657593,
|
| 525 |
+
"rewards/margins": 6.723714351654053,
|
| 526 |
+
"rewards/real": 4.730882167816162,
|
| 527 |
"step": 250
|
| 528 |
},
|
| 529 |
{
|
| 530 |
"epoch": 0.832,
|
| 531 |
+
"grad_norm": 37.766946962270424,
|
| 532 |
+
"learning_rate": 4.0142517814726837e-07,
|
| 533 |
+
"logits/generated": -2.855225086212158,
|
| 534 |
+
"logits/real": -2.8763980865478516,
|
| 535 |
+
"logps/generated": -165.6865997314453,
|
| 536 |
+
"logps/real": -84.22928619384766,
|
| 537 |
+
"loss": 0.1199,
|
| 538 |
"rewards/accuracies": 0.987500011920929,
|
| 539 |
+
"rewards/generated": -2.355945587158203,
|
| 540 |
+
"rewards/margins": 7.509199619293213,
|
| 541 |
+
"rewards/real": 5.153253555297852,
|
| 542 |
"step": 260
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 0.864,
|
| 546 |
+
"grad_norm": 30.45497571981933,
|
| 547 |
+
"learning_rate": 3.9548693586698333e-07,
|
| 548 |
+
"logits/generated": -2.787148952484131,
|
| 549 |
+
"logits/real": -2.8248982429504395,
|
| 550 |
+
"logps/generated": -143.9623260498047,
|
| 551 |
+
"logps/real": -72.55261993408203,
|
| 552 |
+
"loss": 0.176,
|
| 553 |
"rewards/accuracies": 0.987500011920929,
|
| 554 |
+
"rewards/generated": -2.554567813873291,
|
| 555 |
+
"rewards/margins": 7.758988857269287,
|
| 556 |
+
"rewards/real": 5.204421043395996,
|
| 557 |
"step": 270
|
| 558 |
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.8928,
|
| 561 |
+
"eval_logits/generated": -2.805142641067505,
|
| 562 |
+
"eval_logits/real": -2.9279673099517822,
|
| 563 |
+
"eval_logps/generated": -86.06629180908203,
|
| 564 |
+
"eval_logps/real": -108.56407165527344,
|
| 565 |
+
"eval_loss": 0.6700232028961182,
|
| 566 |
+
"eval_rewards/accuracies": 0.7692307829856873,
|
| 567 |
+
"eval_rewards/generated": 2.2447028160095215,
|
| 568 |
+
"eval_rewards/margins": 0.5394153594970703,
|
| 569 |
+
"eval_rewards/real": 2.784118175506592,
|
| 570 |
+
"eval_runtime": 25.461,
|
| 571 |
+
"eval_samples_per_second": 7.855,
|
| 572 |
+
"eval_steps_per_second": 0.511,
|
| 573 |
+
"step": 279
|
| 574 |
+
},
|
| 575 |
{
|
| 576 |
"epoch": 0.896,
|
| 577 |
+
"grad_norm": 12.11949130938464,
|
| 578 |
+
"learning_rate": 3.8954869358669834e-07,
|
| 579 |
+
"logits/generated": -2.8453097343444824,
|
| 580 |
+
"logits/real": -2.9081547260284424,
|
| 581 |
+
"logps/generated": -130.38711547851562,
|
| 582 |
+
"logps/real": -80.25016784667969,
|
| 583 |
+
"loss": 0.2256,
|
| 584 |
+
"rewards/accuracies": 0.987500011920929,
|
| 585 |
+
"rewards/generated": -1.7986886501312256,
|
| 586 |
+
"rewards/margins": 6.868111610412598,
|
| 587 |
+
"rewards/real": 5.069422721862793,
|
| 588 |
"step": 280
|
| 589 |
},
|
| 590 |
{
|
| 591 |
"epoch": 0.928,
|
| 592 |
+
"grad_norm": 8.258057690722529,
|
| 593 |
+
"learning_rate": 3.836104513064133e-07,
|
| 594 |
+
"logits/generated": -2.855512857437134,
|
| 595 |
+
"logits/real": -2.82855224609375,
|
| 596 |
+
"logps/generated": -147.15182495117188,
|
| 597 |
+
"logps/real": -93.78026580810547,
|
| 598 |
+
"loss": 0.144,
|
| 599 |
"rewards/accuracies": 0.9750000238418579,
|
| 600 |
+
"rewards/generated": -2.1502671241760254,
|
| 601 |
+
"rewards/margins": 7.681868553161621,
|
| 602 |
+
"rewards/real": 5.531601905822754,
|
| 603 |
"step": 290
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 0.96,
|
| 607 |
+
"grad_norm": 19.59092219958014,
|
| 608 |
+
"learning_rate": 3.7767220902612825e-07,
|
| 609 |
+
"logits/generated": -2.7956345081329346,
|
| 610 |
+
"logits/real": -2.873706817626953,
|
| 611 |
+
"logps/generated": -146.74185180664062,
|
| 612 |
+
"logps/real": -91.1478500366211,
|
| 613 |
+
"loss": 0.15,
|
| 614 |
+
"rewards/accuracies": 0.987500011920929,
|
| 615 |
+
"rewards/generated": -1.9168081283569336,
|
| 616 |
+
"rewards/margins": 8.325960159301758,
|
| 617 |
+
"rewards/real": 6.40915060043335,
|
| 618 |
"step": 300
|
| 619 |
},
|
| 620 |
{
|
| 621 |
"epoch": 0.992,
|
| 622 |
+
"grad_norm": 15.834799965250946,
|
| 623 |
+
"learning_rate": 3.717339667458432e-07,
|
| 624 |
+
"logits/generated": -2.8749289512634277,
|
| 625 |
+
"logits/real": -2.876892566680908,
|
| 626 |
+
"logps/generated": -163.785400390625,
|
| 627 |
+
"logps/real": -88.39195251464844,
|
| 628 |
+
"loss": 0.1135,
|
| 629 |
+
"rewards/accuracies": 1.0,
|
| 630 |
+
"rewards/generated": -2.48066782951355,
|
| 631 |
+
"rewards/margins": 8.187207221984863,
|
| 632 |
+
"rewards/real": 5.706539154052734,
|
| 633 |
+
"step": 310
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.992,
|
| 637 |
+
"eval_logits/generated": -2.802443027496338,
|
| 638 |
+
"eval_logits/real": -2.9202871322631836,
|
| 639 |
+
"eval_logps/generated": -90.31492614746094,
|
| 640 |
+
"eval_logps/real": -112.55606842041016,
|
| 641 |
+
"eval_loss": 0.6956190466880798,
|
| 642 |
+
"eval_rewards/accuracies": 0.7884615659713745,
|
| 643 |
+
"eval_rewards/generated": 1.8198401927947998,
|
| 644 |
+
"eval_rewards/margins": 0.565077543258667,
|
| 645 |
+
"eval_rewards/real": 2.384917974472046,
|
| 646 |
+
"eval_runtime": 25.3747,
|
| 647 |
+
"eval_samples_per_second": 7.882,
|
| 648 |
+
"eval_steps_per_second": 0.512,
|
| 649 |
"step": 310
|
| 650 |
},
|
| 651 |
{
|
| 652 |
+
"epoch": 1.024,
|
| 653 |
+
"grad_norm": 3.2127566683583244,
|
| 654 |
+
"learning_rate": 3.6579572446555817e-07,
|
| 655 |
+
"logits/generated": -2.8369505405426025,
|
| 656 |
+
"logits/real": -2.8505897521972656,
|
| 657 |
+
"logps/generated": -159.44688415527344,
|
| 658 |
+
"logps/real": -78.3135986328125,
|
| 659 |
+
"loss": 0.1157,
|
| 660 |
+
"rewards/accuracies": 1.0,
|
| 661 |
+
"rewards/generated": -2.547872543334961,
|
| 662 |
+
"rewards/margins": 8.247138023376465,
|
| 663 |
+
"rewards/real": 5.6992669105529785,
|
| 664 |
+
"step": 320
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.056,
|
| 668 |
+
"grad_norm": 5.785222044107326,
|
| 669 |
+
"learning_rate": 3.598574821852731e-07,
|
| 670 |
+
"logits/generated": -2.6663429737091064,
|
| 671 |
+
"logits/real": -2.7017171382904053,
|
| 672 |
+
"logps/generated": -150.9597625732422,
|
| 673 |
+
"logps/real": -76.73829650878906,
|
| 674 |
+
"loss": 0.1132,
|
| 675 |
+
"rewards/accuracies": 1.0,
|
| 676 |
+
"rewards/generated": -2.2275402545928955,
|
| 677 |
+
"rewards/margins": 8.38085651397705,
|
| 678 |
+
"rewards/real": 6.153315544128418,
|
| 679 |
+
"step": 330
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.088,
|
| 683 |
+
"grad_norm": 2.5305457609113042,
|
| 684 |
+
"learning_rate": 3.5391923990498813e-07,
|
| 685 |
+
"logits/generated": -2.5699503421783447,
|
| 686 |
+
"logits/real": -2.605731964111328,
|
| 687 |
+
"logps/generated": -153.1111297607422,
|
| 688 |
+
"logps/real": -74.11537170410156,
|
| 689 |
+
"loss": 0.1221,
|
| 690 |
+
"rewards/accuracies": 1.0,
|
| 691 |
+
"rewards/generated": -2.6568379402160645,
|
| 692 |
+
"rewards/margins": 8.866227149963379,
|
| 693 |
+
"rewards/real": 6.209389686584473,
|
| 694 |
+
"step": 340
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 1.0912,
|
| 698 |
+
"eval_logits/generated": -2.5885517597198486,
|
| 699 |
+
"eval_logits/real": -2.7364766597747803,
|
| 700 |
+
"eval_logps/generated": -92.37077331542969,
|
| 701 |
+
"eval_logps/real": -114.35930633544922,
|
| 702 |
+
"eval_loss": 0.7313989400863647,
|
| 703 |
+
"eval_rewards/accuracies": 0.7307692170143127,
|
| 704 |
+
"eval_rewards/generated": 1.6142545938491821,
|
| 705 |
+
"eval_rewards/margins": 0.5903396010398865,
|
| 706 |
+
"eval_rewards/real": 2.204594612121582,
|
| 707 |
+
"eval_runtime": 24.3069,
|
| 708 |
+
"eval_samples_per_second": 8.228,
|
| 709 |
+
"eval_steps_per_second": 0.535,
|
| 710 |
+
"step": 341
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.12,
|
| 714 |
+
"grad_norm": 3.541237222198075,
|
| 715 |
+
"learning_rate": 3.479809976247031e-07,
|
| 716 |
+
"logits/generated": -2.6312408447265625,
|
| 717 |
+
"logits/real": -2.6744143962860107,
|
| 718 |
+
"logps/generated": -142.3855743408203,
|
| 719 |
+
"logps/real": -71.9720230102539,
|
| 720 |
+
"loss": 0.108,
|
| 721 |
+
"rewards/accuracies": 0.987500011920929,
|
| 722 |
+
"rewards/generated": -2.536757469177246,
|
| 723 |
+
"rewards/margins": 8.74314022064209,
|
| 724 |
+
"rewards/real": 6.206382751464844,
|
| 725 |
+
"step": 350
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 1.152,
|
| 729 |
+
"grad_norm": 3.62080901586099,
|
| 730 |
+
"learning_rate": 3.42042755344418e-07,
|
| 731 |
+
"logits/generated": -2.5982825756073,
|
| 732 |
+
"logits/real": -2.579300880432129,
|
| 733 |
+
"logps/generated": -146.4971466064453,
|
| 734 |
+
"logps/real": -72.2068099975586,
|
| 735 |
+
"loss": 0.0771,
|
| 736 |
+
"rewards/accuracies": 1.0,
|
| 737 |
+
"rewards/generated": -2.3290672302246094,
|
| 738 |
+
"rewards/margins": 9.054376602172852,
|
| 739 |
+
"rewards/real": 6.7253098487854,
|
| 740 |
+
"step": 360
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.184,
|
| 744 |
+
"grad_norm": 19.96975206692693,
|
| 745 |
+
"learning_rate": 3.36104513064133e-07,
|
| 746 |
+
"logits/generated": -2.5062308311462402,
|
| 747 |
+
"logits/real": -2.631631851196289,
|
| 748 |
+
"logps/generated": -162.20138549804688,
|
| 749 |
+
"logps/real": -84.9935531616211,
|
| 750 |
+
"loss": 0.0864,
|
| 751 |
+
"rewards/accuracies": 1.0,
|
| 752 |
+
"rewards/generated": -3.1360862255096436,
|
| 753 |
+
"rewards/margins": 9.973578453063965,
|
| 754 |
+
"rewards/real": 6.837491512298584,
|
| 755 |
+
"step": 370
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.1904,
|
| 759 |
+
"eval_logits/generated": -2.535508871078491,
|
| 760 |
+
"eval_logits/real": -2.7014498710632324,
|
| 761 |
+
"eval_logps/generated": -89.05432891845703,
|
| 762 |
+
"eval_logps/real": -113.19943237304688,
|
| 763 |
+
"eval_loss": 0.7718254327774048,
|
| 764 |
+
"eval_rewards/accuracies": 0.6346153616905212,
|
| 765 |
+
"eval_rewards/generated": 1.945899248123169,
|
| 766 |
+
"eval_rewards/margins": 0.37468206882476807,
|
| 767 |
+
"eval_rewards/real": 2.3205814361572266,
|
| 768 |
+
"eval_runtime": 25.3769,
|
| 769 |
+
"eval_samples_per_second": 7.881,
|
| 770 |
+
"eval_steps_per_second": 0.512,
|
| 771 |
+
"step": 372
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 1.216,
|
| 775 |
+
"grad_norm": 1.77932953998119,
|
| 776 |
+
"learning_rate": 3.3016627078384796e-07,
|
| 777 |
+
"logits/generated": -2.683812379837036,
|
| 778 |
+
"logits/real": -2.721485137939453,
|
| 779 |
+
"logps/generated": -150.50462341308594,
|
| 780 |
+
"logps/real": -63.769615173339844,
|
| 781 |
+
"loss": 0.1072,
|
| 782 |
+
"rewards/accuracies": 1.0,
|
| 783 |
+
"rewards/generated": -2.774550437927246,
|
| 784 |
+
"rewards/margins": 9.467794418334961,
|
| 785 |
+
"rewards/real": 6.693244934082031,
|
| 786 |
+
"step": 380
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 1.248,
|
| 790 |
+
"grad_norm": 2.282468411046649,
|
| 791 |
+
"learning_rate": 3.2422802850356297e-07,
|
| 792 |
+
"logits/generated": -2.680248260498047,
|
| 793 |
+
"logits/real": -2.712040424346924,
|
| 794 |
+
"logps/generated": -159.68753051757812,
|
| 795 |
+
"logps/real": -74.8614501953125,
|
| 796 |
+
"loss": 0.0841,
|
| 797 |
+
"rewards/accuracies": 1.0,
|
| 798 |
+
"rewards/generated": -3.0862269401550293,
|
| 799 |
+
"rewards/margins": 10.16987133026123,
|
| 800 |
+
"rewards/real": 7.083643913269043,
|
| 801 |
+
"step": 390
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 1.28,
|
| 805 |
+
"grad_norm": 7.624913421372154,
|
| 806 |
+
"learning_rate": 3.182897862232779e-07,
|
| 807 |
+
"logits/generated": -2.617255926132202,
|
| 808 |
+
"logits/real": -2.6958706378936768,
|
| 809 |
+
"logps/generated": -152.98448181152344,
|
| 810 |
+
"logps/real": -74.80540466308594,
|
| 811 |
+
"loss": 0.0871,
|
| 812 |
+
"rewards/accuracies": 0.987500011920929,
|
| 813 |
+
"rewards/generated": -2.7490036487579346,
|
| 814 |
+
"rewards/margins": 9.90536880493164,
|
| 815 |
+
"rewards/real": 7.156365394592285,
|
| 816 |
+
"step": 400
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 1.2896,
|
| 820 |
+
"eval_logits/generated": -2.5240492820739746,
|
| 821 |
+
"eval_logits/real": -2.683285713195801,
|
| 822 |
+
"eval_logps/generated": -91.45059967041016,
|
| 823 |
+
"eval_logps/real": -116.5322265625,
|
| 824 |
+
"eval_loss": 0.8230924010276794,
|
| 825 |
+
"eval_rewards/accuracies": 0.5961538553237915,
|
| 826 |
+
"eval_rewards/generated": 1.7062723636627197,
|
| 827 |
+
"eval_rewards/margins": 0.28103044629096985,
|
| 828 |
+
"eval_rewards/real": 1.9873030185699463,
|
| 829 |
+
"eval_runtime": 25.4198,
|
| 830 |
+
"eval_samples_per_second": 7.868,
|
| 831 |
+
"eval_steps_per_second": 0.511,
|
| 832 |
+
"step": 403
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.312,
|
| 836 |
+
"grad_norm": 2.2179908898867304,
|
| 837 |
+
"learning_rate": 3.1235154394299283e-07,
|
| 838 |
+
"logits/generated": -2.590919017791748,
|
| 839 |
+
"logits/real": -2.6620683670043945,
|
| 840 |
+
"logps/generated": -154.9713897705078,
|
| 841 |
+
"logps/real": -64.6716537475586,
|
| 842 |
+
"loss": 0.1223,
|
| 843 |
+
"rewards/accuracies": 1.0,
|
| 844 |
+
"rewards/generated": -3.2485358715057373,
|
| 845 |
+
"rewards/margins": 11.025705337524414,
|
| 846 |
+
"rewards/real": 7.777169704437256,
|
| 847 |
+
"step": 410
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.3439999999999999,
|
| 851 |
+
"grad_norm": 2.1123570407415557,
|
| 852 |
+
"learning_rate": 3.0641330166270784e-07,
|
| 853 |
+
"logits/generated": -2.5320184230804443,
|
| 854 |
+
"logits/real": -2.605661153793335,
|
| 855 |
+
"logps/generated": -157.86102294921875,
|
| 856 |
+
"logps/real": -64.73320007324219,
|
| 857 |
+
"loss": 0.1132,
|
| 858 |
+
"rewards/accuracies": 0.987500011920929,
|
| 859 |
+
"rewards/generated": -3.422405242919922,
|
| 860 |
+
"rewards/margins": 10.930196762084961,
|
| 861 |
+
"rewards/real": 7.5077924728393555,
|
| 862 |
+
"step": 420
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.376,
|
| 866 |
+
"grad_norm": 52.0477538818839,
|
| 867 |
+
"learning_rate": 3.004750593824228e-07,
|
| 868 |
+
"logits/generated": -2.479769468307495,
|
| 869 |
+
"logits/real": -2.5996639728546143,
|
| 870 |
+
"logps/generated": -170.76138305664062,
|
| 871 |
+
"logps/real": -64.69879150390625,
|
| 872 |
+
"loss": 0.1454,
|
| 873 |
+
"rewards/accuracies": 1.0,
|
| 874 |
+
"rewards/generated": -4.175970554351807,
|
| 875 |
+
"rewards/margins": 11.573558807373047,
|
| 876 |
+
"rewards/real": 7.397587776184082,
|
| 877 |
+
"step": 430
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.3888,
|
| 881 |
+
"eval_logits/generated": -2.4324543476104736,
|
| 882 |
+
"eval_logits/real": -2.611969232559204,
|
| 883 |
+
"eval_logps/generated": -95.73091888427734,
|
| 884 |
+
"eval_logps/real": -119.04707336425781,
|
| 885 |
+
"eval_loss": 0.7979809641838074,
|
| 886 |
+
"eval_rewards/accuracies": 0.6730769276618958,
|
| 887 |
+
"eval_rewards/generated": 1.2782399654388428,
|
| 888 |
+
"eval_rewards/margins": 0.4575783908367157,
|
| 889 |
+
"eval_rewards/real": 1.7358182668685913,
|
| 890 |
+
"eval_runtime": 25.2977,
|
| 891 |
+
"eval_samples_per_second": 7.906,
|
| 892 |
+
"eval_steps_per_second": 0.514,
|
| 893 |
+
"step": 434
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 1.408,
|
| 897 |
+
"grad_norm": 3.7531043248353426,
|
| 898 |
+
"learning_rate": 2.9453681710213776e-07,
|
| 899 |
+
"logits/generated": -2.6189701557159424,
|
| 900 |
+
"logits/real": -2.6460072994232178,
|
| 901 |
+
"logps/generated": -178.8316192626953,
|
| 902 |
+
"logps/real": -66.7384033203125,
|
| 903 |
+
"loss": 0.07,
|
| 904 |
+
"rewards/accuracies": 1.0,
|
| 905 |
+
"rewards/generated": -3.621222734451294,
|
| 906 |
+
"rewards/margins": 11.845494270324707,
|
| 907 |
+
"rewards/real": 8.224271774291992,
|
| 908 |
+
"step": 440
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 1.44,
|
| 912 |
+
"grad_norm": 2.986230537501553,
|
| 913 |
+
"learning_rate": 2.885985748218527e-07,
|
| 914 |
+
"logits/generated": -2.5055761337280273,
|
| 915 |
+
"logits/real": -2.4730076789855957,
|
| 916 |
+
"logps/generated": -173.35125732421875,
|
| 917 |
+
"logps/real": -54.73099899291992,
|
| 918 |
+
"loss": 0.0934,
|
| 919 |
+
"rewards/accuracies": 1.0,
|
| 920 |
+
"rewards/generated": -4.109053611755371,
|
| 921 |
+
"rewards/margins": 11.3353271484375,
|
| 922 |
+
"rewards/real": 7.2262725830078125,
|
| 923 |
+
"step": 450
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 1.472,
|
| 927 |
+
"grad_norm": 2.618123223732552,
|
| 928 |
+
"learning_rate": 2.8266033254156767e-07,
|
| 929 |
+
"logits/generated": -2.4285778999328613,
|
| 930 |
+
"logits/real": -2.516470432281494,
|
| 931 |
+
"logps/generated": -151.6107177734375,
|
| 932 |
+
"logps/real": -55.497283935546875,
|
| 933 |
+
"loss": 0.0747,
|
| 934 |
+
"rewards/accuracies": 0.987500011920929,
|
| 935 |
+
"rewards/generated": -3.275545835494995,
|
| 936 |
+
"rewards/margins": 11.137293815612793,
|
| 937 |
+
"rewards/real": 7.861748695373535,
|
| 938 |
+
"step": 460
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 1.488,
|
| 942 |
+
"eval_logits/generated": -2.3556630611419678,
|
| 943 |
+
"eval_logits/real": -2.568284034729004,
|
| 944 |
+
"eval_logps/generated": -93.57498931884766,
|
| 945 |
+
"eval_logps/real": -117.37250518798828,
|
| 946 |
+
"eval_loss": 0.8086085319519043,
|
| 947 |
+
"eval_rewards/accuracies": 0.6538461446762085,
|
| 948 |
+
"eval_rewards/generated": 1.4938340187072754,
|
| 949 |
+
"eval_rewards/margins": 0.40944167971611023,
|
| 950 |
+
"eval_rewards/real": 1.9032756090164185,
|
| 951 |
+
"eval_runtime": 25.2837,
|
| 952 |
+
"eval_samples_per_second": 7.91,
|
| 953 |
+
"eval_steps_per_second": 0.514,
|
| 954 |
+
"step": 465
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.504,
|
| 958 |
+
"grad_norm": 2.3062948097679628,
|
| 959 |
+
"learning_rate": 2.7672209026128263e-07,
|
| 960 |
+
"logits/generated": -2.356971263885498,
|
| 961 |
+
"logits/real": -2.3984694480895996,
|
| 962 |
+
"logps/generated": -161.08523559570312,
|
| 963 |
+
"logps/real": -62.349212646484375,
|
| 964 |
+
"loss": 0.1099,
|
| 965 |
+
"rewards/accuracies": 0.987500011920929,
|
| 966 |
+
"rewards/generated": -3.898419141769409,
|
| 967 |
+
"rewards/margins": 12.482399940490723,
|
| 968 |
+
"rewards/real": 8.583980560302734,
|
| 969 |
+
"step": 470
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 1.536,
|
| 973 |
+
"grad_norm": 2.1831722984217468,
|
| 974 |
+
"learning_rate": 2.7078384798099764e-07,
|
| 975 |
+
"logits/generated": -2.434488296508789,
|
| 976 |
+
"logits/real": -2.4292871952056885,
|
| 977 |
+
"logps/generated": -165.63919067382812,
|
| 978 |
+
"logps/real": -53.660057067871094,
|
| 979 |
+
"loss": 0.0923,
|
| 980 |
+
"rewards/accuracies": 0.987500011920929,
|
| 981 |
+
"rewards/generated": -3.511655330657959,
|
| 982 |
+
"rewards/margins": 11.318243980407715,
|
| 983 |
+
"rewards/real": 7.806589603424072,
|
| 984 |
+
"step": 480
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.568,
|
| 988 |
+
"grad_norm": 1.9381609438311436,
|
| 989 |
+
"learning_rate": 2.648456057007126e-07,
|
| 990 |
+
"logits/generated": -2.395662784576416,
|
| 991 |
+
"logits/real": -2.446293354034424,
|
| 992 |
+
"logps/generated": -171.5785675048828,
|
| 993 |
+
"logps/real": -52.84147262573242,
|
| 994 |
+
"loss": 0.0882,
|
| 995 |
+
"rewards/accuracies": 1.0,
|
| 996 |
+
"rewards/generated": -3.7677719593048096,
|
| 997 |
+
"rewards/margins": 13.188079833984375,
|
| 998 |
+
"rewards/real": 9.420308113098145,
|
| 999 |
+
"step": 490
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 1.5872000000000002,
|
| 1003 |
+
"eval_logits/generated": -2.272164821624756,
|
| 1004 |
+
"eval_logits/real": -2.4783287048339844,
|
| 1005 |
+
"eval_logps/generated": -103.67977142333984,
|
| 1006 |
+
"eval_logps/real": -128.15371704101562,
|
| 1007 |
+
"eval_loss": 0.928070068359375,
|
| 1008 |
+
"eval_rewards/accuracies": 0.5192307829856873,
|
| 1009 |
+
"eval_rewards/generated": 0.483354777097702,
|
| 1010 |
+
"eval_rewards/margins": 0.34179702401161194,
|
| 1011 |
+
"eval_rewards/real": 0.8251518607139587,
|
| 1012 |
+
"eval_runtime": 25.574,
|
| 1013 |
+
"eval_samples_per_second": 7.82,
|
| 1014 |
+
"eval_steps_per_second": 0.508,
|
| 1015 |
+
"step": 496
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.6,
|
| 1019 |
+
"grad_norm": 2.2605824412655213,
|
| 1020 |
+
"learning_rate": 2.589073634204275e-07,
|
| 1021 |
+
"logits/generated": -2.3627734184265137,
|
| 1022 |
+
"logits/real": -2.4548180103302,
|
| 1023 |
+
"logps/generated": -169.9461212158203,
|
| 1024 |
+
"logps/real": -53.60149383544922,
|
| 1025 |
+
"loss": 0.0556,
|
| 1026 |
+
"rewards/accuracies": 1.0,
|
| 1027 |
+
"rewards/generated": -4.064536094665527,
|
| 1028 |
+
"rewards/margins": 12.720796585083008,
|
| 1029 |
+
"rewards/real": 8.656259536743164,
|
| 1030 |
+
"step": 500
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.6320000000000001,
|
| 1034 |
+
"grad_norm": 2.2813650068674387,
|
| 1035 |
+
"learning_rate": 2.529691211401425e-07,
|
| 1036 |
+
"logits/generated": -2.449333906173706,
|
| 1037 |
+
"logits/real": -2.473750352859497,
|
| 1038 |
+
"logps/generated": -169.88919067382812,
|
| 1039 |
+
"logps/real": -48.773372650146484,
|
| 1040 |
+
"loss": 0.0795,
|
| 1041 |
+
"rewards/accuracies": 1.0,
|
| 1042 |
+
"rewards/generated": -4.045998573303223,
|
| 1043 |
+
"rewards/margins": 13.330732345581055,
|
| 1044 |
+
"rewards/real": 9.284735679626465,
|
| 1045 |
+
"step": 510
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.6640000000000001,
|
| 1049 |
+
"grad_norm": 1.9756005158149215,
|
| 1050 |
+
"learning_rate": 2.4703087885985747e-07,
|
| 1051 |
+
"logits/generated": -2.2918307781219482,
|
| 1052 |
+
"logits/real": -2.4153807163238525,
|
| 1053 |
+
"logps/generated": -180.82943725585938,
|
| 1054 |
+
"logps/real": -64.54903411865234,
|
| 1055 |
+
"loss": 0.0693,
|
| 1056 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1057 |
+
"rewards/generated": -4.285412788391113,
|
| 1058 |
+
"rewards/margins": 13.515165328979492,
|
| 1059 |
+
"rewards/real": 9.229752540588379,
|
| 1060 |
+
"step": 520
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 1.6864,
|
| 1064 |
+
"eval_logits/generated": -2.139906644821167,
|
| 1065 |
+
"eval_logits/real": -2.368074893951416,
|
| 1066 |
+
"eval_logps/generated": -108.95226287841797,
|
| 1067 |
+
"eval_logps/real": -131.37367248535156,
|
| 1068 |
+
"eval_loss": 0.8953699469566345,
|
| 1069 |
+
"eval_rewards/accuracies": 0.6153846383094788,
|
| 1070 |
+
"eval_rewards/generated": -0.04389484226703644,
|
| 1071 |
+
"eval_rewards/margins": 0.5470523238182068,
|
| 1072 |
+
"eval_rewards/real": 0.5031575560569763,
|
| 1073 |
+
"eval_runtime": 25.3846,
|
| 1074 |
+
"eval_samples_per_second": 7.879,
|
| 1075 |
+
"eval_steps_per_second": 0.512,
|
| 1076 |
+
"step": 527
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 1.696,
|
| 1080 |
+
"grad_norm": 2.077906587896759,
|
| 1081 |
+
"learning_rate": 2.410926365795724e-07,
|
| 1082 |
+
"logits/generated": -2.4040651321411133,
|
| 1083 |
+
"logits/real": -2.4135825634002686,
|
| 1084 |
+
"logps/generated": -208.2981414794922,
|
| 1085 |
+
"logps/real": -51.23198699951172,
|
| 1086 |
+
"loss": 0.0774,
|
| 1087 |
+
"rewards/accuracies": 1.0,
|
| 1088 |
+
"rewards/generated": -5.633874416351318,
|
| 1089 |
+
"rewards/margins": 15.663395881652832,
|
| 1090 |
+
"rewards/real": 10.029520988464355,
|
| 1091 |
+
"step": 530
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 1.728,
|
| 1095 |
+
"grad_norm": 15.329093500904815,
|
| 1096 |
+
"learning_rate": 2.351543942992874e-07,
|
| 1097 |
+
"logits/generated": -2.3569560050964355,
|
| 1098 |
+
"logits/real": -2.3737831115722656,
|
| 1099 |
+
"logps/generated": -171.27365112304688,
|
| 1100 |
+
"logps/real": -41.608699798583984,
|
| 1101 |
+
"loss": 0.0459,
|
| 1102 |
+
"rewards/accuracies": 1.0,
|
| 1103 |
+
"rewards/generated": -4.523318767547607,
|
| 1104 |
+
"rewards/margins": 14.89915943145752,
|
| 1105 |
+
"rewards/real": 10.375839233398438,
|
| 1106 |
+
"step": 540
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 1.76,
|
| 1110 |
+
"grad_norm": 1.5653317010766996,
|
| 1111 |
+
"learning_rate": 2.2921615201900234e-07,
|
| 1112 |
+
"logits/generated": -2.2786147594451904,
|
| 1113 |
+
"logits/real": -2.380913496017456,
|
| 1114 |
+
"logps/generated": -178.86477661132812,
|
| 1115 |
+
"logps/real": -51.02559280395508,
|
| 1116 |
+
"loss": 0.0982,
|
| 1117 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1118 |
+
"rewards/generated": -4.9379401206970215,
|
| 1119 |
+
"rewards/margins": 14.276674270629883,
|
| 1120 |
+
"rewards/real": 9.338732719421387,
|
| 1121 |
+
"step": 550
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.7856,
|
| 1125 |
+
"eval_logits/generated": -2.332611560821533,
|
| 1126 |
+
"eval_logits/real": -2.5183045864105225,
|
| 1127 |
+
"eval_logps/generated": -103.10279846191406,
|
| 1128 |
+
"eval_logps/real": -126.28335571289062,
|
| 1129 |
+
"eval_loss": 0.8776556253433228,
|
| 1130 |
+
"eval_rewards/accuracies": 0.6538461446762085,
|
| 1131 |
+
"eval_rewards/generated": 0.5410510897636414,
|
| 1132 |
+
"eval_rewards/margins": 0.47113847732543945,
|
| 1133 |
+
"eval_rewards/real": 1.0121896266937256,
|
| 1134 |
+
"eval_runtime": 24.3313,
|
| 1135 |
+
"eval_samples_per_second": 8.22,
|
| 1136 |
+
"eval_steps_per_second": 0.534,
|
| 1137 |
+
"step": 558
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.792,
|
| 1141 |
+
"grad_norm": 23.493623835378564,
|
| 1142 |
+
"learning_rate": 2.2327790973871732e-07,
|
| 1143 |
+
"logits/generated": -2.4433507919311523,
|
| 1144 |
+
"logits/real": -2.4358441829681396,
|
| 1145 |
+
"logps/generated": -192.2283935546875,
|
| 1146 |
+
"logps/real": -47.60938262939453,
|
| 1147 |
+
"loss": 0.1072,
|
| 1148 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1149 |
+
"rewards/generated": -5.960423469543457,
|
| 1150 |
+
"rewards/margins": 14.969975471496582,
|
| 1151 |
+
"rewards/real": 9.009553909301758,
|
| 1152 |
+
"step": 560
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 1.8239999999999998,
|
| 1156 |
+
"grad_norm": 2.4384516635536024,
|
| 1157 |
+
"learning_rate": 2.173396674584323e-07,
|
| 1158 |
+
"logits/generated": -2.429056406021118,
|
| 1159 |
+
"logits/real": -2.447208881378174,
|
| 1160 |
+
"logps/generated": -178.07447814941406,
|
| 1161 |
+
"logps/real": -44.7133903503418,
|
| 1162 |
+
"loss": 0.0462,
|
| 1163 |
+
"rewards/accuracies": 1.0,
|
| 1164 |
+
"rewards/generated": -5.000073432922363,
|
| 1165 |
+
"rewards/margins": 14.771443367004395,
|
| 1166 |
+
"rewards/real": 9.771368980407715,
|
| 1167 |
+
"step": 570
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 1.8559999999999999,
|
| 1171 |
+
"grad_norm": 2.9085865122927945,
|
| 1172 |
+
"learning_rate": 2.1140142517814726e-07,
|
| 1173 |
+
"logits/generated": -2.404963254928589,
|
| 1174 |
+
"logits/real": -2.481902599334717,
|
| 1175 |
+
"logps/generated": -188.43467712402344,
|
| 1176 |
+
"logps/real": -44.5074462890625,
|
| 1177 |
+
"loss": 0.0674,
|
| 1178 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1179 |
+
"rewards/generated": -5.706393718719482,
|
| 1180 |
+
"rewards/margins": 16.334299087524414,
|
| 1181 |
+
"rewards/real": 10.627904891967773,
|
| 1182 |
+
"step": 580
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 1.8848,
|
| 1186 |
+
"eval_logits/generated": -2.302616834640503,
|
| 1187 |
+
"eval_logits/real": -2.4848058223724365,
|
| 1188 |
+
"eval_logps/generated": -113.82382202148438,
|
| 1189 |
+
"eval_logps/real": -136.99197387695312,
|
| 1190 |
+
"eval_loss": 0.9359613060951233,
|
| 1191 |
+
"eval_rewards/accuracies": 0.5961538553237915,
|
| 1192 |
+
"eval_rewards/generated": -0.5310505628585815,
|
| 1193 |
+
"eval_rewards/margins": 0.4723770022392273,
|
| 1194 |
+
"eval_rewards/real": -0.058673594146966934,
|
| 1195 |
+
"eval_runtime": 25.5072,
|
| 1196 |
+
"eval_samples_per_second": 7.841,
|
| 1197 |
+
"eval_steps_per_second": 0.51,
|
| 1198 |
+
"step": 589
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 1.888,
|
| 1202 |
+
"grad_norm": 2.224003882967406,
|
| 1203 |
+
"learning_rate": 2.0546318289786222e-07,
|
| 1204 |
+
"logits/generated": -2.4189112186431885,
|
| 1205 |
+
"logits/real": -2.4926321506500244,
|
| 1206 |
+
"logps/generated": -189.12107849121094,
|
| 1207 |
+
"logps/real": -34.886329650878906,
|
| 1208 |
+
"loss": 0.0404,
|
| 1209 |
+
"rewards/accuracies": 1.0,
|
| 1210 |
+
"rewards/generated": -6.180232048034668,
|
| 1211 |
+
"rewards/margins": 15.565508842468262,
|
| 1212 |
+
"rewards/real": 9.385276794433594,
|
| 1213 |
+
"step": 590
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.92,
|
| 1217 |
+
"grad_norm": 2.459225729639833,
|
| 1218 |
+
"learning_rate": 1.9952494061757718e-07,
|
| 1219 |
+
"logits/generated": -2.362020492553711,
|
| 1220 |
+
"logits/real": -2.3705716133117676,
|
| 1221 |
+
"logps/generated": -185.7289581298828,
|
| 1222 |
+
"logps/real": -29.566946029663086,
|
| 1223 |
+
"loss": 0.0835,
|
| 1224 |
+
"rewards/accuracies": 1.0,
|
| 1225 |
+
"rewards/generated": -6.159688472747803,
|
| 1226 |
+
"rewards/margins": 16.430912017822266,
|
| 1227 |
+
"rewards/real": 10.271220207214355,
|
| 1228 |
+
"step": 600
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.952,
|
| 1232 |
+
"grad_norm": 2.454360280136539,
|
| 1233 |
+
"learning_rate": 1.9358669833729216e-07,
|
| 1234 |
+
"logits/generated": -2.4703481197357178,
|
| 1235 |
+
"logits/real": -2.499268054962158,
|
| 1236 |
+
"logps/generated": -194.42752075195312,
|
| 1237 |
+
"logps/real": -35.10074996948242,
|
| 1238 |
+
"loss": 0.0584,
|
| 1239 |
+
"rewards/accuracies": 1.0,
|
| 1240 |
+
"rewards/generated": -6.010377407073975,
|
| 1241 |
+
"rewards/margins": 17.025707244873047,
|
| 1242 |
+
"rewards/real": 11.015329360961914,
|
| 1243 |
+
"step": 610
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.984,
|
| 1247 |
+
"grad_norm": 2.4169242550497096,
|
| 1248 |
+
"learning_rate": 1.876484560570071e-07,
|
| 1249 |
+
"logits/generated": -2.390986919403076,
|
| 1250 |
+
"logits/real": -2.453716993331909,
|
| 1251 |
+
"logps/generated": -187.9002227783203,
|
| 1252 |
+
"logps/real": -46.34991455078125,
|
| 1253 |
+
"loss": 0.0424,
|
| 1254 |
+
"rewards/accuracies": 1.0,
|
| 1255 |
+
"rewards/generated": -6.053334712982178,
|
| 1256 |
+
"rewards/margins": 17.14385986328125,
|
| 1257 |
+
"rewards/real": 11.090524673461914,
|
| 1258 |
+
"step": 620
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 1.984,
|
| 1262 |
+
"eval_logits/generated": -2.295470952987671,
|
| 1263 |
+
"eval_logits/real": -2.4845852851867676,
|
| 1264 |
+
"eval_logps/generated": -115.48155975341797,
|
| 1265 |
+
"eval_logps/real": -138.99147033691406,
|
| 1266 |
+
"eval_loss": 0.9420953989028931,
|
| 1267 |
+
"eval_rewards/accuracies": 0.5769230723381042,
|
| 1268 |
+
"eval_rewards/generated": -0.6968246102333069,
|
| 1269 |
+
"eval_rewards/margins": 0.4382031559944153,
|
| 1270 |
+
"eval_rewards/real": -0.2586213946342468,
|
| 1271 |
+
"eval_runtime": 25.416,
|
| 1272 |
+
"eval_samples_per_second": 7.869,
|
| 1273 |
+
"eval_steps_per_second": 0.511,
|
| 1274 |
+
"step": 620
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.016,
|
| 1278 |
+
"grad_norm": 13.261771791971471,
|
| 1279 |
+
"learning_rate": 1.8171021377672207e-07,
|
| 1280 |
+
"logits/generated": -2.4547436237335205,
|
| 1281 |
+
"logits/real": -2.3958754539489746,
|
| 1282 |
+
"logps/generated": -199.30731201171875,
|
| 1283 |
+
"logps/real": -31.5421142578125,
|
| 1284 |
+
"loss": 0.0561,
|
| 1285 |
+
"rewards/accuracies": 1.0,
|
| 1286 |
+
"rewards/generated": -6.895923614501953,
|
| 1287 |
+
"rewards/margins": 18.920942306518555,
|
| 1288 |
+
"rewards/real": 12.025017738342285,
|
| 1289 |
+
"step": 630
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 2.048,
|
| 1293 |
+
"grad_norm": 3.1173265935906165,
|
| 1294 |
+
"learning_rate": 1.7577197149643706e-07,
|
| 1295 |
+
"logits/generated": -2.331156015396118,
|
| 1296 |
+
"logits/real": -2.406820774078369,
|
| 1297 |
+
"logps/generated": -202.47897338867188,
|
| 1298 |
+
"logps/real": -29.300609588623047,
|
| 1299 |
+
"loss": 0.067,
|
| 1300 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1301 |
+
"rewards/generated": -6.904895782470703,
|
| 1302 |
+
"rewards/margins": 18.91607093811035,
|
| 1303 |
+
"rewards/real": 12.011175155639648,
|
| 1304 |
+
"step": 640
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.08,
|
| 1308 |
+
"grad_norm": 2.33318980437696,
|
| 1309 |
+
"learning_rate": 1.6983372921615202e-07,
|
| 1310 |
+
"logits/generated": -2.3845553398132324,
|
| 1311 |
+
"logits/real": -2.3674371242523193,
|
| 1312 |
+
"logps/generated": -214.82986450195312,
|
| 1313 |
+
"logps/real": -25.64461898803711,
|
| 1314 |
+
"loss": 0.0235,
|
| 1315 |
+
"rewards/accuracies": 1.0,
|
| 1316 |
+
"rewards/generated": -7.688128471374512,
|
| 1317 |
+
"rewards/margins": 19.164615631103516,
|
| 1318 |
+
"rewards/real": 11.476486206054688,
|
| 1319 |
+
"step": 650
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 2.0832,
|
| 1323 |
+
"eval_logits/generated": -2.2115392684936523,
|
| 1324 |
+
"eval_logits/real": -2.3974390029907227,
|
| 1325 |
+
"eval_logps/generated": -128.70654296875,
|
| 1326 |
+
"eval_logps/real": -153.1709442138672,
|
| 1327 |
+
"eval_loss": 1.0938998460769653,
|
| 1328 |
+
"eval_rewards/accuracies": 0.5,
|
| 1329 |
+
"eval_rewards/generated": -2.019321918487549,
|
| 1330 |
+
"eval_rewards/margins": 0.3427533805370331,
|
| 1331 |
+
"eval_rewards/real": -1.6765685081481934,
|
| 1332 |
+
"eval_runtime": 24.3647,
|
| 1333 |
+
"eval_samples_per_second": 8.209,
|
| 1334 |
+
"eval_steps_per_second": 0.534,
|
| 1335 |
+
"step": 651
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 2.112,
|
| 1339 |
+
"grad_norm": 2.851649001912384,
|
| 1340 |
+
"learning_rate": 1.6389548693586697e-07,
|
| 1341 |
+
"logits/generated": -2.297102689743042,
|
| 1342 |
+
"logits/real": -2.319133758544922,
|
| 1343 |
+
"logps/generated": -201.3473663330078,
|
| 1344 |
+
"logps/real": -25.802215576171875,
|
| 1345 |
+
"loss": 0.0541,
|
| 1346 |
+
"rewards/accuracies": 1.0,
|
| 1347 |
+
"rewards/generated": -7.507266044616699,
|
| 1348 |
+
"rewards/margins": 18.991275787353516,
|
| 1349 |
+
"rewards/real": 11.484009742736816,
|
| 1350 |
+
"step": 660
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.144,
|
| 1354 |
+
"grad_norm": 2.090545045938182,
|
| 1355 |
+
"learning_rate": 1.5795724465558193e-07,
|
| 1356 |
+
"logits/generated": -2.4693615436553955,
|
| 1357 |
+
"logits/real": -2.4613473415374756,
|
| 1358 |
+
"logps/generated": -200.6858367919922,
|
| 1359 |
+
"logps/real": -39.16984558105469,
|
| 1360 |
+
"loss": 0.0572,
|
| 1361 |
+
"rewards/accuracies": 1.0,
|
| 1362 |
+
"rewards/generated": -7.124711513519287,
|
| 1363 |
+
"rewards/margins": 20.449026107788086,
|
| 1364 |
+
"rewards/real": 13.324316024780273,
|
| 1365 |
+
"step": 670
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 2.176,
|
| 1369 |
+
"grad_norm": 2.1939273942211277,
|
| 1370 |
+
"learning_rate": 1.520190023752969e-07,
|
| 1371 |
+
"logits/generated": -2.344393253326416,
|
| 1372 |
+
"logits/real": -2.3299365043640137,
|
| 1373 |
+
"logps/generated": -212.4817352294922,
|
| 1374 |
+
"logps/real": -20.405345916748047,
|
| 1375 |
+
"loss": 0.024,
|
| 1376 |
+
"rewards/accuracies": 1.0,
|
| 1377 |
+
"rewards/generated": -7.623618125915527,
|
| 1378 |
+
"rewards/margins": 18.887493133544922,
|
| 1379 |
+
"rewards/real": 11.263875007629395,
|
| 1380 |
+
"step": 680
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 2.1824,
|
| 1384 |
+
"eval_logits/generated": -2.204942226409912,
|
| 1385 |
+
"eval_logits/real": -2.393613338470459,
|
| 1386 |
+
"eval_logps/generated": -133.9093017578125,
|
| 1387 |
+
"eval_logps/real": -157.97007751464844,
|
| 1388 |
+
"eval_loss": 1.1490856409072876,
|
| 1389 |
+
"eval_rewards/accuracies": 0.5,
|
| 1390 |
+
"eval_rewards/generated": -2.539597749710083,
|
| 1391 |
+
"eval_rewards/margins": 0.3831157386302948,
|
| 1392 |
+
"eval_rewards/real": -2.156482219696045,
|
| 1393 |
+
"eval_runtime": 25.5214,
|
| 1394 |
+
"eval_samples_per_second": 7.837,
|
| 1395 |
+
"eval_steps_per_second": 0.509,
|
| 1396 |
+
"step": 682
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 2.208,
|
| 1400 |
+
"grad_norm": 2.6699514134692506,
|
| 1401 |
+
"learning_rate": 1.4608076009501184e-07,
|
| 1402 |
+
"logits/generated": -2.3223159313201904,
|
| 1403 |
+
"logits/real": -2.267909526824951,
|
| 1404 |
+
"logps/generated": -209.5951690673828,
|
| 1405 |
+
"logps/real": -22.11200714111328,
|
| 1406 |
+
"loss": 0.0273,
|
| 1407 |
+
"rewards/accuracies": 1.0,
|
| 1408 |
+
"rewards/generated": -7.6561784744262695,
|
| 1409 |
+
"rewards/margins": 18.957958221435547,
|
| 1410 |
+
"rewards/real": 11.301776885986328,
|
| 1411 |
+
"step": 690
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 2.24,
|
| 1415 |
+
"grad_norm": 2.245581599718528,
|
| 1416 |
+
"learning_rate": 1.4014251781472683e-07,
|
| 1417 |
+
"logits/generated": -2.293184757232666,
|
| 1418 |
+
"logits/real": -2.2608914375305176,
|
| 1419 |
+
"logps/generated": -195.58767700195312,
|
| 1420 |
+
"logps/real": -30.79043197631836,
|
| 1421 |
+
"loss": 0.0365,
|
| 1422 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1423 |
+
"rewards/generated": -6.686369895935059,
|
| 1424 |
+
"rewards/margins": 18.34311294555664,
|
| 1425 |
+
"rewards/real": 11.656742095947266,
|
| 1426 |
+
"step": 700
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 2.2720000000000002,
|
| 1430 |
+
"grad_norm": 3.154191159829218,
|
| 1431 |
+
"learning_rate": 1.342042755344418e-07,
|
| 1432 |
+
"logits/generated": -2.2829322814941406,
|
| 1433 |
+
"logits/real": -2.370777130126953,
|
| 1434 |
+
"logps/generated": -205.3447723388672,
|
| 1435 |
+
"logps/real": -21.945812225341797,
|
| 1436 |
+
"loss": 0.0469,
|
| 1437 |
+
"rewards/accuracies": 1.0,
|
| 1438 |
+
"rewards/generated": -7.478776454925537,
|
| 1439 |
+
"rewards/margins": 19.887378692626953,
|
| 1440 |
+
"rewards/real": 12.408601760864258,
|
| 1441 |
+
"step": 710
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.2816,
|
| 1445 |
+
"eval_logits/generated": -2.2161128520965576,
|
| 1446 |
+
"eval_logits/real": -2.409364938735962,
|
| 1447 |
+
"eval_logps/generated": -133.31402587890625,
|
| 1448 |
+
"eval_logps/real": -157.023193359375,
|
| 1449 |
+
"eval_loss": 1.1324479579925537,
|
| 1450 |
+
"eval_rewards/accuracies": 0.5,
|
| 1451 |
+
"eval_rewards/generated": -2.480071544647217,
|
| 1452 |
+
"eval_rewards/margins": 0.41827720403671265,
|
| 1453 |
+
"eval_rewards/real": -2.0617942810058594,
|
| 1454 |
+
"eval_runtime": 25.3651,
|
| 1455 |
+
"eval_samples_per_second": 7.885,
|
| 1456 |
+
"eval_steps_per_second": 0.513,
|
| 1457 |
+
"step": 713
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 2.304,
|
| 1461 |
+
"grad_norm": 1.70496744806456,
|
| 1462 |
+
"learning_rate": 1.2826603325415677e-07,
|
| 1463 |
+
"logits/generated": -2.3763372898101807,
|
| 1464 |
+
"logits/real": -2.365422248840332,
|
| 1465 |
+
"logps/generated": -207.6084442138672,
|
| 1466 |
+
"logps/real": -18.752004623413086,
|
| 1467 |
+
"loss": 0.0623,
|
| 1468 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1469 |
+
"rewards/generated": -7.463881015777588,
|
| 1470 |
+
"rewards/margins": 19.985795974731445,
|
| 1471 |
+
"rewards/real": 12.5219144821167,
|
| 1472 |
+
"step": 720
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 2.336,
|
| 1476 |
+
"grad_norm": 1.7034616606078241,
|
| 1477 |
+
"learning_rate": 1.2232779097387173e-07,
|
| 1478 |
+
"logits/generated": -2.3761379718780518,
|
| 1479 |
+
"logits/real": -2.344998836517334,
|
| 1480 |
+
"logps/generated": -213.1655731201172,
|
| 1481 |
+
"logps/real": -18.98818016052246,
|
| 1482 |
+
"loss": 0.0191,
|
| 1483 |
+
"rewards/accuracies": 1.0,
|
| 1484 |
+
"rewards/generated": -8.140850067138672,
|
| 1485 |
+
"rewards/margins": 20.51372528076172,
|
| 1486 |
+
"rewards/real": 12.37287425994873,
|
| 1487 |
+
"step": 730
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 2.368,
|
| 1491 |
+
"grad_norm": 2.470062940286456,
|
| 1492 |
+
"learning_rate": 1.163895486935867e-07,
|
| 1493 |
+
"logits/generated": -2.342379093170166,
|
| 1494 |
+
"logits/real": -2.342834949493408,
|
| 1495 |
+
"logps/generated": -207.5922393798828,
|
| 1496 |
+
"logps/real": -26.875713348388672,
|
| 1497 |
+
"loss": 0.0328,
|
| 1498 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1499 |
+
"rewards/generated": -7.883567810058594,
|
| 1500 |
+
"rewards/margins": 20.540231704711914,
|
| 1501 |
+
"rewards/real": 12.656661987304688,
|
| 1502 |
+
"step": 740
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 2.3808,
|
| 1506 |
+
"eval_logits/generated": -2.207951068878174,
|
| 1507 |
+
"eval_logits/real": -2.395209550857544,
|
| 1508 |
+
"eval_logps/generated": -136.21505737304688,
|
| 1509 |
+
"eval_logps/real": -160.9390411376953,
|
| 1510 |
+
"eval_loss": 1.183669090270996,
|
| 1511 |
+
"eval_rewards/accuracies": 0.48076921701431274,
|
| 1512 |
+
"eval_rewards/generated": -2.770172119140625,
|
| 1513 |
+
"eval_rewards/margins": 0.316792368888855,
|
| 1514 |
+
"eval_rewards/real": -2.4533796310424805,
|
| 1515 |
+
"eval_runtime": 24.3566,
|
| 1516 |
+
"eval_samples_per_second": 8.211,
|
| 1517 |
+
"eval_steps_per_second": 0.534,
|
| 1518 |
+
"step": 744
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 2.4,
|
| 1522 |
+
"grad_norm": 2.0406895368605653,
|
| 1523 |
+
"learning_rate": 1.1045130641330165e-07,
|
| 1524 |
+
"logits/generated": -2.338466167449951,
|
| 1525 |
+
"logits/real": -2.2994284629821777,
|
| 1526 |
+
"logps/generated": -220.8417510986328,
|
| 1527 |
+
"logps/real": -22.894880294799805,
|
| 1528 |
+
"loss": 0.0269,
|
| 1529 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1530 |
+
"rewards/generated": -8.605572700500488,
|
| 1531 |
+
"rewards/margins": 20.708206176757812,
|
| 1532 |
+
"rewards/real": 12.102633476257324,
|
| 1533 |
+
"step": 750
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 2.432,
|
| 1537 |
+
"grad_norm": 22.669827351080848,
|
| 1538 |
+
"learning_rate": 1.0451306413301662e-07,
|
| 1539 |
+
"logits/generated": -2.42244291305542,
|
| 1540 |
+
"logits/real": -2.315642833709717,
|
| 1541 |
+
"logps/generated": -221.71664428710938,
|
| 1542 |
+
"logps/real": -19.839298248291016,
|
| 1543 |
+
"loss": 0.0358,
|
| 1544 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1545 |
+
"rewards/generated": -8.81576919555664,
|
| 1546 |
+
"rewards/margins": 21.98428726196289,
|
| 1547 |
+
"rewards/real": 13.168516159057617,
|
| 1548 |
+
"step": 760
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 2.464,
|
| 1552 |
+
"grad_norm": 2.524402783352166,
|
| 1553 |
+
"learning_rate": 9.857482185273158e-08,
|
| 1554 |
+
"logits/generated": -2.350756883621216,
|
| 1555 |
+
"logits/real": -2.3330166339874268,
|
| 1556 |
+
"logps/generated": -215.21243286132812,
|
| 1557 |
+
"logps/real": -17.460187911987305,
|
| 1558 |
+
"loss": 0.0367,
|
| 1559 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1560 |
+
"rewards/generated": -7.874646186828613,
|
| 1561 |
+
"rewards/margins": 21.986469268798828,
|
| 1562 |
+
"rewards/real": 14.111824035644531,
|
| 1563 |
+
"step": 770
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 2.48,
|
| 1567 |
+
"eval_logits/generated": -2.1815290451049805,
|
| 1568 |
+
"eval_logits/real": -2.3776841163635254,
|
| 1569 |
+
"eval_logps/generated": -138.23756408691406,
|
| 1570 |
+
"eval_logps/real": -162.54420471191406,
|
| 1571 |
+
"eval_loss": 1.1778903007507324,
|
| 1572 |
+
"eval_rewards/accuracies": 0.48076921701431274,
|
| 1573 |
+
"eval_rewards/generated": -2.972423791885376,
|
| 1574 |
+
"eval_rewards/margins": 0.35853010416030884,
|
| 1575 |
+
"eval_rewards/real": -2.613893747329712,
|
| 1576 |
+
"eval_runtime": 25.6643,
|
| 1577 |
+
"eval_samples_per_second": 7.793,
|
| 1578 |
+
"eval_steps_per_second": 0.507,
|
| 1579 |
+
"step": 775
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 2.496,
|
| 1583 |
+
"grad_norm": 2.40420451796189,
|
| 1584 |
+
"learning_rate": 9.263657957244655e-08,
|
| 1585 |
+
"logits/generated": -2.2569501399993896,
|
| 1586 |
+
"logits/real": -2.264871120452881,
|
| 1587 |
+
"logps/generated": -202.81716918945312,
|
| 1588 |
+
"logps/real": -24.641332626342773,
|
| 1589 |
+
"loss": 0.0486,
|
| 1590 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1591 |
+
"rewards/generated": -7.750497341156006,
|
| 1592 |
+
"rewards/margins": 20.739505767822266,
|
| 1593 |
+
"rewards/real": 12.989007949829102,
|
| 1594 |
+
"step": 780
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 2.528,
|
| 1598 |
+
"grad_norm": 22.160283840192612,
|
| 1599 |
+
"learning_rate": 8.669833729216151e-08,
|
| 1600 |
+
"logits/generated": -2.3489882946014404,
|
| 1601 |
+
"logits/real": -2.3021740913391113,
|
| 1602 |
+
"logps/generated": -242.10452270507812,
|
| 1603 |
+
"logps/real": -19.257644653320312,
|
| 1604 |
+
"loss": 0.0673,
|
| 1605 |
+
"rewards/accuracies": 1.0,
|
| 1606 |
+
"rewards/generated": -9.988310813903809,
|
| 1607 |
+
"rewards/margins": 23.60817527770996,
|
| 1608 |
+
"rewards/real": 13.619865417480469,
|
| 1609 |
+
"step": 790
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 2.56,
|
| 1613 |
+
"grad_norm": 2.113351859691927,
|
| 1614 |
+
"learning_rate": 8.076009501187649e-08,
|
| 1615 |
+
"logits/generated": -2.2851662635803223,
|
| 1616 |
+
"logits/real": -2.278385639190674,
|
| 1617 |
+
"logps/generated": -215.65634155273438,
|
| 1618 |
+
"logps/real": -17.226621627807617,
|
| 1619 |
+
"loss": 0.0596,
|
| 1620 |
+
"rewards/accuracies": 1.0,
|
| 1621 |
+
"rewards/generated": -9.116616249084473,
|
| 1622 |
+
"rewards/margins": 22.34113311767578,
|
| 1623 |
+
"rewards/real": 13.224515914916992,
|
| 1624 |
+
"step": 800
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 2.5792,
|
| 1628 |
+
"eval_logits/generated": -2.1523475646972656,
|
| 1629 |
+
"eval_logits/real": -2.3457515239715576,
|
| 1630 |
+
"eval_logps/generated": -144.71934509277344,
|
| 1631 |
+
"eval_logps/real": -169.895263671875,
|
| 1632 |
+
"eval_loss": 1.284745454788208,
|
| 1633 |
+
"eval_rewards/accuracies": 0.42307692766189575,
|
| 1634 |
+
"eval_rewards/generated": -3.6206037998199463,
|
| 1635 |
+
"eval_rewards/margins": 0.2716017961502075,
|
| 1636 |
+
"eval_rewards/real": -3.3490025997161865,
|
| 1637 |
+
"eval_runtime": 25.4228,
|
| 1638 |
+
"eval_samples_per_second": 7.867,
|
| 1639 |
+
"eval_steps_per_second": 0.511,
|
| 1640 |
+
"step": 806
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 2.592,
|
| 1644 |
+
"grad_norm": 11.216326120395912,
|
| 1645 |
+
"learning_rate": 7.482185273159145e-08,
|
| 1646 |
+
"logits/generated": -2.292395830154419,
|
| 1647 |
+
"logits/real": -2.2245354652404785,
|
| 1648 |
+
"logps/generated": -216.6605682373047,
|
| 1649 |
+
"logps/real": -17.53030776977539,
|
| 1650 |
+
"loss": 0.0354,
|
| 1651 |
+
"rewards/accuracies": 1.0,
|
| 1652 |
+
"rewards/generated": -9.128471374511719,
|
| 1653 |
+
"rewards/margins": 21.153453826904297,
|
| 1654 |
+
"rewards/real": 12.024984359741211,
|
| 1655 |
+
"step": 810
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 2.624,
|
| 1659 |
+
"grad_norm": 2.0837019278053304,
|
| 1660 |
+
"learning_rate": 6.88836104513064e-08,
|
| 1661 |
+
"logits/generated": -2.2674856185913086,
|
| 1662 |
+
"logits/real": -2.246483325958252,
|
| 1663 |
+
"logps/generated": -235.4110107421875,
|
| 1664 |
+
"logps/real": -24.103424072265625,
|
| 1665 |
+
"loss": 0.035,
|
| 1666 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1667 |
+
"rewards/generated": -9.786630630493164,
|
| 1668 |
+
"rewards/margins": 22.674787521362305,
|
| 1669 |
+
"rewards/real": 12.888157844543457,
|
| 1670 |
+
"step": 820
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 2.656,
|
| 1674 |
+
"grad_norm": 3.9512882624496712,
|
| 1675 |
+
"learning_rate": 6.294536817102138e-08,
|
| 1676 |
+
"logits/generated": -2.252331256866455,
|
| 1677 |
+
"logits/real": -2.2256457805633545,
|
| 1678 |
+
"logps/generated": -202.04873657226562,
|
| 1679 |
+
"logps/real": -9.093053817749023,
|
| 1680 |
+
"loss": 0.0395,
|
| 1681 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1682 |
+
"rewards/generated": -8.500150680541992,
|
| 1683 |
+
"rewards/margins": 19.9840087890625,
|
| 1684 |
+
"rewards/real": 11.483858108520508,
|
| 1685 |
+
"step": 830
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 2.6784,
|
| 1689 |
+
"eval_logits/generated": -2.139922618865967,
|
| 1690 |
+
"eval_logits/real": -2.3346006870269775,
|
| 1691 |
+
"eval_logps/generated": -147.52369689941406,
|
| 1692 |
+
"eval_logps/real": -172.99374389648438,
|
| 1693 |
+
"eval_loss": 1.3357782363891602,
|
| 1694 |
+
"eval_rewards/accuracies": 0.42307692766189575,
|
| 1695 |
+
"eval_rewards/generated": -3.9010369777679443,
|
| 1696 |
+
"eval_rewards/margins": 0.24218764901161194,
|
| 1697 |
+
"eval_rewards/real": -3.6588494777679443,
|
| 1698 |
+
"eval_runtime": 25.3471,
|
| 1699 |
+
"eval_samples_per_second": 7.89,
|
| 1700 |
+
"eval_steps_per_second": 0.513,
|
| 1701 |
+
"step": 837
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 2.6879999999999997,
|
| 1705 |
+
"grad_norm": 22.1387966335112,
|
| 1706 |
+
"learning_rate": 5.700712589073634e-08,
|
| 1707 |
+
"logits/generated": -2.3124189376831055,
|
| 1708 |
+
"logits/real": -2.2206485271453857,
|
| 1709 |
+
"logps/generated": -229.35366821289062,
|
| 1710 |
+
"logps/real": -10.505355834960938,
|
| 1711 |
+
"loss": 0.0201,
|
| 1712 |
+
"rewards/accuracies": 1.0,
|
| 1713 |
+
"rewards/generated": -10.045392036437988,
|
| 1714 |
+
"rewards/margins": 22.45701026916504,
|
| 1715 |
+
"rewards/real": 12.41161823272705,
|
| 1716 |
+
"step": 840
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 2.7199999999999998,
|
| 1720 |
+
"grad_norm": 1.6992001757462196,
|
| 1721 |
+
"learning_rate": 5.10688836104513e-08,
|
| 1722 |
+
"logits/generated": -2.196502447128296,
|
| 1723 |
+
"logits/real": -2.26035737991333,
|
| 1724 |
+
"logps/generated": -206.0448760986328,
|
| 1725 |
+
"logps/real": -13.079363822937012,
|
| 1726 |
+
"loss": 0.0262,
|
| 1727 |
+
"rewards/accuracies": 1.0,
|
| 1728 |
+
"rewards/generated": -8.79426383972168,
|
| 1729 |
+
"rewards/margins": 21.76181411743164,
|
| 1730 |
+
"rewards/real": 12.967552185058594,
|
| 1731 |
+
"step": 850
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 2.752,
|
| 1735 |
+
"grad_norm": 2.6474508823777674,
|
| 1736 |
+
"learning_rate": 4.5130641330166267e-08,
|
| 1737 |
+
"logits/generated": -2.3978683948516846,
|
| 1738 |
+
"logits/real": -2.248363494873047,
|
| 1739 |
+
"logps/generated": -228.8793487548828,
|
| 1740 |
+
"logps/real": -19.478151321411133,
|
| 1741 |
+
"loss": 0.0302,
|
| 1742 |
+
"rewards/accuracies": 1.0,
|
| 1743 |
+
"rewards/generated": -10.040186882019043,
|
| 1744 |
+
"rewards/margins": 23.719615936279297,
|
| 1745 |
+
"rewards/real": 13.679426193237305,
|
| 1746 |
+
"step": 860
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 2.7776,
|
| 1750 |
+
"eval_logits/generated": -2.152860403060913,
|
| 1751 |
+
"eval_logits/real": -2.347489833831787,
|
| 1752 |
+
"eval_logps/generated": -148.89895629882812,
|
| 1753 |
+
"eval_logps/real": -174.31674194335938,
|
| 1754 |
+
"eval_loss": 1.3724898099899292,
|
| 1755 |
+
"eval_rewards/accuracies": 0.42307692766189575,
|
| 1756 |
+
"eval_rewards/generated": -4.038564205169678,
|
| 1757 |
+
"eval_rewards/margins": 0.24741561710834503,
|
| 1758 |
+
"eval_rewards/real": -3.7911486625671387,
|
| 1759 |
+
"eval_runtime": 24.3356,
|
| 1760 |
+
"eval_samples_per_second": 8.218,
|
| 1761 |
+
"eval_steps_per_second": 0.534,
|
| 1762 |
+
"step": 868
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 2.784,
|
| 1766 |
+
"grad_norm": 1.7441479405692193,
|
| 1767 |
+
"learning_rate": 3.919239904988123e-08,
|
| 1768 |
+
"logits/generated": -2.349801778793335,
|
| 1769 |
+
"logits/real": -2.2702341079711914,
|
| 1770 |
+
"logps/generated": -234.3719940185547,
|
| 1771 |
+
"logps/real": -14.089096069335938,
|
| 1772 |
+
"loss": 0.0369,
|
| 1773 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1774 |
+
"rewards/generated": -10.196078300476074,
|
| 1775 |
+
"rewards/margins": 22.43720245361328,
|
| 1776 |
+
"rewards/real": 12.241125106811523,
|
| 1777 |
+
"step": 870
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 2.816,
|
| 1781 |
+
"grad_norm": 2.2386762278529777,
|
| 1782 |
+
"learning_rate": 3.32541567695962e-08,
|
| 1783 |
+
"logits/generated": -2.2697014808654785,
|
| 1784 |
+
"logits/real": -2.2320122718811035,
|
| 1785 |
+
"logps/generated": -235.64669799804688,
|
| 1786 |
+
"logps/real": -9.42314338684082,
|
| 1787 |
+
"loss": 0.0392,
|
| 1788 |
+
"rewards/accuracies": 1.0,
|
| 1789 |
+
"rewards/generated": -10.039624214172363,
|
| 1790 |
+
"rewards/margins": 22.644527435302734,
|
| 1791 |
+
"rewards/real": 12.604903221130371,
|
| 1792 |
+
"step": 880
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 2.848,
|
| 1796 |
+
"grad_norm": 1.5468486075007237,
|
| 1797 |
+
"learning_rate": 2.7315914489311164e-08,
|
| 1798 |
+
"logits/generated": -2.2128958702087402,
|
| 1799 |
+
"logits/real": -2.260349988937378,
|
| 1800 |
+
"logps/generated": -205.9203643798828,
|
| 1801 |
+
"logps/real": -14.784475326538086,
|
| 1802 |
+
"loss": 0.0132,
|
| 1803 |
+
"rewards/accuracies": 1.0,
|
| 1804 |
+
"rewards/generated": -9.317214012145996,
|
| 1805 |
+
"rewards/margins": 22.827068328857422,
|
| 1806 |
+
"rewards/real": 13.509855270385742,
|
| 1807 |
+
"step": 890
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 2.8768000000000002,
|
| 1811 |
+
"eval_logits/generated": -2.1226818561553955,
|
| 1812 |
+
"eval_logits/real": -2.317802667617798,
|
| 1813 |
+
"eval_logps/generated": -154.74986267089844,
|
| 1814 |
+
"eval_logps/real": -181.03439331054688,
|
| 1815 |
+
"eval_loss": 1.4968535900115967,
|
| 1816 |
+
"eval_rewards/accuracies": 0.4423076808452606,
|
| 1817 |
+
"eval_rewards/generated": -4.623655319213867,
|
| 1818 |
+
"eval_rewards/margins": 0.16074196994304657,
|
| 1819 |
+
"eval_rewards/real": -4.462913513183594,
|
| 1820 |
+
"eval_runtime": 25.3278,
|
| 1821 |
+
"eval_samples_per_second": 7.896,
|
| 1822 |
+
"eval_steps_per_second": 0.513,
|
| 1823 |
+
"step": 899
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.88,
|
| 1827 |
+
"grad_norm": 5.589462535052379,
|
| 1828 |
+
"learning_rate": 2.1377672209026125e-08,
|
| 1829 |
+
"logits/generated": -2.1872811317443848,
|
| 1830 |
+
"logits/real": -2.190425395965576,
|
| 1831 |
+
"logps/generated": -222.4118194580078,
|
| 1832 |
+
"logps/real": -12.531976699829102,
|
| 1833 |
+
"loss": 0.0144,
|
| 1834 |
+
"rewards/accuracies": 1.0,
|
| 1835 |
+
"rewards/generated": -10.07731819152832,
|
| 1836 |
+
"rewards/margins": 22.0319881439209,
|
| 1837 |
+
"rewards/real": 11.954668045043945,
|
| 1838 |
+
"step": 900
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 2.912,
|
| 1842 |
+
"grad_norm": 1.1002291429170004,
|
| 1843 |
+
"learning_rate": 1.5439429928741092e-08,
|
| 1844 |
+
"logits/generated": -2.267789363861084,
|
| 1845 |
+
"logits/real": -2.174405574798584,
|
| 1846 |
+
"logps/generated": -258.24346923828125,
|
| 1847 |
+
"logps/real": -17.154361724853516,
|
| 1848 |
+
"loss": 0.0169,
|
| 1849 |
+
"rewards/accuracies": 1.0,
|
| 1850 |
+
"rewards/generated": -11.40230655670166,
|
| 1851 |
+
"rewards/margins": 25.64777183532715,
|
| 1852 |
+
"rewards/real": 14.245463371276855,
|
| 1853 |
+
"step": 910
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 2.944,
|
| 1857 |
+
"grad_norm": 2.533629319876756,
|
| 1858 |
+
"learning_rate": 9.501187648456057e-09,
|
| 1859 |
+
"logits/generated": -2.2727675437927246,
|
| 1860 |
+
"logits/real": -2.2528767585754395,
|
| 1861 |
+
"logps/generated": -228.3806915283203,
|
| 1862 |
+
"logps/real": -12.532684326171875,
|
| 1863 |
+
"loss": 0.029,
|
| 1864 |
+
"rewards/accuracies": 1.0,
|
| 1865 |
+
"rewards/generated": -10.578569412231445,
|
| 1866 |
+
"rewards/margins": 23.38239097595215,
|
| 1867 |
+
"rewards/real": 12.80382251739502,
|
| 1868 |
+
"step": 920
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 2.976,
|
| 1872 |
+
"grad_norm": 22.025786598083638,
|
| 1873 |
+
"learning_rate": 3.562945368171021e-09,
|
| 1874 |
+
"logits/generated": -2.2085366249084473,
|
| 1875 |
+
"logits/real": -2.227689743041992,
|
| 1876 |
+
"logps/generated": -219.00088500976562,
|
| 1877 |
+
"logps/real": -16.957204818725586,
|
| 1878 |
+
"loss": 0.034,
|
| 1879 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1880 |
+
"rewards/generated": -9.949075698852539,
|
| 1881 |
+
"rewards/margins": 22.433155059814453,
|
| 1882 |
+
"rewards/real": 12.484078407287598,
|
| 1883 |
+
"step": 930
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 2.976,
|
| 1887 |
+
"eval_logits/generated": -2.120302200317383,
|
| 1888 |
+
"eval_logits/real": -2.3164217472076416,
|
| 1889 |
+
"eval_logps/generated": -155.21237182617188,
|
| 1890 |
+
"eval_logps/real": -181.53463745117188,
|
| 1891 |
+
"eval_loss": 1.4922109842300415,
|
| 1892 |
+
"eval_rewards/accuracies": 0.4423076808452606,
|
| 1893 |
+
"eval_rewards/generated": -4.669904708862305,
|
| 1894 |
+
"eval_rewards/margins": 0.15696734189987183,
|
| 1895 |
+
"eval_rewards/real": -4.512937545776367,
|
| 1896 |
+
"eval_runtime": 25.074,
|
| 1897 |
+
"eval_samples_per_second": 7.976,
|
| 1898 |
+
"eval_steps_per_second": 0.518,
|
| 1899 |
+
"step": 930
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 2.9952,
|
| 1903 |
+
"step": 936,
|
| 1904 |
"total_flos": 0.0,
|
| 1905 |
+
"train_loss": 0.16162516827830392,
|
| 1906 |
+
"train_runtime": 16659.2082,
|
| 1907 |
+
"train_samples_per_second": 1.8,
|
| 1908 |
+
"train_steps_per_second": 0.056
|
| 1909 |
}
|
| 1910 |
],
|
| 1911 |
"logging_steps": 10,
|
| 1912 |
+
"max_steps": 936,
|
| 1913 |
"num_input_tokens_seen": 0,
|
| 1914 |
+
"num_train_epochs": 3,
|
| 1915 |
+
"save_steps": 31,
|
| 1916 |
"stateful_callbacks": {
|
| 1917 |
"TrainerControl": {
|
| 1918 |
"args": {
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6392
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36e25769f4bb9d40c6d0fe2d6f74f4924eae565684061dc7895da21ea737b025
|
| 3 |
size 6392
|