Commit
•
249e3b5
1
Parent(s):
2190ad7
Model save
Browse files- README.md +87 -0
- adapter_model.safetensors +1 -1
- all_results.json +21 -0
- eval_results.json +16 -0
- runs/Apr24_18-06-13_gpu4-119-5/events.out.tfevents.1713946169.gpu4-119-5.4164190.0 +2 -2
- runs/Apr24_18-06-13_gpu4-119-5/events.out.tfevents.1713968856.gpu4-119-5.4164190.1 +3 -0
- train_results.json +8 -0
- trainer_state.json +2468 -0
README.md
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: microsoft/phi-2
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-gpo-renew2-b0.001-extra-v2-i1
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# phi-2-gpo-renew2-b0.001-extra-v2-i1
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on the None dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.0388
|
| 22 |
+
- Rewards/chosen: 0.0266
|
| 23 |
+
- Rewards/rejected: -0.0126
|
| 24 |
+
- Rewards/accuracies: 0.6070
|
| 25 |
+
- Rewards/margins: 0.0392
|
| 26 |
+
- Logps/rejected: -379.8497
|
| 27 |
+
- Logps/chosen: -369.7509
|
| 28 |
+
- Logits/rejected: -0.9196
|
| 29 |
+
- Logits/chosen: -0.9539
|
| 30 |
+
|
| 31 |
+
## Model description
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Intended uses & limitations
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Training and evaluation data
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Training procedure
|
| 44 |
+
|
| 45 |
+
### Training hyperparameters
|
| 46 |
+
|
| 47 |
+
The following hyperparameters were used during training:
|
| 48 |
+
- learning_rate: 5e-06
|
| 49 |
+
- train_batch_size: 4
|
| 50 |
+
- eval_batch_size: 4
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- gradient_accumulation_steps: 4
|
| 54 |
+
- total_train_batch_size: 16
|
| 55 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 56 |
+
- lr_scheduler_type: cosine
|
| 57 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 58 |
+
- num_epochs: 1
|
| 59 |
+
|
| 60 |
+
### Training results
|
| 61 |
+
|
| 62 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 63 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 64 |
+
| 0.098 | 0.06 | 100 | 0.0533 | -0.0029 | -0.0036 | 0.4980 | 0.0007 | -370.8433 | -399.2503 | -0.7225 | -0.8171 |
|
| 65 |
+
| 0.094 | 0.13 | 200 | 0.0491 | -0.0390 | -0.0525 | 0.5525 | 0.0135 | -419.6949 | -435.2693 | -1.0754 | -1.1388 |
|
| 66 |
+
| 0.0898 | 0.19 | 300 | 0.0452 | -0.0184 | -0.0403 | 0.5780 | 0.0218 | -407.5088 | -414.7480 | -1.0291 | -1.0858 |
|
| 67 |
+
| 0.0731 | 0.26 | 400 | 0.0430 | -0.0069 | -0.0331 | 0.5970 | 0.0262 | -400.2979 | -403.1916 | -0.9864 | -1.0412 |
|
| 68 |
+
| 0.0787 | 0.32 | 500 | 0.0422 | -0.0122 | -0.0473 | 0.6070 | 0.0351 | -414.4887 | -408.4566 | -1.0587 | -1.0975 |
|
| 69 |
+
| 0.0742 | 0.38 | 600 | 0.0406 | 0.0135 | -0.0175 | 0.6085 | 0.0309 | -384.7105 | -382.8363 | -0.9872 | -1.0246 |
|
| 70 |
+
| 0.0635 | 0.45 | 700 | 0.0401 | 0.0166 | -0.0188 | 0.6095 | 0.0354 | -386.0258 | -379.6696 | -0.9903 | -1.0225 |
|
| 71 |
+
| 0.0881 | 0.51 | 800 | 0.0395 | 0.0250 | -0.0102 | 0.6085 | 0.0352 | -377.4323 | -371.2672 | -0.9658 | -0.9975 |
|
| 72 |
+
| 0.0753 | 0.58 | 900 | 0.0393 | 0.0304 | -0.0046 | 0.5990 | 0.0350 | -371.7872 | -365.8699 | -0.9026 | -0.9456 |
|
| 73 |
+
| 0.0922 | 0.64 | 1000 | 0.0390 | 0.0286 | -0.0075 | 0.5990 | 0.0361 | -374.7669 | -367.7319 | -0.8801 | -0.9184 |
|
| 74 |
+
| 0.0703 | 0.7 | 1100 | 0.0389 | 0.0227 | -0.0161 | 0.6000 | 0.0387 | -383.3026 | -373.6226 | -0.9300 | -0.9602 |
|
| 75 |
+
| 0.0746 | 0.77 | 1200 | 0.0388 | 0.0226 | -0.0179 | 0.6050 | 0.0405 | -385.1601 | -373.7153 | -0.8944 | -0.9306 |
|
| 76 |
+
| 0.0925 | 0.83 | 1300 | 0.0387 | 0.0263 | -0.0131 | 0.6030 | 0.0393 | -380.3072 | -370.0340 | -0.9171 | -0.9494 |
|
| 77 |
+
| 0.0863 | 0.9 | 1400 | 0.0387 | 0.0269 | -0.0123 | 0.6055 | 0.0392 | -379.5608 | -369.4450 | -0.9121 | -0.9447 |
|
| 78 |
+
| 0.0904 | 0.96 | 1500 | 0.0386 | 0.0268 | -0.0124 | 0.6045 | 0.0392 | -379.6000 | -369.4944 | -0.9203 | -0.9536 |
|
| 79 |
+
|
| 80 |
+
|
| 81 |
+
### Framework versions
|
| 82 |
+
|
| 83 |
+
- PEFT 0.7.1
|
| 84 |
+
- Transformers 4.36.2
|
| 85 |
+
- Pytorch 2.1.2
|
| 86 |
+
- Datasets 2.14.6
|
| 87 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 167807296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:575c533757ec39ed9b3aec4f881a7caece463a6ba3c82324f139f9e776da1955
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_logits/chosen": -0.9538615942001343,
|
| 4 |
+
"eval_logits/rejected": -0.9195562601089478,
|
| 5 |
+
"eval_logps/chosen": -369.7508544921875,
|
| 6 |
+
"eval_logps/rejected": -379.8497314453125,
|
| 7 |
+
"eval_loss": 0.03875432908535004,
|
| 8 |
+
"eval_rewards/accuracies": 0.6069999933242798,
|
| 9 |
+
"eval_rewards/chosen": 0.02655443549156189,
|
| 10 |
+
"eval_rewards/margins": 0.03916873782873154,
|
| 11 |
+
"eval_rewards/rejected": -0.012614301405847073,
|
| 12 |
+
"eval_runtime": 539.9666,
|
| 13 |
+
"eval_samples": 2000,
|
| 14 |
+
"eval_samples_per_second": 3.704,
|
| 15 |
+
"eval_steps_per_second": 0.926,
|
| 16 |
+
"train_loss": 0.08419492204701015,
|
| 17 |
+
"train_runtime": 22113.8812,
|
| 18 |
+
"train_samples": 61135,
|
| 19 |
+
"train_samples_per_second": 1.131,
|
| 20 |
+
"train_steps_per_second": 0.071
|
| 21 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_logits/chosen": -0.9538615942001343,
|
| 4 |
+
"eval_logits/rejected": -0.9195562601089478,
|
| 5 |
+
"eval_logps/chosen": -369.7508544921875,
|
| 6 |
+
"eval_logps/rejected": -379.8497314453125,
|
| 7 |
+
"eval_loss": 0.03875432908535004,
|
| 8 |
+
"eval_rewards/accuracies": 0.6069999933242798,
|
| 9 |
+
"eval_rewards/chosen": 0.02655443549156189,
|
| 10 |
+
"eval_rewards/margins": 0.03916873782873154,
|
| 11 |
+
"eval_rewards/rejected": -0.012614301405847073,
|
| 12 |
+
"eval_runtime": 539.9666,
|
| 13 |
+
"eval_samples": 2000,
|
| 14 |
+
"eval_samples_per_second": 3.704,
|
| 15 |
+
"eval_steps_per_second": 0.926
|
| 16 |
+
}
|
runs/Apr24_18-06-13_gpu4-119-5/events.out.tfevents.1713946169.gpu4-119-5.4164190.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c41419a5d7825d35d06123f941ef2e00fd6b7364cfd3971c346e41154d7c8570
|
| 3 |
+
size 115710
|
runs/Apr24_18-06-13_gpu4-119-5/events.out.tfevents.1713968856.gpu4-119-5.4164190.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2ba85ec9de94c1c5352e8c32d4be230ba59d073d2464df6dad9c69552ff59e9
|
| 3 |
+
size 828
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.08419492204701015,
|
| 4 |
+
"train_runtime": 22113.8812,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 1.131,
|
| 7 |
+
"train_steps_per_second": 0.071
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2468 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.99968,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 1562,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 3.184713375796179e-08,
|
| 14 |
+
"logits/chosen": -0.9295870065689087,
|
| 15 |
+
"logits/rejected": -0.43873703479766846,
|
| 16 |
+
"logps/chosen": -320.5160827636719,
|
| 17 |
+
"logps/rejected": -293.4969482421875,
|
| 18 |
+
"loss": 0.0436,
|
| 19 |
+
"rewards/accuracies": 0.0,
|
| 20 |
+
"rewards/chosen": 0.0,
|
| 21 |
+
"rewards/margins": 0.0,
|
| 22 |
+
"rewards/rejected": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 3.1847133757961787e-07,
|
| 28 |
+
"logits/chosen": -0.8594987988471985,
|
| 29 |
+
"logits/rejected": -0.762122631072998,
|
| 30 |
+
"logps/chosen": -363.58154296875,
|
| 31 |
+
"logps/rejected": -320.7466735839844,
|
| 32 |
+
"loss": 0.1005,
|
| 33 |
+
"rewards/accuracies": 0.4444444477558136,
|
| 34 |
+
"rewards/chosen": 0.00017958095122594386,
|
| 35 |
+
"rewards/margins": 0.00031936122104525566,
|
| 36 |
+
"rewards/rejected": -0.0001397802698193118,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"learning_rate": 6.369426751592357e-07,
|
| 42 |
+
"logits/chosen": -0.8261665105819702,
|
| 43 |
+
"logits/rejected": -0.8099607229232788,
|
| 44 |
+
"logps/chosen": -329.95257568359375,
|
| 45 |
+
"logps/rejected": -333.58843994140625,
|
| 46 |
+
"loss": 0.0881,
|
| 47 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 48 |
+
"rewards/chosen": 1.5792587873875163e-05,
|
| 49 |
+
"rewards/margins": 9.325530118076131e-05,
|
| 50 |
+
"rewards/rejected": -7.746272603981197e-05,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 9.554140127388537e-07,
|
| 56 |
+
"logits/chosen": -0.8844400644302368,
|
| 57 |
+
"logits/rejected": -0.8693345189094543,
|
| 58 |
+
"logps/chosen": -266.14617919921875,
|
| 59 |
+
"logps/rejected": -239.8477020263672,
|
| 60 |
+
"loss": 0.1076,
|
| 61 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 62 |
+
"rewards/chosen": -0.00019076233729720116,
|
| 63 |
+
"rewards/margins": -0.00011497503146529198,
|
| 64 |
+
"rewards/rejected": -7.578730583190918e-05,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03,
|
| 69 |
+
"learning_rate": 1.2738853503184715e-06,
|
| 70 |
+
"logits/chosen": -0.9354842305183411,
|
| 71 |
+
"logits/rejected": -0.8044538497924805,
|
| 72 |
+
"logps/chosen": -337.16436767578125,
|
| 73 |
+
"logps/rejected": -294.0509338378906,
|
| 74 |
+
"loss": 0.1014,
|
| 75 |
+
"rewards/accuracies": 0.40625,
|
| 76 |
+
"rewards/chosen": -0.0002183823671657592,
|
| 77 |
+
"rewards/margins": 0.00015057336713653058,
|
| 78 |
+
"rewards/rejected": -0.00036895571975037456,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.03,
|
| 83 |
+
"learning_rate": 1.5923566878980892e-06,
|
| 84 |
+
"logits/chosen": -0.9226251840591431,
|
| 85 |
+
"logits/rejected": -0.7269760966300964,
|
| 86 |
+
"logps/chosen": -291.8105773925781,
|
| 87 |
+
"logps/rejected": -249.9472198486328,
|
| 88 |
+
"loss": 0.0984,
|
| 89 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 90 |
+
"rewards/chosen": -1.6449857866973616e-05,
|
| 91 |
+
"rewards/margins": -4.999707016395405e-05,
|
| 92 |
+
"rewards/rejected": 3.354722139192745e-05,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.04,
|
| 97 |
+
"learning_rate": 1.9108280254777074e-06,
|
| 98 |
+
"logits/chosen": -0.7843996286392212,
|
| 99 |
+
"logits/rejected": -0.8503357768058777,
|
| 100 |
+
"logps/chosen": -314.31842041015625,
|
| 101 |
+
"logps/rejected": -272.9629821777344,
|
| 102 |
+
"loss": 0.0831,
|
| 103 |
+
"rewards/accuracies": 0.375,
|
| 104 |
+
"rewards/chosen": -0.0004891738062724471,
|
| 105 |
+
"rewards/margins": 3.405969255254604e-05,
|
| 106 |
+
"rewards/rejected": -0.0005232334951870143,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.04,
|
| 111 |
+
"learning_rate": 2.229299363057325e-06,
|
| 112 |
+
"logits/chosen": -0.8444937467575073,
|
| 113 |
+
"logits/rejected": -0.8687158823013306,
|
| 114 |
+
"logps/chosen": -346.148193359375,
|
| 115 |
+
"logps/rejected": -314.0673828125,
|
| 116 |
+
"loss": 0.101,
|
| 117 |
+
"rewards/accuracies": 0.4375,
|
| 118 |
+
"rewards/chosen": -0.0008814434404484928,
|
| 119 |
+
"rewards/margins": 0.0003942731418646872,
|
| 120 |
+
"rewards/rejected": -0.00127571658231318,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.05,
|
| 125 |
+
"learning_rate": 2.547770700636943e-06,
|
| 126 |
+
"logits/chosen": -0.7672846913337708,
|
| 127 |
+
"logits/rejected": -0.8321496844291687,
|
| 128 |
+
"logps/chosen": -308.66973876953125,
|
| 129 |
+
"logps/rejected": -290.4983215332031,
|
| 130 |
+
"loss": 0.0897,
|
| 131 |
+
"rewards/accuracies": 0.375,
|
| 132 |
+
"rewards/chosen": -0.0009476385894231498,
|
| 133 |
+
"rewards/margins": 0.0005286627565510571,
|
| 134 |
+
"rewards/rejected": -0.0014763014623895288,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.06,
|
| 139 |
+
"learning_rate": 2.8662420382165605e-06,
|
| 140 |
+
"logits/chosen": -0.9202607870101929,
|
| 141 |
+
"logits/rejected": -0.6941283345222473,
|
| 142 |
+
"logps/chosen": -294.6612854003906,
|
| 143 |
+
"logps/rejected": -279.28497314453125,
|
| 144 |
+
"loss": 0.0967,
|
| 145 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 146 |
+
"rewards/chosen": -0.0010531718144193292,
|
| 147 |
+
"rewards/margins": 0.0009010445210151374,
|
| 148 |
+
"rewards/rejected": -0.0019542162772268057,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.06,
|
| 153 |
+
"learning_rate": 3.1847133757961785e-06,
|
| 154 |
+
"logits/chosen": -0.8830461502075195,
|
| 155 |
+
"logits/rejected": -0.7728676795959473,
|
| 156 |
+
"logps/chosen": -346.04144287109375,
|
| 157 |
+
"logps/rejected": -351.6465759277344,
|
| 158 |
+
"loss": 0.098,
|
| 159 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 160 |
+
"rewards/chosen": -0.0019343973835930228,
|
| 161 |
+
"rewards/margins": 0.0005748984985984862,
|
| 162 |
+
"rewards/rejected": -0.00250929594039917,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.06,
|
| 167 |
+
"eval_logits/chosen": -0.8170500993728638,
|
| 168 |
+
"eval_logits/rejected": -0.7225118279457092,
|
| 169 |
+
"eval_logps/chosen": -399.25030517578125,
|
| 170 |
+
"eval_logps/rejected": -370.84332275390625,
|
| 171 |
+
"eval_loss": 0.053326018154621124,
|
| 172 |
+
"eval_rewards/accuracies": 0.49799999594688416,
|
| 173 |
+
"eval_rewards/chosen": -0.002945071319118142,
|
| 174 |
+
"eval_rewards/margins": 0.0006628122646361589,
|
| 175 |
+
"eval_rewards/rejected": -0.0036078833509236574,
|
| 176 |
+
"eval_runtime": 539.8002,
|
| 177 |
+
"eval_samples_per_second": 3.705,
|
| 178 |
+
"eval_steps_per_second": 0.926,
|
| 179 |
+
"step": 100
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.07,
|
| 183 |
+
"learning_rate": 3.5031847133757964e-06,
|
| 184 |
+
"logits/chosen": -0.8748549222946167,
|
| 185 |
+
"logits/rejected": -0.8573201298713684,
|
| 186 |
+
"logps/chosen": -356.620361328125,
|
| 187 |
+
"logps/rejected": -334.7717590332031,
|
| 188 |
+
"loss": 0.0783,
|
| 189 |
+
"rewards/accuracies": 0.46875,
|
| 190 |
+
"rewards/chosen": -0.0018858186667785048,
|
| 191 |
+
"rewards/margins": 0.0009479810250923038,
|
| 192 |
+
"rewards/rejected": -0.0028337999247014523,
|
| 193 |
+
"step": 110
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.08,
|
| 197 |
+
"learning_rate": 3.821656050955415e-06,
|
| 198 |
+
"logits/chosen": -0.9678497314453125,
|
| 199 |
+
"logits/rejected": -0.7722519636154175,
|
| 200 |
+
"logps/chosen": -313.03741455078125,
|
| 201 |
+
"logps/rejected": -280.0140686035156,
|
| 202 |
+
"loss": 0.1122,
|
| 203 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 204 |
+
"rewards/chosen": -0.0022745749447494745,
|
| 205 |
+
"rewards/margins": 0.0005548128974623978,
|
| 206 |
+
"rewards/rejected": -0.0028293877840042114,
|
| 207 |
+
"step": 120
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.08,
|
| 211 |
+
"learning_rate": 4.140127388535032e-06,
|
| 212 |
+
"logits/chosen": -0.8600385785102844,
|
| 213 |
+
"logits/rejected": -0.8846877813339233,
|
| 214 |
+
"logps/chosen": -355.20513916015625,
|
| 215 |
+
"logps/rejected": -346.33135986328125,
|
| 216 |
+
"loss": 0.098,
|
| 217 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 218 |
+
"rewards/chosen": -0.005060167517513037,
|
| 219 |
+
"rewards/margins": -0.00011312137212371454,
|
| 220 |
+
"rewards/rejected": -0.004947046283632517,
|
| 221 |
+
"step": 130
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.09,
|
| 225 |
+
"learning_rate": 4.45859872611465e-06,
|
| 226 |
+
"logits/chosen": -0.848353385925293,
|
| 227 |
+
"logits/rejected": -0.8213680386543274,
|
| 228 |
+
"logps/chosen": -355.6943054199219,
|
| 229 |
+
"logps/rejected": -313.1587829589844,
|
| 230 |
+
"loss": 0.0975,
|
| 231 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 232 |
+
"rewards/chosen": -0.005685682408511639,
|
| 233 |
+
"rewards/margins": 0.0019586696289479733,
|
| 234 |
+
"rewards/rejected": -0.007644351571798325,
|
| 235 |
+
"step": 140
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.1,
|
| 239 |
+
"learning_rate": 4.777070063694268e-06,
|
| 240 |
+
"logits/chosen": -0.9287412762641907,
|
| 241 |
+
"logits/rejected": -0.8769465684890747,
|
| 242 |
+
"logps/chosen": -329.2186584472656,
|
| 243 |
+
"logps/rejected": -298.8802795410156,
|
| 244 |
+
"loss": 0.0712,
|
| 245 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 246 |
+
"rewards/chosen": -0.00595915038138628,
|
| 247 |
+
"rewards/margins": 0.0024761247914284468,
|
| 248 |
+
"rewards/rejected": -0.00843527540564537,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.1,
|
| 253 |
+
"learning_rate": 4.999943753177818e-06,
|
| 254 |
+
"logits/chosen": -1.005480170249939,
|
| 255 |
+
"logits/rejected": -0.9711716771125793,
|
| 256 |
+
"logps/chosen": -311.71209716796875,
|
| 257 |
+
"logps/rejected": -299.3721923828125,
|
| 258 |
+
"loss": 0.0886,
|
| 259 |
+
"rewards/accuracies": 0.4375,
|
| 260 |
+
"rewards/chosen": -0.00825573317706585,
|
| 261 |
+
"rewards/margins": 0.0037352764047682285,
|
| 262 |
+
"rewards/rejected": -0.011991010047495365,
|
| 263 |
+
"step": 160
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.11,
|
| 267 |
+
"learning_rate": 4.998943880079481e-06,
|
| 268 |
+
"logits/chosen": -0.9299044609069824,
|
| 269 |
+
"logits/rejected": -0.9373615384101868,
|
| 270 |
+
"logps/chosen": -343.8020324707031,
|
| 271 |
+
"logps/rejected": -338.0310974121094,
|
| 272 |
+
"loss": 0.0924,
|
| 273 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 274 |
+
"rewards/chosen": -0.01192331500351429,
|
| 275 |
+
"rewards/margins": 0.0040515875443816185,
|
| 276 |
+
"rewards/rejected": -0.015974899753928185,
|
| 277 |
+
"step": 170
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.12,
|
| 281 |
+
"learning_rate": 4.99669465299937e-06,
|
| 282 |
+
"logits/chosen": -0.9855674505233765,
|
| 283 |
+
"logits/rejected": -0.9634091258049011,
|
| 284 |
+
"logps/chosen": -309.28851318359375,
|
| 285 |
+
"logps/rejected": -315.94561767578125,
|
| 286 |
+
"loss": 0.0928,
|
| 287 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 288 |
+
"rewards/chosen": -0.015265477821230888,
|
| 289 |
+
"rewards/margins": 0.005659168586134911,
|
| 290 |
+
"rewards/rejected": -0.0209246464073658,
|
| 291 |
+
"step": 180
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.12,
|
| 295 |
+
"learning_rate": 4.993197196444851e-06,
|
| 296 |
+
"logits/chosen": -1.0295236110687256,
|
| 297 |
+
"logits/rejected": -1.1208027601242065,
|
| 298 |
+
"logps/chosen": -317.75494384765625,
|
| 299 |
+
"logps/rejected": -328.9816589355469,
|
| 300 |
+
"loss": 0.0805,
|
| 301 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 302 |
+
"rewards/chosen": -0.020896239206194878,
|
| 303 |
+
"rewards/margins": 0.007452984340488911,
|
| 304 |
+
"rewards/rejected": -0.028349224478006363,
|
| 305 |
+
"step": 190
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.13,
|
| 309 |
+
"learning_rate": 4.988453258979111e-06,
|
| 310 |
+
"logits/chosen": -1.1082799434661865,
|
| 311 |
+
"logits/rejected": -1.0890588760375977,
|
| 312 |
+
"logps/chosen": -350.9955139160156,
|
| 313 |
+
"logps/rejected": -370.35955810546875,
|
| 314 |
+
"loss": 0.094,
|
| 315 |
+
"rewards/accuracies": 0.518750011920929,
|
| 316 |
+
"rewards/chosen": -0.02722758986055851,
|
| 317 |
+
"rewards/margins": 0.009304001927375793,
|
| 318 |
+
"rewards/rejected": -0.036531589925289154,
|
| 319 |
+
"step": 200
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.13,
|
| 323 |
+
"eval_logits/chosen": -1.1387817859649658,
|
| 324 |
+
"eval_logits/rejected": -1.0754390954971313,
|
| 325 |
+
"eval_logps/chosen": -435.2693176269531,
|
| 326 |
+
"eval_logps/rejected": -419.6948547363281,
|
| 327 |
+
"eval_loss": 0.04912900552153587,
|
| 328 |
+
"eval_rewards/accuracies": 0.5525000095367432,
|
| 329 |
+
"eval_rewards/chosen": -0.03896407037973404,
|
| 330 |
+
"eval_rewards/margins": 0.013495376333594322,
|
| 331 |
+
"eval_rewards/rejected": -0.05245944485068321,
|
| 332 |
+
"eval_runtime": 539.703,
|
| 333 |
+
"eval_samples_per_second": 3.706,
|
| 334 |
+
"eval_steps_per_second": 0.926,
|
| 335 |
+
"step": 200
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.13,
|
| 339 |
+
"learning_rate": 4.982465212346954e-06,
|
| 340 |
+
"logits/chosen": -1.1744426488876343,
|
| 341 |
+
"logits/rejected": -1.174753189086914,
|
| 342 |
+
"logps/chosen": -324.98370361328125,
|
| 343 |
+
"logps/rejected": -325.42767333984375,
|
| 344 |
+
"loss": 0.1019,
|
| 345 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 346 |
+
"rewards/chosen": -0.031165916472673416,
|
| 347 |
+
"rewards/margins": 0.018221553415060043,
|
| 348 |
+
"rewards/rejected": -0.04938746988773346,
|
| 349 |
+
"step": 210
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.14,
|
| 353 |
+
"learning_rate": 4.975236050289041e-06,
|
| 354 |
+
"logits/chosen": -1.1808674335479736,
|
| 355 |
+
"logits/rejected": -1.0834966897964478,
|
| 356 |
+
"logps/chosen": -334.90087890625,
|
| 357 |
+
"logps/rejected": -318.37060546875,
|
| 358 |
+
"loss": 0.0697,
|
| 359 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 360 |
+
"rewards/chosen": -0.028991717845201492,
|
| 361 |
+
"rewards/margins": 0.01802077516913414,
|
| 362 |
+
"rewards/rejected": -0.04701249301433563,
|
| 363 |
+
"step": 220
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.15,
|
| 367 |
+
"learning_rate": 4.96676938704516e-06,
|
| 368 |
+
"logits/chosen": -1.1029061079025269,
|
| 369 |
+
"logits/rejected": -1.1640408039093018,
|
| 370 |
+
"logps/chosen": -435.73687744140625,
|
| 371 |
+
"logps/rejected": -423.72900390625,
|
| 372 |
+
"loss": 0.0749,
|
| 373 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 374 |
+
"rewards/chosen": -0.040695372968912125,
|
| 375 |
+
"rewards/margins": 0.014833291992545128,
|
| 376 |
+
"rewards/rejected": -0.0555286630988121,
|
| 377 |
+
"step": 230
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.15,
|
| 381 |
+
"learning_rate": 4.95706945554728e-06,
|
| 382 |
+
"logits/chosen": -1.1478805541992188,
|
| 383 |
+
"logits/rejected": -1.2186776399612427,
|
| 384 |
+
"logps/chosen": -370.34503173828125,
|
| 385 |
+
"logps/rejected": -392.4799499511719,
|
| 386 |
+
"loss": 0.0734,
|
| 387 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 388 |
+
"rewards/chosen": -0.0380186066031456,
|
| 389 |
+
"rewards/margins": 0.017158757895231247,
|
| 390 |
+
"rewards/rejected": -0.05517736077308655,
|
| 391 |
+
"step": 240
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.16,
|
| 395 |
+
"learning_rate": 4.9461411053032805e-06,
|
| 396 |
+
"logits/chosen": -1.1441445350646973,
|
| 397 |
+
"logits/rejected": -1.1573059558868408,
|
| 398 |
+
"logps/chosen": -357.66925048828125,
|
| 399 |
+
"logps/rejected": -347.88848876953125,
|
| 400 |
+
"loss": 0.1096,
|
| 401 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 402 |
+
"rewards/chosen": -0.03414061293005943,
|
| 403 |
+
"rewards/margins": 0.009822173975408077,
|
| 404 |
+
"rewards/rejected": -0.043962787836790085,
|
| 405 |
+
"step": 250
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.17,
|
| 409 |
+
"learning_rate": 4.933989799972431e-06,
|
| 410 |
+
"logits/chosen": -1.1844251155853271,
|
| 411 |
+
"logits/rejected": -1.1023800373077393,
|
| 412 |
+
"logps/chosen": -362.12890625,
|
| 413 |
+
"logps/rejected": -344.55694580078125,
|
| 414 |
+
"loss": 0.087,
|
| 415 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 416 |
+
"rewards/chosen": -0.023841477930545807,
|
| 417 |
+
"rewards/margins": 0.01648232527077198,
|
| 418 |
+
"rewards/rejected": -0.04032380133867264,
|
| 419 |
+
"step": 260
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.17,
|
| 423 |
+
"learning_rate": 4.920621614633815e-06,
|
| 424 |
+
"logits/chosen": -1.158454418182373,
|
| 425 |
+
"logits/rejected": -1.1034467220306396,
|
| 426 |
+
"logps/chosen": -350.0291748046875,
|
| 427 |
+
"logps/rejected": -357.4220886230469,
|
| 428 |
+
"loss": 0.094,
|
| 429 |
+
"rewards/accuracies": 0.46875,
|
| 430 |
+
"rewards/chosen": -0.02827698551118374,
|
| 431 |
+
"rewards/margins": 0.01563875935971737,
|
| 432 |
+
"rewards/rejected": -0.04391574487090111,
|
| 433 |
+
"step": 270
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.18,
|
| 437 |
+
"learning_rate": 4.906043232749081e-06,
|
| 438 |
+
"logits/chosen": -1.1796420812606812,
|
| 439 |
+
"logits/rejected": -0.9934653043746948,
|
| 440 |
+
"logps/chosen": -337.7796325683594,
|
| 441 |
+
"logps/rejected": -324.3747863769531,
|
| 442 |
+
"loss": 0.0808,
|
| 443 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 444 |
+
"rewards/chosen": -0.02835283800959587,
|
| 445 |
+
"rewards/margins": 0.013374457135796547,
|
| 446 |
+
"rewards/rejected": -0.04172729700803757,
|
| 447 |
+
"step": 280
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.19,
|
| 451 |
+
"learning_rate": 4.890261942821023e-06,
|
| 452 |
+
"logits/chosen": -1.0947716236114502,
|
| 453 |
+
"logits/rejected": -1.134075403213501,
|
| 454 |
+
"logps/chosen": -331.79107666015625,
|
| 455 |
+
"logps/rejected": -309.49609375,
|
| 456 |
+
"loss": 0.1014,
|
| 457 |
+
"rewards/accuracies": 0.46875,
|
| 458 |
+
"rewards/chosen": -0.024125058203935623,
|
| 459 |
+
"rewards/margins": 0.01503480039536953,
|
| 460 |
+
"rewards/rejected": -0.0391598604619503,
|
| 461 |
+
"step": 290
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.19,
|
| 465 |
+
"learning_rate": 4.873285634749678e-06,
|
| 466 |
+
"logits/chosen": -1.1100094318389893,
|
| 467 |
+
"logits/rejected": -1.0816996097564697,
|
| 468 |
+
"logps/chosen": -311.70867919921875,
|
| 469 |
+
"logps/rejected": -315.1243591308594,
|
| 470 |
+
"loss": 0.0898,
|
| 471 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 472 |
+
"rewards/chosen": -0.017298461869359016,
|
| 473 |
+
"rewards/margins": 0.013444353826344013,
|
| 474 |
+
"rewards/rejected": -0.030742818489670753,
|
| 475 |
+
"step": 300
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.19,
|
| 479 |
+
"eval_logits/chosen": -1.0857516527175903,
|
| 480 |
+
"eval_logits/rejected": -1.0291430950164795,
|
| 481 |
+
"eval_logps/chosen": -414.7480163574219,
|
| 482 |
+
"eval_logps/rejected": -407.5087585449219,
|
| 483 |
+
"eval_loss": 0.045198723673820496,
|
| 484 |
+
"eval_rewards/accuracies": 0.578000009059906,
|
| 485 |
+
"eval_rewards/chosen": -0.018442772328853607,
|
| 486 |
+
"eval_rewards/margins": 0.021830614656209946,
|
| 487 |
+
"eval_rewards/rejected": -0.04027338698506355,
|
| 488 |
+
"eval_runtime": 539.5827,
|
| 489 |
+
"eval_samples_per_second": 3.707,
|
| 490 |
+
"eval_steps_per_second": 0.927,
|
| 491 |
+
"step": 300
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.2,
|
| 495 |
+
"learning_rate": 4.855122795887746e-06,
|
| 496 |
+
"logits/chosen": -1.1289124488830566,
|
| 497 |
+
"logits/rejected": -1.06770658493042,
|
| 498 |
+
"logps/chosen": -297.93218994140625,
|
| 499 |
+
"logps/rejected": -279.66107177734375,
|
| 500 |
+
"loss": 0.0967,
|
| 501 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 502 |
+
"rewards/chosen": -0.0071363793686032295,
|
| 503 |
+
"rewards/margins": 0.01606178656220436,
|
| 504 |
+
"rewards/rejected": -0.023198166862130165,
|
| 505 |
+
"step": 310
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.2,
|
| 509 |
+
"learning_rate": 4.83578250679731e-06,
|
| 510 |
+
"logits/chosen": -1.0398236513137817,
|
| 511 |
+
"logits/rejected": -1.0716559886932373,
|
| 512 |
+
"logps/chosen": -323.6569519042969,
|
| 513 |
+
"logps/rejected": -321.41845703125,
|
| 514 |
+
"loss": 0.1007,
|
| 515 |
+
"rewards/accuracies": 0.5,
|
| 516 |
+
"rewards/chosen": -0.01684439741075039,
|
| 517 |
+
"rewards/margins": 0.016590449959039688,
|
| 518 |
+
"rewards/rejected": -0.03343484550714493,
|
| 519 |
+
"step": 320
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.21,
|
| 523 |
+
"learning_rate": 4.8152744367099935e-06,
|
| 524 |
+
"logits/chosen": -1.13577139377594,
|
| 525 |
+
"logits/rejected": -1.0597126483917236,
|
| 526 |
+
"logps/chosen": -253.326171875,
|
| 527 |
+
"logps/rejected": -260.37652587890625,
|
| 528 |
+
"loss": 0.1324,
|
| 529 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 530 |
+
"rewards/chosen": -0.011873602867126465,
|
| 531 |
+
"rewards/margins": 0.021937990561127663,
|
| 532 |
+
"rewards/rejected": -0.03381159156560898,
|
| 533 |
+
"step": 330
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.22,
|
| 537 |
+
"learning_rate": 4.793608838692792e-06,
|
| 538 |
+
"logits/chosen": -1.05599844455719,
|
| 539 |
+
"logits/rejected": -1.1052017211914062,
|
| 540 |
+
"logps/chosen": -330.0927734375,
|
| 541 |
+
"logps/rejected": -354.5337829589844,
|
| 542 |
+
"loss": 0.0709,
|
| 543 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 544 |
+
"rewards/chosen": -0.014163943938910961,
|
| 545 |
+
"rewards/margins": 0.0242028646171093,
|
| 546 |
+
"rewards/rejected": -0.038366809487342834,
|
| 547 |
+
"step": 340
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 0.22,
|
| 551 |
+
"learning_rate": 4.770796544522026e-06,
|
| 552 |
+
"logits/chosen": -1.1162309646606445,
|
| 553 |
+
"logits/rejected": -1.0977243185043335,
|
| 554 |
+
"logps/chosen": -298.46929931640625,
|
| 555 |
+
"logps/rejected": -301.0343017578125,
|
| 556 |
+
"loss": 0.0828,
|
| 557 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 558 |
+
"rewards/chosen": -0.014557460322976112,
|
| 559 |
+
"rewards/margins": 0.010886872187256813,
|
| 560 |
+
"rewards/rejected": -0.025444332510232925,
|
| 561 |
+
"step": 350
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.23,
|
| 565 |
+
"learning_rate": 4.746848959267968e-06,
|
| 566 |
+
"logits/chosen": -1.0554795265197754,
|
| 567 |
+
"logits/rejected": -0.9742960929870605,
|
| 568 |
+
"logps/chosen": -337.4122314453125,
|
| 569 |
+
"logps/rejected": -334.58367919921875,
|
| 570 |
+
"loss": 0.0999,
|
| 571 |
+
"rewards/accuracies": 0.46875,
|
| 572 |
+
"rewards/chosen": -0.002840764820575714,
|
| 573 |
+
"rewards/margins": 0.01675793156027794,
|
| 574 |
+
"rewards/rejected": -0.019598694518208504,
|
| 575 |
+
"step": 360
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.24,
|
| 579 |
+
"learning_rate": 4.721778055592841e-06,
|
| 580 |
+
"logits/chosen": -0.995721161365509,
|
| 581 |
+
"logits/rejected": -1.009927749633789,
|
| 582 |
+
"logps/chosen": -307.5436706542969,
|
| 583 |
+
"logps/rejected": -305.3434753417969,
|
| 584 |
+
"loss": 0.0958,
|
| 585 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 586 |
+
"rewards/chosen": 0.0018125835340470076,
|
| 587 |
+
"rewards/margins": 0.01936521753668785,
|
| 588 |
+
"rewards/rejected": -0.017552632838487625,
|
| 589 |
+
"step": 370
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.24,
|
| 593 |
+
"learning_rate": 4.695596367765054e-06,
|
| 594 |
+
"logits/chosen": -0.9923852682113647,
|
| 595 |
+
"logits/rejected": -0.9759656190872192,
|
| 596 |
+
"logps/chosen": -369.8957824707031,
|
| 597 |
+
"logps/rejected": -335.42694091796875,
|
| 598 |
+
"loss": 0.0605,
|
| 599 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 600 |
+
"rewards/chosen": -0.0068010808899998665,
|
| 601 |
+
"rewards/margins": 0.011534233577549458,
|
| 602 |
+
"rewards/rejected": -0.018335314467549324,
|
| 603 |
+
"step": 380
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.25,
|
| 607 |
+
"learning_rate": 4.6683169853926615e-06,
|
| 608 |
+
"logits/chosen": -1.056213617324829,
|
| 609 |
+
"logits/rejected": -1.1043860912322998,
|
| 610 |
+
"logps/chosen": -302.78570556640625,
|
| 611 |
+
"logps/rejected": -315.38604736328125,
|
| 612 |
+
"loss": 0.0863,
|
| 613 |
+
"rewards/accuracies": 0.5,
|
| 614 |
+
"rewards/chosen": -0.00549090001732111,
|
| 615 |
+
"rewards/margins": 0.01604645326733589,
|
| 616 |
+
"rewards/rejected": -0.021537352353334427,
|
| 617 |
+
"step": 390
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.26,
|
| 621 |
+
"learning_rate": 4.639953546879173e-06,
|
| 622 |
+
"logits/chosen": -1.011988639831543,
|
| 623 |
+
"logits/rejected": -0.9677278399467468,
|
| 624 |
+
"logps/chosen": -352.2360534667969,
|
| 625 |
+
"logps/rejected": -355.6038513183594,
|
| 626 |
+
"loss": 0.0731,
|
| 627 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 628 |
+
"rewards/chosen": -0.0009747882140800357,
|
| 629 |
+
"rewards/margins": 0.02269767038524151,
|
| 630 |
+
"rewards/rejected": -0.02367245778441429,
|
| 631 |
+
"step": 400
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.26,
|
| 635 |
+
"eval_logits/chosen": -1.0412452220916748,
|
| 636 |
+
"eval_logits/rejected": -0.9863678812980652,
|
| 637 |
+
"eval_logps/chosen": -403.1915588378906,
|
| 638 |
+
"eval_logps/rejected": -400.2978820800781,
|
| 639 |
+
"eval_loss": 0.04299690201878548,
|
| 640 |
+
"eval_rewards/accuracies": 0.597000002861023,
|
| 641 |
+
"eval_rewards/chosen": -0.006886274088174105,
|
| 642 |
+
"eval_rewards/margins": 0.026176199316978455,
|
| 643 |
+
"eval_rewards/rejected": -0.03306247293949127,
|
| 644 |
+
"eval_runtime": 539.7275,
|
| 645 |
+
"eval_samples_per_second": 3.706,
|
| 646 |
+
"eval_steps_per_second": 0.926,
|
| 647 |
+
"step": 400
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.26,
|
| 651 |
+
"learning_rate": 4.610520232605e-06,
|
| 652 |
+
"logits/chosen": -1.0571563243865967,
|
| 653 |
+
"logits/rejected": -0.9747453927993774,
|
| 654 |
+
"logps/chosen": -360.03936767578125,
|
| 655 |
+
"logps/rejected": -312.3787536621094,
|
| 656 |
+
"loss": 0.0706,
|
| 657 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 658 |
+
"rewards/chosen": -0.007904710248112679,
|
| 659 |
+
"rewards/margins": 0.012304016388952732,
|
| 660 |
+
"rewards/rejected": -0.020208725705742836,
|
| 661 |
+
"step": 410
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.27,
|
| 665 |
+
"learning_rate": 4.580031757837931e-06,
|
| 666 |
+
"logits/chosen": -1.0687059164047241,
|
| 667 |
+
"logits/rejected": -1.028510570526123,
|
| 668 |
+
"logps/chosen": -309.14532470703125,
|
| 669 |
+
"logps/rejected": -316.378662109375,
|
| 670 |
+
"loss": 0.0713,
|
| 671 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 672 |
+
"rewards/chosen": -0.0021046344190835953,
|
| 673 |
+
"rewards/margins": 0.0198093019425869,
|
| 674 |
+
"rewards/rejected": -0.021913940086960793,
|
| 675 |
+
"step": 420
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.28,
|
| 679 |
+
"learning_rate": 4.5485033653761936e-06,
|
| 680 |
+
"logits/chosen": -1.0271378755569458,
|
| 681 |
+
"logits/rejected": -1.0654436349868774,
|
| 682 |
+
"logps/chosen": -289.8176574707031,
|
| 683 |
+
"logps/rejected": -310.6886291503906,
|
| 684 |
+
"loss": 0.0907,
|
| 685 |
+
"rewards/accuracies": 0.5,
|
| 686 |
+
"rewards/chosen": 0.001297330716624856,
|
| 687 |
+
"rewards/margins": 0.02493324503302574,
|
| 688 |
+
"rewards/rejected": -0.02363591641187668,
|
| 689 |
+
"step": 430
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.28,
|
| 693 |
+
"learning_rate": 4.5159508179277775e-06,
|
| 694 |
+
"logits/chosen": -1.0649276971817017,
|
| 695 |
+
"logits/rejected": -1.0275962352752686,
|
| 696 |
+
"logps/chosen": -287.9720458984375,
|
| 697 |
+
"logps/rejected": -285.8517761230469,
|
| 698 |
+
"loss": 0.0857,
|
| 699 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 700 |
+
"rewards/chosen": 0.0025724810548126698,
|
| 701 |
+
"rewards/margins": 0.02007809281349182,
|
| 702 |
+
"rewards/rejected": -0.01750561222434044,
|
| 703 |
+
"step": 440
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.29,
|
| 707 |
+
"learning_rate": 4.48239039022982e-06,
|
| 708 |
+
"logits/chosen": -1.0931072235107422,
|
| 709 |
+
"logits/rejected": -0.9774026870727539,
|
| 710 |
+
"logps/chosen": -302.41497802734375,
|
| 711 |
+
"logps/rejected": -302.8403625488281,
|
| 712 |
+
"loss": 0.0786,
|
| 713 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 714 |
+
"rewards/chosen": 0.0017650052905082703,
|
| 715 |
+
"rewards/margins": 0.017620224505662918,
|
| 716 |
+
"rewards/rejected": -0.015855219215154648,
|
| 717 |
+
"step": 450
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.29,
|
| 721 |
+
"learning_rate": 4.447838860912011e-06,
|
| 722 |
+
"logits/chosen": -1.0981853008270264,
|
| 723 |
+
"logits/rejected": -0.9999715685844421,
|
| 724 |
+
"logps/chosen": -287.58599853515625,
|
| 725 |
+
"logps/rejected": -293.4311218261719,
|
| 726 |
+
"loss": 0.0808,
|
| 727 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 728 |
+
"rewards/chosen": -0.007796216756105423,
|
| 729 |
+
"rewards/margins": 0.02398526482284069,
|
| 730 |
+
"rewards/rejected": -0.03178148344159126,
|
| 731 |
+
"step": 460
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.3,
|
| 735 |
+
"learning_rate": 4.412313504108062e-06,
|
| 736 |
+
"logits/chosen": -1.1234943866729736,
|
| 737 |
+
"logits/rejected": -1.0496309995651245,
|
| 738 |
+
"logps/chosen": -343.08856201171875,
|
| 739 |
+
"logps/rejected": -333.7749938964844,
|
| 740 |
+
"loss": 0.0816,
|
| 741 |
+
"rewards/accuracies": 0.5,
|
| 742 |
+
"rewards/chosen": -0.00918793398886919,
|
| 743 |
+
"rewards/margins": 0.024025408551096916,
|
| 744 |
+
"rewards/rejected": -0.03321333974599838,
|
| 745 |
+
"step": 470
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.31,
|
| 749 |
+
"learning_rate": 4.375832080819465e-06,
|
| 750 |
+
"logits/chosen": -1.1054179668426514,
|
| 751 |
+
"logits/rejected": -1.063127040863037,
|
| 752 |
+
"logps/chosen": -352.350341796875,
|
| 753 |
+
"logps/rejected": -334.8752746582031,
|
| 754 |
+
"loss": 0.0853,
|
| 755 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 756 |
+
"rewards/chosen": -0.00490077119320631,
|
| 757 |
+
"rewards/margins": 0.02163395844399929,
|
| 758 |
+
"rewards/rejected": -0.026534726843237877,
|
| 759 |
+
"step": 480
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.31,
|
| 763 |
+
"learning_rate": 4.338412830035823e-06,
|
| 764 |
+
"logits/chosen": -1.0628612041473389,
|
| 765 |
+
"logits/rejected": -1.0398916006088257,
|
| 766 |
+
"logps/chosen": -347.806884765625,
|
| 767 |
+
"logps/rejected": -326.5227355957031,
|
| 768 |
+
"loss": 0.0819,
|
| 769 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 770 |
+
"rewards/chosen": 0.0002868110022973269,
|
| 771 |
+
"rewards/margins": 0.024012237787246704,
|
| 772 |
+
"rewards/rejected": -0.02372542954981327,
|
| 773 |
+
"step": 490
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.32,
|
| 777 |
+
"learning_rate": 4.300074459616216e-06,
|
| 778 |
+
"logits/chosen": -1.083707332611084,
|
| 779 |
+
"logits/rejected": -1.1182454824447632,
|
| 780 |
+
"logps/chosen": -304.49078369140625,
|
| 781 |
+
"logps/rejected": -335.3125,
|
| 782 |
+
"loss": 0.0787,
|
| 783 |
+
"rewards/accuracies": 0.5625,
|
| 784 |
+
"rewards/chosen": -0.0009085664642043412,
|
| 785 |
+
"rewards/margins": 0.03841399401426315,
|
| 786 |
+
"rewards/rejected": -0.039322562515735626,
|
| 787 |
+
"step": 500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.32,
|
| 791 |
+
"eval_logits/chosen": -1.0974537134170532,
|
| 792 |
+
"eval_logits/rejected": -1.0587471723556519,
|
| 793 |
+
"eval_logps/chosen": -408.4565734863281,
|
| 794 |
+
"eval_logps/rejected": -414.48870849609375,
|
| 795 |
+
"eval_loss": 0.0421723797917366,
|
| 796 |
+
"eval_rewards/accuracies": 0.6069999933242798,
|
| 797 |
+
"eval_rewards/chosen": -0.012151296250522137,
|
| 798 |
+
"eval_rewards/margins": 0.0351020023226738,
|
| 799 |
+
"eval_rewards/rejected": -0.04725329577922821,
|
| 800 |
+
"eval_runtime": 539.5972,
|
| 801 |
+
"eval_samples_per_second": 3.706,
|
| 802 |
+
"eval_steps_per_second": 0.927,
|
| 803 |
+
"step": 500
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.33,
|
| 807 |
+
"learning_rate": 4.260836136936159e-06,
|
| 808 |
+
"logits/chosen": -1.0836219787597656,
|
| 809 |
+
"logits/rejected": -1.0971637964248657,
|
| 810 |
+
"logps/chosen": -287.1617431640625,
|
| 811 |
+
"logps/rejected": -287.58148193359375,
|
| 812 |
+
"loss": 0.0893,
|
| 813 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 814 |
+
"rewards/chosen": -0.0011848447611555457,
|
| 815 |
+
"rewards/margins": 0.033604733645915985,
|
| 816 |
+
"rewards/rejected": -0.03478958457708359,
|
| 817 |
+
"step": 510
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.33,
|
| 821 |
+
"learning_rate": 4.220717479304816e-06,
|
| 822 |
+
"logits/chosen": -1.0768189430236816,
|
| 823 |
+
"logits/rejected": -1.0736405849456787,
|
| 824 |
+
"logps/chosen": -351.4105529785156,
|
| 825 |
+
"logps/rejected": -374.5068359375,
|
| 826 |
+
"loss": 0.092,
|
| 827 |
+
"rewards/accuracies": 0.550000011920929,
|
| 828 |
+
"rewards/chosen": -0.007005206309258938,
|
| 829 |
+
"rewards/margins": 0.034257300198078156,
|
| 830 |
+
"rewards/rejected": -0.04126249998807907,
|
| 831 |
+
"step": 520
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.34,
|
| 835 |
+
"learning_rate": 4.179738544157272e-06,
|
| 836 |
+
"logits/chosen": -1.0806314945220947,
|
| 837 |
+
"logits/rejected": -1.0778720378875732,
|
| 838 |
+
"logps/chosen": -275.35101318359375,
|
| 839 |
+
"logps/rejected": -270.8028869628906,
|
| 840 |
+
"loss": 0.0745,
|
| 841 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 842 |
+
"rewards/chosen": 0.0068043931387364864,
|
| 843 |
+
"rewards/margins": 0.013285738416016102,
|
| 844 |
+
"rewards/rejected": -0.006481344345957041,
|
| 845 |
+
"step": 530
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.35,
|
| 849 |
+
"learning_rate": 4.137919819026762e-06,
|
| 850 |
+
"logits/chosen": -1.0614066123962402,
|
| 851 |
+
"logits/rejected": -0.9721907377243042,
|
| 852 |
+
"logps/chosen": -313.3482971191406,
|
| 853 |
+
"logps/rejected": -321.7149963378906,
|
| 854 |
+
"loss": 0.076,
|
| 855 |
+
"rewards/accuracies": 0.518750011920929,
|
| 856 |
+
"rewards/chosen": 0.009386066347360611,
|
| 857 |
+
"rewards/margins": 0.01596895419061184,
|
| 858 |
+
"rewards/rejected": -0.006582888774573803,
|
| 859 |
+
"step": 540
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.35,
|
| 863 |
+
"learning_rate": 4.09528221130187e-06,
|
| 864 |
+
"logits/chosen": -1.0457605123519897,
|
| 865 |
+
"logits/rejected": -1.0901412963867188,
|
| 866 |
+
"logps/chosen": -314.82086181640625,
|
| 867 |
+
"logps/rejected": -286.0039367675781,
|
| 868 |
+
"loss": 0.0912,
|
| 869 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 870 |
+
"rewards/chosen": 0.009857202880084515,
|
| 871 |
+
"rewards/margins": 0.024335484951734543,
|
| 872 |
+
"rewards/rejected": -0.014478283934295177,
|
| 873 |
+
"step": 550
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.36,
|
| 877 |
+
"learning_rate": 4.0518470377738274e-06,
|
| 878 |
+
"logits/chosen": -1.0764890909194946,
|
| 879 |
+
"logits/rejected": -1.1156491041183472,
|
| 880 |
+
"logps/chosen": -293.7644348144531,
|
| 881 |
+
"logps/rejected": -315.3866882324219,
|
| 882 |
+
"loss": 0.0867,
|
| 883 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 884 |
+
"rewards/chosen": 0.011628219857811928,
|
| 885 |
+
"rewards/margins": 0.028168832883238792,
|
| 886 |
+
"rewards/rejected": -0.016540613025426865,
|
| 887 |
+
"step": 560
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.36,
|
| 891 |
+
"learning_rate": 4.0076360139791155e-06,
|
| 892 |
+
"logits/chosen": -1.0172516107559204,
|
| 893 |
+
"logits/rejected": -1.0550997257232666,
|
| 894 |
+
"logps/chosen": -340.00164794921875,
|
| 895 |
+
"logps/rejected": -341.64068603515625,
|
| 896 |
+
"loss": 0.0722,
|
| 897 |
+
"rewards/accuracies": 0.5625,
|
| 898 |
+
"rewards/chosen": 0.017406892031431198,
|
| 899 |
+
"rewards/margins": 0.03364910930395126,
|
| 900 |
+
"rewards/rejected": -0.016242217272520065,
|
| 901 |
+
"step": 570
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.37,
|
| 905 |
+
"learning_rate": 3.962671243342728e-06,
|
| 906 |
+
"logits/chosen": -1.089564323425293,
|
| 907 |
+
"logits/rejected": -1.038475513458252,
|
| 908 |
+
"logps/chosen": -314.32012939453125,
|
| 909 |
+
"logps/rejected": -327.6108703613281,
|
| 910 |
+
"loss": 0.0795,
|
| 911 |
+
"rewards/accuracies": 0.46875,
|
| 912 |
+
"rewards/chosen": 0.00651575718075037,
|
| 913 |
+
"rewards/margins": 0.018204618245363235,
|
| 914 |
+
"rewards/rejected": -0.011688861064612865,
|
| 915 |
+
"step": 580
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.38,
|
| 919 |
+
"learning_rate": 3.916975206127501e-06,
|
| 920 |
+
"logits/chosen": -1.0772429704666138,
|
| 921 |
+
"logits/rejected": -1.0009437799453735,
|
| 922 |
+
"logps/chosen": -317.83563232421875,
|
| 923 |
+
"logps/rejected": -302.42535400390625,
|
| 924 |
+
"loss": 0.0735,
|
| 925 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 926 |
+
"rewards/chosen": 0.006415791809558868,
|
| 927 |
+
"rewards/margins": 0.03005843423306942,
|
| 928 |
+
"rewards/rejected": -0.0236426442861557,
|
| 929 |
+
"step": 590
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.38,
|
| 933 |
+
"learning_rate": 3.870570748195039e-06,
|
| 934 |
+
"logits/chosen": -1.1022416353225708,
|
| 935 |
+
"logits/rejected": -1.1078603267669678,
|
| 936 |
+
"logps/chosen": -338.0223083496094,
|
| 937 |
+
"logps/rejected": -388.9841003417969,
|
| 938 |
+
"loss": 0.0742,
|
| 939 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 940 |
+
"rewards/chosen": 0.012317690066993237,
|
| 941 |
+
"rewards/margins": 0.049174439162015915,
|
| 942 |
+
"rewards/rejected": -0.036856748163700104,
|
| 943 |
+
"step": 600
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.38,
|
| 947 |
+
"eval_logits/chosen": -1.0245832204818726,
|
| 948 |
+
"eval_logits/rejected": -0.9872007966041565,
|
| 949 |
+
"eval_logps/chosen": -382.8363342285156,
|
| 950 |
+
"eval_logps/rejected": -384.7105407714844,
|
| 951 |
+
"eval_loss": 0.04055745154619217,
|
| 952 |
+
"eval_rewards/accuracies": 0.6085000038146973,
|
| 953 |
+
"eval_rewards/chosen": 0.013468942604959011,
|
| 954 |
+
"eval_rewards/margins": 0.030944030731916428,
|
| 955 |
+
"eval_rewards/rejected": -0.017475087195634842,
|
| 956 |
+
"eval_runtime": 539.6208,
|
| 957 |
+
"eval_samples_per_second": 3.706,
|
| 958 |
+
"eval_steps_per_second": 0.927,
|
| 959 |
+
"step": 600
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.39,
|
| 963 |
+
"learning_rate": 3.823481069583869e-06,
|
| 964 |
+
"logits/chosen": -1.0397017002105713,
|
| 965 |
+
"logits/rejected": -0.9971572756767273,
|
| 966 |
+
"logps/chosen": -305.09442138671875,
|
| 967 |
+
"logps/rejected": -314.68621826171875,
|
| 968 |
+
"loss": 0.08,
|
| 969 |
+
"rewards/accuracies": 0.5,
|
| 970 |
+
"rewards/chosen": 0.01682308129966259,
|
| 971 |
+
"rewards/margins": 0.026519659906625748,
|
| 972 |
+
"rewards/rejected": -0.009696578606963158,
|
| 973 |
+
"step": 610
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.4,
|
| 977 |
+
"learning_rate": 3.7757297129105087e-06,
|
| 978 |
+
"logits/chosen": -1.1045721769332886,
|
| 979 |
+
"logits/rejected": -1.0196996927261353,
|
| 980 |
+
"logps/chosen": -272.93206787109375,
|
| 981 |
+
"logps/rejected": -281.6795349121094,
|
| 982 |
+
"loss": 0.0878,
|
| 983 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 984 |
+
"rewards/chosen": 0.009568464942276478,
|
| 985 |
+
"rewards/margins": 0.01820019818842411,
|
| 986 |
+
"rewards/rejected": -0.008631732314825058,
|
| 987 |
+
"step": 620
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.4,
|
| 991 |
+
"learning_rate": 3.7273405515992785e-06,
|
| 992 |
+
"logits/chosen": -1.0682175159454346,
|
| 993 |
+
"logits/rejected": -1.0213497877120972,
|
| 994 |
+
"logps/chosen": -284.39495849609375,
|
| 995 |
+
"logps/rejected": -300.61920166015625,
|
| 996 |
+
"loss": 0.0707,
|
| 997 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 998 |
+
"rewards/chosen": 0.01212714146822691,
|
| 999 |
+
"rewards/margins": 0.01712757535278797,
|
| 1000 |
+
"rewards/rejected": -0.005000432953238487,
|
| 1001 |
+
"step": 630
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 0.41,
|
| 1005 |
+
"learning_rate": 3.678337777946706e-06,
|
| 1006 |
+
"logits/chosen": -1.0374300479888916,
|
| 1007 |
+
"logits/rejected": -0.9567736387252808,
|
| 1008 |
+
"logps/chosen": -334.5086669921875,
|
| 1009 |
+
"logps/rejected": -323.7910461425781,
|
| 1010 |
+
"loss": 0.0941,
|
| 1011 |
+
"rewards/accuracies": 0.5625,
|
| 1012 |
+
"rewards/chosen": 0.02154025062918663,
|
| 1013 |
+
"rewards/margins": 0.038109757006168365,
|
| 1014 |
+
"rewards/rejected": -0.016569510102272034,
|
| 1015 |
+
"step": 640
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.42,
|
| 1019 |
+
"learning_rate": 3.6287458910265293e-06,
|
| 1020 |
+
"logits/chosen": -1.0035964250564575,
|
| 1021 |
+
"logits/rejected": -0.9258484840393066,
|
| 1022 |
+
"logps/chosen": -309.08038330078125,
|
| 1023 |
+
"logps/rejected": -294.2804260253906,
|
| 1024 |
+
"loss": 0.0892,
|
| 1025 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1026 |
+
"rewards/chosen": 0.011333728209137917,
|
| 1027 |
+
"rewards/margins": 0.01899079605937004,
|
| 1028 |
+
"rewards/rejected": -0.0076570697128772736,
|
| 1029 |
+
"step": 650
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.42,
|
| 1033 |
+
"learning_rate": 3.57858968444131e-06,
|
| 1034 |
+
"logits/chosen": -1.0853677988052368,
|
| 1035 |
+
"logits/rejected": -0.9975327253341675,
|
| 1036 |
+
"logps/chosen": -330.30255126953125,
|
| 1037 |
+
"logps/rejected": -312.9807434082031,
|
| 1038 |
+
"loss": 0.0737,
|
| 1039 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1040 |
+
"rewards/chosen": 0.014233666472136974,
|
| 1041 |
+
"rewards/margins": 0.025601008906960487,
|
| 1042 |
+
"rewards/rejected": -0.011367343366146088,
|
| 1043 |
+
"step": 660
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.43,
|
| 1047 |
+
"learning_rate": 3.5278942339268034e-06,
|
| 1048 |
+
"logits/chosen": -1.0091218948364258,
|
| 1049 |
+
"logits/rejected": -1.0755088329315186,
|
| 1050 |
+
"logps/chosen": -276.8922424316406,
|
| 1051 |
+
"logps/rejected": -313.3753967285156,
|
| 1052 |
+
"loss": 0.0777,
|
| 1053 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1054 |
+
"rewards/chosen": 0.016201717779040337,
|
| 1055 |
+
"rewards/margins": 0.032345883548259735,
|
| 1056 |
+
"rewards/rejected": -0.016144167631864548,
|
| 1057 |
+
"step": 670
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.44,
|
| 1061 |
+
"learning_rate": 3.476684884815279e-06,
|
| 1062 |
+
"logits/chosen": -1.0601623058319092,
|
| 1063 |
+
"logits/rejected": -0.9890631437301636,
|
| 1064 |
+
"logps/chosen": -280.58538818359375,
|
| 1065 |
+
"logps/rejected": -280.070068359375,
|
| 1066 |
+
"loss": 0.1082,
|
| 1067 |
+
"rewards/accuracies": 0.375,
|
| 1068 |
+
"rewards/chosen": 0.00014938972890377045,
|
| 1069 |
+
"rewards/margins": 0.00774354999884963,
|
| 1070 |
+
"rewards/rejected": -0.007594159804284573,
|
| 1071 |
+
"step": 680
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.44,
|
| 1075 |
+
"learning_rate": 3.424987239364044e-06,
|
| 1076 |
+
"logits/chosen": -1.0564385652542114,
|
| 1077 |
+
"logits/rejected": -1.0079165697097778,
|
| 1078 |
+
"logps/chosen": -290.0457458496094,
|
| 1079 |
+
"logps/rejected": -317.35809326171875,
|
| 1080 |
+
"loss": 0.0844,
|
| 1081 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1082 |
+
"rewards/chosen": 0.01850438304245472,
|
| 1083 |
+
"rewards/margins": 0.03815234825015068,
|
| 1084 |
+
"rewards/rejected": -0.01964796707034111,
|
| 1085 |
+
"step": 690
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.45,
|
| 1089 |
+
"learning_rate": 3.3728271439555277e-06,
|
| 1090 |
+
"logits/chosen": -1.061989426612854,
|
| 1091 |
+
"logits/rejected": -1.0436265468597412,
|
| 1092 |
+
"logps/chosen": -300.4573669433594,
|
| 1093 |
+
"logps/rejected": -291.84320068359375,
|
| 1094 |
+
"loss": 0.0635,
|
| 1095 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1096 |
+
"rewards/chosen": 0.012897541746497154,
|
| 1097 |
+
"rewards/margins": 0.022769348695874214,
|
| 1098 |
+
"rewards/rejected": -0.009871806018054485,
|
| 1099 |
+
"step": 700
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.45,
|
| 1103 |
+
"eval_logits/chosen": -1.022524118423462,
|
| 1104 |
+
"eval_logits/rejected": -0.9903200268745422,
|
| 1105 |
+
"eval_logps/chosen": -379.6695556640625,
|
| 1106 |
+
"eval_logps/rejected": -386.0258483886719,
|
| 1107 |
+
"eval_loss": 0.04010883718729019,
|
| 1108 |
+
"eval_rewards/accuracies": 0.609499990940094,
|
| 1109 |
+
"eval_rewards/chosen": 0.01663573645055294,
|
| 1110 |
+
"eval_rewards/margins": 0.035426173359155655,
|
| 1111 |
+
"eval_rewards/rejected": -0.018790436908602715,
|
| 1112 |
+
"eval_runtime": 539.7506,
|
| 1113 |
+
"eval_samples_per_second": 3.705,
|
| 1114 |
+
"eval_steps_per_second": 0.926,
|
| 1115 |
+
"step": 700
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.45,
|
| 1119 |
+
"learning_rate": 3.3202306761753078e-06,
|
| 1120 |
+
"logits/chosen": -1.056015133857727,
|
| 1121 |
+
"logits/rejected": -1.044135570526123,
|
| 1122 |
+
"logps/chosen": -291.650634765625,
|
| 1123 |
+
"logps/rejected": -287.6162414550781,
|
| 1124 |
+
"loss": 0.0908,
|
| 1125 |
+
"rewards/accuracies": 0.46875,
|
| 1126 |
+
"rewards/chosen": 0.012986931018531322,
|
| 1127 |
+
"rewards/margins": 0.0346546433866024,
|
| 1128 |
+
"rewards/rejected": -0.021667715162038803,
|
| 1129 |
+
"step": 710
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.46,
|
| 1133 |
+
"learning_rate": 3.2672241317745513e-06,
|
| 1134 |
+
"logits/chosen": -0.9736151695251465,
|
| 1135 |
+
"logits/rejected": -0.9646120071411133,
|
| 1136 |
+
"logps/chosen": -311.49346923828125,
|
| 1137 |
+
"logps/rejected": -341.27496337890625,
|
| 1138 |
+
"loss": 0.0756,
|
| 1139 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 1140 |
+
"rewards/chosen": 0.017372317612171173,
|
| 1141 |
+
"rewards/margins": 0.028918754309415817,
|
| 1142 |
+
"rewards/rejected": -0.011546434834599495,
|
| 1143 |
+
"step": 720
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.47,
|
| 1147 |
+
"learning_rate": 3.213834011523378e-06,
|
| 1148 |
+
"logits/chosen": -1.1018531322479248,
|
| 1149 |
+
"logits/rejected": -1.025577187538147,
|
| 1150 |
+
"logps/chosen": -319.1675720214844,
|
| 1151 |
+
"logps/rejected": -316.7040100097656,
|
| 1152 |
+
"loss": 0.0765,
|
| 1153 |
+
"rewards/accuracies": 0.543749988079071,
|
| 1154 |
+
"rewards/chosen": 0.014244127087295055,
|
| 1155 |
+
"rewards/margins": 0.03847852349281311,
|
| 1156 |
+
"rewards/rejected": -0.02423439547419548,
|
| 1157 |
+
"step": 730
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.47,
|
| 1161 |
+
"learning_rate": 3.160087007961724e-06,
|
| 1162 |
+
"logits/chosen": -1.048353910446167,
|
| 1163 |
+
"logits/rejected": -0.990998387336731,
|
| 1164 |
+
"logps/chosen": -310.2376403808594,
|
| 1165 |
+
"logps/rejected": -319.0958557128906,
|
| 1166 |
+
"loss": 0.0929,
|
| 1167 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 1168 |
+
"rewards/chosen": 0.01772814430296421,
|
| 1169 |
+
"rewards/margins": 0.029912665486335754,
|
| 1170 |
+
"rewards/rejected": -0.012184521183371544,
|
| 1171 |
+
"step": 740
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.48,
|
| 1175 |
+
"learning_rate": 3.1060099920543404e-06,
|
| 1176 |
+
"logits/chosen": -0.9734305143356323,
|
| 1177 |
+
"logits/rejected": -1.053117275238037,
|
| 1178 |
+
"logps/chosen": -253.1719207763672,
|
| 1179 |
+
"logps/rejected": -268.72723388671875,
|
| 1180 |
+
"loss": 0.1126,
|
| 1181 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1182 |
+
"rewards/chosen": 0.019246799871325493,
|
| 1183 |
+
"rewards/margins": 0.028338003903627396,
|
| 1184 |
+
"rewards/rejected": -0.009091204032301903,
|
| 1185 |
+
"step": 750
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.49,
|
| 1189 |
+
"learning_rate": 3.0516299997565675e-06,
|
| 1190 |
+
"logits/chosen": -0.9801149368286133,
|
| 1191 |
+
"logits/rejected": -1.053255558013916,
|
| 1192 |
+
"logps/chosen": -300.00152587890625,
|
| 1193 |
+
"logps/rejected": -316.54656982421875,
|
| 1194 |
+
"loss": 0.0736,
|
| 1195 |
+
"rewards/accuracies": 0.518750011920929,
|
| 1196 |
+
"rewards/chosen": 0.013104112818837166,
|
| 1197 |
+
"rewards/margins": 0.0340370312333107,
|
| 1198 |
+
"rewards/rejected": -0.020932912826538086,
|
| 1199 |
+
"step": 760
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.49,
|
| 1203 |
+
"learning_rate": 2.996974218497643e-06,
|
| 1204 |
+
"logits/chosen": -0.9665297269821167,
|
| 1205 |
+
"logits/rejected": -1.033067226409912,
|
| 1206 |
+
"logps/chosen": -327.8155822753906,
|
| 1207 |
+
"logps/rejected": -291.17840576171875,
|
| 1208 |
+
"loss": 0.0969,
|
| 1209 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 1210 |
+
"rewards/chosen": 0.013064196333289146,
|
| 1211 |
+
"rewards/margins": 0.014280739240348339,
|
| 1212 |
+
"rewards/rejected": -0.0012165403459221125,
|
| 1213 |
+
"step": 770
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.5,
|
| 1217 |
+
"learning_rate": 2.9420699735882673e-06,
|
| 1218 |
+
"logits/chosen": -1.103208065032959,
|
| 1219 |
+
"logits/rejected": -1.0667420625686646,
|
| 1220 |
+
"logps/chosen": -284.13043212890625,
|
| 1221 |
+
"logps/rejected": -288.5975646972656,
|
| 1222 |
+
"loss": 0.0726,
|
| 1223 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1224 |
+
"rewards/chosen": 0.021614065393805504,
|
| 1225 |
+
"rewards/margins": 0.022522414103150368,
|
| 1226 |
+
"rewards/rejected": -0.000908347952645272,
|
| 1227 |
+
"step": 780
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.51,
|
| 1231 |
+
"learning_rate": 2.8869447145592345e-06,
|
| 1232 |
+
"logits/chosen": -0.9419771432876587,
|
| 1233 |
+
"logits/rejected": -0.9063804745674133,
|
| 1234 |
+
"logps/chosen": -275.3219299316406,
|
| 1235 |
+
"logps/rejected": -281.0148620605469,
|
| 1236 |
+
"loss": 0.0885,
|
| 1237 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 1238 |
+
"rewards/chosen": 0.021688418462872505,
|
| 1239 |
+
"rewards/margins": 0.02460251934826374,
|
| 1240 |
+
"rewards/rejected": -0.0029140994884073734,
|
| 1241 |
+
"step": 790
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.51,
|
| 1245 |
+
"learning_rate": 2.831626001437969e-06,
|
| 1246 |
+
"logits/chosen": -0.9940840601921082,
|
| 1247 |
+
"logits/rejected": -0.9761594533920288,
|
| 1248 |
+
"logps/chosen": -284.7848205566406,
|
| 1249 |
+
"logps/rejected": -305.55279541015625,
|
| 1250 |
+
"loss": 0.0881,
|
| 1251 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1252 |
+
"rewards/chosen": 0.02503793314099312,
|
| 1253 |
+
"rewards/margins": 0.03414962440729141,
|
| 1254 |
+
"rewards/rejected": -0.009111693128943443,
|
| 1255 |
+
"step": 800
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.51,
|
| 1259 |
+
"eval_logits/chosen": -0.9974508881568909,
|
| 1260 |
+
"eval_logits/rejected": -0.9657922387123108,
|
| 1261 |
+
"eval_logps/chosen": -371.2671813964844,
|
| 1262 |
+
"eval_logps/rejected": -377.4323425292969,
|
| 1263 |
+
"eval_loss": 0.03950659930706024,
|
| 1264 |
+
"eval_rewards/accuracies": 0.6085000038146973,
|
| 1265 |
+
"eval_rewards/chosen": 0.025038093328475952,
|
| 1266 |
+
"eval_rewards/margins": 0.03523498401045799,
|
| 1267 |
+
"eval_rewards/rejected": -0.010196887888014317,
|
| 1268 |
+
"eval_runtime": 539.4827,
|
| 1269 |
+
"eval_samples_per_second": 3.707,
|
| 1270 |
+
"eval_steps_per_second": 0.927,
|
| 1271 |
+
"step": 800
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 0.52,
|
| 1275 |
+
"learning_rate": 2.776141490969808e-06,
|
| 1276 |
+
"logits/chosen": -0.9662303924560547,
|
| 1277 |
+
"logits/rejected": -1.007033109664917,
|
| 1278 |
+
"logps/chosen": -273.5274963378906,
|
| 1279 |
+
"logps/rejected": -297.3504943847656,
|
| 1280 |
+
"loss": 0.0811,
|
| 1281 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 1282 |
+
"rewards/chosen": 0.017237264662981033,
|
| 1283 |
+
"rewards/margins": 0.020813334733247757,
|
| 1284 |
+
"rewards/rejected": -0.0035760707687586546,
|
| 1285 |
+
"step": 810
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.52,
|
| 1289 |
+
"learning_rate": 2.720518922790937e-06,
|
| 1290 |
+
"logits/chosen": -1.0530205965042114,
|
| 1291 |
+
"logits/rejected": -0.9808292388916016,
|
| 1292 |
+
"logps/chosen": -260.7068176269531,
|
| 1293 |
+
"logps/rejected": -254.8417205810547,
|
| 1294 |
+
"loss": 0.1111,
|
| 1295 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1296 |
+
"rewards/chosen": 0.015834391117095947,
|
| 1297 |
+
"rewards/margins": 0.030566086992621422,
|
| 1298 |
+
"rewards/rejected": -0.014731695875525475,
|
| 1299 |
+
"step": 820
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.53,
|
| 1303 |
+
"learning_rate": 2.66478610555988e-06,
|
| 1304 |
+
"logits/chosen": -0.973135769367218,
|
| 1305 |
+
"logits/rejected": -1.0151408910751343,
|
| 1306 |
+
"logps/chosen": -315.51763916015625,
|
| 1307 |
+
"logps/rejected": -324.4259338378906,
|
| 1308 |
+
"loss": 0.0997,
|
| 1309 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 1310 |
+
"rewards/chosen": 0.022444238886237144,
|
| 1311 |
+
"rewards/margins": 0.034130193293094635,
|
| 1312 |
+
"rewards/rejected": -0.011685955338180065,
|
| 1313 |
+
"step": 830
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.54,
|
| 1317 |
+
"learning_rate": 2.608970903054482e-06,
|
| 1318 |
+
"logits/chosen": -0.9921044111251831,
|
| 1319 |
+
"logits/rejected": -0.996437668800354,
|
| 1320 |
+
"logps/chosen": -298.2916564941406,
|
| 1321 |
+
"logps/rejected": -307.67156982421875,
|
| 1322 |
+
"loss": 0.0765,
|
| 1323 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1324 |
+
"rewards/chosen": 0.020961161702871323,
|
| 1325 |
+
"rewards/margins": 0.020338475704193115,
|
| 1326 |
+
"rewards/rejected": 0.0006226839614100754,
|
| 1327 |
+
"step": 840
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.54,
|
| 1331 |
+
"learning_rate": 2.553101220241337e-06,
|
| 1332 |
+
"logits/chosen": -0.955792248249054,
|
| 1333 |
+
"logits/rejected": -0.9853283166885376,
|
| 1334 |
+
"logps/chosen": -322.03619384765625,
|
| 1335 |
+
"logps/rejected": -341.2371520996094,
|
| 1336 |
+
"loss": 0.0739,
|
| 1337 |
+
"rewards/accuracies": 0.46875,
|
| 1338 |
+
"rewards/chosen": 0.02565811201930046,
|
| 1339 |
+
"rewards/margins": 0.01996433734893799,
|
| 1340 |
+
"rewards/rejected": 0.005693775601685047,
|
| 1341 |
+
"step": 850
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.55,
|
| 1345 |
+
"learning_rate": 2.4972049893246218e-06,
|
| 1346 |
+
"logits/chosen": -0.9531941413879395,
|
| 1347 |
+
"logits/rejected": -0.9866918325424194,
|
| 1348 |
+
"logps/chosen": -299.243896484375,
|
| 1349 |
+
"logps/rejected": -302.88336181640625,
|
| 1350 |
+
"loss": 0.0889,
|
| 1351 |
+
"rewards/accuracies": 0.5,
|
| 1352 |
+
"rewards/chosen": 0.02744477614760399,
|
| 1353 |
+
"rewards/margins": 0.03810085728764534,
|
| 1354 |
+
"rewards/rejected": -0.0106560830026865,
|
| 1355 |
+
"step": 860
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 0.56,
|
| 1359 |
+
"learning_rate": 2.4413101557813095e-06,
|
| 1360 |
+
"logits/chosen": -0.9942695498466492,
|
| 1361 |
+
"logits/rejected": -1.0122566223144531,
|
| 1362 |
+
"logps/chosen": -291.08892822265625,
|
| 1363 |
+
"logps/rejected": -301.3489074707031,
|
| 1364 |
+
"loss": 0.076,
|
| 1365 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 1366 |
+
"rewards/chosen": 0.027957703918218613,
|
| 1367 |
+
"rewards/margins": 0.030374949797987938,
|
| 1368 |
+
"rewards/rejected": -0.0024172496050596237,
|
| 1369 |
+
"step": 870
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 0.56,
|
| 1373 |
+
"learning_rate": 2.3854446643897566e-06,
|
| 1374 |
+
"logits/chosen": -0.935819149017334,
|
| 1375 |
+
"logits/rejected": -0.8908926248550415,
|
| 1376 |
+
"logps/chosen": -274.655517578125,
|
| 1377 |
+
"logps/rejected": -301.8233642578125,
|
| 1378 |
+
"loss": 0.093,
|
| 1379 |
+
"rewards/accuracies": 0.5,
|
| 1380 |
+
"rewards/chosen": 0.02359806001186371,
|
| 1381 |
+
"rewards/margins": 0.040579214692115784,
|
| 1382 |
+
"rewards/rejected": -0.016981154680252075,
|
| 1383 |
+
"step": 880
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 0.57,
|
| 1387 |
+
"learning_rate": 2.3296364452586246e-06,
|
| 1388 |
+
"logits/chosen": -0.9394947290420532,
|
| 1389 |
+
"logits/rejected": -1.015801191329956,
|
| 1390 |
+
"logps/chosen": -244.6657257080078,
|
| 1391 |
+
"logps/rejected": -255.62155151367188,
|
| 1392 |
+
"loss": 0.0973,
|
| 1393 |
+
"rewards/accuracies": 0.40625,
|
| 1394 |
+
"rewards/chosen": 0.02046729251742363,
|
| 1395 |
+
"rewards/margins": 0.011587701737880707,
|
| 1396 |
+
"rewards/rejected": 0.008879591710865498,
|
| 1397 |
+
"step": 890
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 0.58,
|
| 1401 |
+
"learning_rate": 2.273913399863151e-06,
|
| 1402 |
+
"logits/chosen": -0.9255924224853516,
|
| 1403 |
+
"logits/rejected": -0.9949308633804321,
|
| 1404 |
+
"logps/chosen": -319.0769958496094,
|
| 1405 |
+
"logps/rejected": -335.22125244140625,
|
| 1406 |
+
"loss": 0.0753,
|
| 1407 |
+
"rewards/accuracies": 0.5,
|
| 1408 |
+
"rewards/chosen": 0.0340295247733593,
|
| 1409 |
+
"rewards/margins": 0.03226994723081589,
|
| 1410 |
+
"rewards/rejected": 0.0017595753306522965,
|
| 1411 |
+
"step": 900
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 0.58,
|
| 1415 |
+
"eval_logits/chosen": -0.9455928802490234,
|
| 1416 |
+
"eval_logits/rejected": -0.9026016592979431,
|
| 1417 |
+
"eval_logps/chosen": -365.869873046875,
|
| 1418 |
+
"eval_logps/rejected": -371.78717041015625,
|
| 1419 |
+
"eval_loss": 0.03934764117002487,
|
| 1420 |
+
"eval_rewards/accuracies": 0.5989999771118164,
|
| 1421 |
+
"eval_rewards/chosen": 0.0304353516548872,
|
| 1422 |
+
"eval_rewards/margins": 0.03498707711696625,
|
| 1423 |
+
"eval_rewards/rejected": -0.00455172173678875,
|
| 1424 |
+
"eval_runtime": 539.932,
|
| 1425 |
+
"eval_samples_per_second": 3.704,
|
| 1426 |
+
"eval_steps_per_second": 0.926,
|
| 1427 |
+
"step": 900
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.58,
|
| 1431 |
+
"learning_rate": 2.2183033870957237e-06,
|
| 1432 |
+
"logits/chosen": -0.9681524038314819,
|
| 1433 |
+
"logits/rejected": -0.9325584173202515,
|
| 1434 |
+
"logps/chosen": -338.40325927734375,
|
| 1435 |
+
"logps/rejected": -328.984619140625,
|
| 1436 |
+
"loss": 0.0758,
|
| 1437 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1438 |
+
"rewards/chosen": 0.017296601086854935,
|
| 1439 |
+
"rewards/margins": 0.03395865857601166,
|
| 1440 |
+
"rewards/rejected": -0.016662055626511574,
|
| 1441 |
+
"step": 910
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.59,
|
| 1445 |
+
"learning_rate": 2.1628342093377533e-06,
|
| 1446 |
+
"logits/chosen": -0.9232437014579773,
|
| 1447 |
+
"logits/rejected": -0.9427019357681274,
|
| 1448 |
+
"logps/chosen": -301.0130615234375,
|
| 1449 |
+
"logps/rejected": -300.59088134765625,
|
| 1450 |
+
"loss": 0.0675,
|
| 1451 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1452 |
+
"rewards/chosen": 0.03031068481504917,
|
| 1453 |
+
"rewards/margins": 0.030671527609229088,
|
| 1454 |
+
"rewards/rejected": -0.00036084355087950826,
|
| 1455 |
+
"step": 920
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 0.6,
|
| 1459 |
+
"learning_rate": 2.1075335985597954e-06,
|
| 1460 |
+
"logits/chosen": -0.9461824297904968,
|
| 1461 |
+
"logits/rejected": -0.9701375961303711,
|
| 1462 |
+
"logps/chosen": -309.2986755371094,
|
| 1463 |
+
"logps/rejected": -296.10052490234375,
|
| 1464 |
+
"loss": 0.0781,
|
| 1465 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1466 |
+
"rewards/chosen": 0.02147866040468216,
|
| 1467 |
+
"rewards/margins": 0.023677725344896317,
|
| 1468 |
+
"rewards/rejected": -0.002199061680585146,
|
| 1469 |
+
"step": 930
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 0.6,
|
| 1473 |
+
"learning_rate": 2.0524292024568687e-06,
|
| 1474 |
+
"logits/chosen": -0.8931276202201843,
|
| 1475 |
+
"logits/rejected": -0.9411805272102356,
|
| 1476 |
+
"logps/chosen": -305.31793212890625,
|
| 1477 |
+
"logps/rejected": -297.4659729003906,
|
| 1478 |
+
"loss": 0.076,
|
| 1479 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 1480 |
+
"rewards/chosen": 0.024056371301412582,
|
| 1481 |
+
"rewards/margins": 0.024243740364909172,
|
| 1482 |
+
"rewards/rejected": -0.00018737166828941554,
|
| 1483 |
+
"step": 940
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 0.61,
|
| 1487 |
+
"learning_rate": 1.9975485706259194e-06,
|
| 1488 |
+
"logits/chosen": -0.9346133470535278,
|
| 1489 |
+
"logits/rejected": -0.9174652099609375,
|
| 1490 |
+
"logps/chosen": -243.5562744140625,
|
| 1491 |
+
"logps/rejected": -260.65185546875,
|
| 1492 |
+
"loss": 0.0812,
|
| 1493 |
+
"rewards/accuracies": 0.5,
|
| 1494 |
+
"rewards/chosen": 0.036083243787288666,
|
| 1495 |
+
"rewards/margins": 0.028012529015541077,
|
| 1496 |
+
"rewards/rejected": 0.00807071290910244,
|
| 1497 |
+
"step": 950
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 0.61,
|
| 1501 |
+
"learning_rate": 1.942919140792319e-06,
|
| 1502 |
+
"logits/chosen": -0.9425550699234009,
|
| 1503 |
+
"logits/rejected": -0.8944212198257446,
|
| 1504 |
+
"logps/chosen": -263.16070556640625,
|
| 1505 |
+
"logps/rejected": -296.932861328125,
|
| 1506 |
+
"loss": 0.0774,
|
| 1507 |
+
"rewards/accuracies": 0.53125,
|
| 1508 |
+
"rewards/chosen": 0.03240296244621277,
|
| 1509 |
+
"rewards/margins": 0.03853650763630867,
|
| 1510 |
+
"rewards/rejected": -0.006133544258773327,
|
| 1511 |
+
"step": 960
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 0.62,
|
| 1515 |
+
"learning_rate": 1.888568225092296e-06,
|
| 1516 |
+
"logits/chosen": -0.9160875082015991,
|
| 1517 |
+
"logits/rejected": -0.9029878377914429,
|
| 1518 |
+
"logps/chosen": -259.0009460449219,
|
| 1519 |
+
"logps/rejected": -284.0046081542969,
|
| 1520 |
+
"loss": 0.071,
|
| 1521 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1522 |
+
"rewards/chosen": 0.031025772914290428,
|
| 1523 |
+
"rewards/margins": 0.026529574766755104,
|
| 1524 |
+
"rewards/rejected": 0.004496193490922451,
|
| 1525 |
+
"step": 970
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 0.63,
|
| 1529 |
+
"learning_rate": 1.8345229964181628e-06,
|
| 1530 |
+
"logits/chosen": -1.0157150030136108,
|
| 1531 |
+
"logits/rejected": -0.9280340075492859,
|
| 1532 |
+
"logps/chosen": -280.16455078125,
|
| 1533 |
+
"logps/rejected": -297.15655517578125,
|
| 1534 |
+
"loss": 0.0771,
|
| 1535 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 1536 |
+
"rewards/chosen": 0.02662072703242302,
|
| 1537 |
+
"rewards/margins": 0.036351434886455536,
|
| 1538 |
+
"rewards/rejected": -0.009730703197419643,
|
| 1539 |
+
"step": 980
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 0.63,
|
| 1543 |
+
"learning_rate": 1.7808104748331459e-06,
|
| 1544 |
+
"logits/chosen": -0.9648904800415039,
|
| 1545 |
+
"logits/rejected": -0.9164671897888184,
|
| 1546 |
+
"logps/chosen": -303.514892578125,
|
| 1547 |
+
"logps/rejected": -292.10748291015625,
|
| 1548 |
+
"loss": 0.0935,
|
| 1549 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1550 |
+
"rewards/chosen": 0.025951769202947617,
|
| 1551 |
+
"rewards/margins": 0.030179208144545555,
|
| 1552 |
+
"rewards/rejected": -0.004227438475936651,
|
| 1553 |
+
"step": 990
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 0.64,
|
| 1557 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1558 |
+
"logits/chosen": -1.048002004623413,
|
| 1559 |
+
"logits/rejected": -0.9596036672592163,
|
| 1560 |
+
"logps/chosen": -297.2816467285156,
|
| 1561 |
+
"logps/rejected": -316.36981201171875,
|
| 1562 |
+
"loss": 0.0922,
|
| 1563 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 1564 |
+
"rewards/chosen": 0.01601681485772133,
|
| 1565 |
+
"rewards/margins": 0.02387891337275505,
|
| 1566 |
+
"rewards/rejected": -0.007862097583711147,
|
| 1567 |
+
"step": 1000
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 0.64,
|
| 1571 |
+
"eval_logits/chosen": -0.9184301495552063,
|
| 1572 |
+
"eval_logits/rejected": -0.8800927996635437,
|
| 1573 |
+
"eval_logps/chosen": -367.73187255859375,
|
| 1574 |
+
"eval_logps/rejected": -374.7669372558594,
|
| 1575 |
+
"eval_loss": 0.03902236372232437,
|
| 1576 |
+
"eval_rewards/accuracies": 0.5989999771118164,
|
| 1577 |
+
"eval_rewards/chosen": 0.028573375195264816,
|
| 1578 |
+
"eval_rewards/margins": 0.03610490262508392,
|
| 1579 |
+
"eval_rewards/rejected": -0.007531528826802969,
|
| 1580 |
+
"eval_runtime": 539.7163,
|
| 1581 |
+
"eval_samples_per_second": 3.706,
|
| 1582 |
+
"eval_steps_per_second": 0.926,
|
| 1583 |
+
"step": 1000
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 0.65,
|
| 1587 |
+
"learning_rate": 1.6744907880685735e-06,
|
| 1588 |
+
"logits/chosen": -0.9194186925888062,
|
| 1589 |
+
"logits/rejected": -0.8924249410629272,
|
| 1590 |
+
"logps/chosen": -283.1637878417969,
|
| 1591 |
+
"logps/rejected": -281.9581604003906,
|
| 1592 |
+
"loss": 0.0699,
|
| 1593 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 1594 |
+
"rewards/chosen": 0.02727659046649933,
|
| 1595 |
+
"rewards/margins": 0.020740380510687828,
|
| 1596 |
+
"rewards/rejected": 0.006536208093166351,
|
| 1597 |
+
"step": 1010
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 0.65,
|
| 1601 |
+
"learning_rate": 1.6219367777137652e-06,
|
| 1602 |
+
"logits/chosen": -0.9643144607543945,
|
| 1603 |
+
"logits/rejected": -0.9084329605102539,
|
| 1604 |
+
"logps/chosen": -296.56732177734375,
|
| 1605 |
+
"logps/rejected": -299.80718994140625,
|
| 1606 |
+
"loss": 0.0756,
|
| 1607 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 1608 |
+
"rewards/chosen": 0.02761662006378174,
|
| 1609 |
+
"rewards/margins": 0.022495564073324203,
|
| 1610 |
+
"rewards/rejected": 0.00512105505913496,
|
| 1611 |
+
"step": 1020
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.66,
|
| 1615 |
+
"learning_rate": 1.569821757522666e-06,
|
| 1616 |
+
"logits/chosen": -0.9250129461288452,
|
| 1617 |
+
"logits/rejected": -0.8530179262161255,
|
| 1618 |
+
"logps/chosen": -305.6438903808594,
|
| 1619 |
+
"logps/rejected": -310.8564758300781,
|
| 1620 |
+
"loss": 0.0796,
|
| 1621 |
+
"rewards/accuracies": 0.518750011920929,
|
| 1622 |
+
"rewards/chosen": 0.023295477032661438,
|
| 1623 |
+
"rewards/margins": 0.03615058213472366,
|
| 1624 |
+
"rewards/rejected": -0.012855103239417076,
|
| 1625 |
+
"step": 1030
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 0.67,
|
| 1629 |
+
"learning_rate": 1.5181717825453732e-06,
|
| 1630 |
+
"logits/chosen": -0.9920689463615417,
|
| 1631 |
+
"logits/rejected": -0.9287697076797485,
|
| 1632 |
+
"logps/chosen": -268.88922119140625,
|
| 1633 |
+
"logps/rejected": -297.29791259765625,
|
| 1634 |
+
"loss": 0.0907,
|
| 1635 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1636 |
+
"rewards/chosen": 0.025286570191383362,
|
| 1637 |
+
"rewards/margins": 0.03552253916859627,
|
| 1638 |
+
"rewards/rejected": -0.010235967114567757,
|
| 1639 |
+
"step": 1040
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 0.67,
|
| 1643 |
+
"learning_rate": 1.4670126753313286e-06,
|
| 1644 |
+
"logits/chosen": -0.9325569272041321,
|
| 1645 |
+
"logits/rejected": -0.9226212501525879,
|
| 1646 |
+
"logps/chosen": -273.0399169921875,
|
| 1647 |
+
"logps/rejected": -300.64337158203125,
|
| 1648 |
+
"loss": 0.0829,
|
| 1649 |
+
"rewards/accuracies": 0.5,
|
| 1650 |
+
"rewards/chosen": 0.02183723822236061,
|
| 1651 |
+
"rewards/margins": 0.02524760365486145,
|
| 1652 |
+
"rewards/rejected": -0.0034103658981621265,
|
| 1653 |
+
"step": 1050
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.68,
|
| 1657 |
+
"learning_rate": 1.4163700130192627e-06,
|
| 1658 |
+
"logits/chosen": -0.9867246747016907,
|
| 1659 |
+
"logits/rejected": -0.9270626306533813,
|
| 1660 |
+
"logps/chosen": -240.3236846923828,
|
| 1661 |
+
"logps/rejected": -277.5113525390625,
|
| 1662 |
+
"loss": 0.1007,
|
| 1663 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 1664 |
+
"rewards/chosen": 0.0198238343000412,
|
| 1665 |
+
"rewards/margins": 0.027476917952299118,
|
| 1666 |
+
"rewards/rejected": -0.007653082255274057,
|
| 1667 |
+
"step": 1060
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 0.68,
|
| 1671 |
+
"learning_rate": 1.366269114549833e-06,
|
| 1672 |
+
"logits/chosen": -0.9571270942687988,
|
| 1673 |
+
"logits/rejected": -0.9630721807479858,
|
| 1674 |
+
"logps/chosen": -282.31561279296875,
|
| 1675 |
+
"logps/rejected": -273.7547607421875,
|
| 1676 |
+
"loss": 0.0832,
|
| 1677 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 1678 |
+
"rewards/chosen": 0.014912809245288372,
|
| 1679 |
+
"rewards/margins": 0.02561432123184204,
|
| 1680 |
+
"rewards/rejected": -0.01070151012390852,
|
| 1681 |
+
"step": 1070
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 0.69,
|
| 1685 |
+
"learning_rate": 1.3167350280073514e-06,
|
| 1686 |
+
"logits/chosen": -0.9449627995491028,
|
| 1687 |
+
"logits/rejected": -0.9669274091720581,
|
| 1688 |
+
"logps/chosen": -290.40899658203125,
|
| 1689 |
+
"logps/rejected": -305.1846923828125,
|
| 1690 |
+
"loss": 0.0881,
|
| 1691 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1692 |
+
"rewards/chosen": 0.020313743501901627,
|
| 1693 |
+
"rewards/margins": 0.03765721619129181,
|
| 1694 |
+
"rewards/rejected": -0.017343472689390182,
|
| 1695 |
+
"step": 1080
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 0.7,
|
| 1699 |
+
"learning_rate": 1.267792518096918e-06,
|
| 1700 |
+
"logits/chosen": -0.954971194267273,
|
| 1701 |
+
"logits/rejected": -1.0019423961639404,
|
| 1702 |
+
"logps/chosen": -314.25335693359375,
|
| 1703 |
+
"logps/rejected": -375.009521484375,
|
| 1704 |
+
"loss": 0.0652,
|
| 1705 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1706 |
+
"rewards/chosen": 0.023724760860204697,
|
| 1707 |
+
"rewards/margins": 0.04069076478481293,
|
| 1708 |
+
"rewards/rejected": -0.01696600392460823,
|
| 1709 |
+
"step": 1090
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 0.7,
|
| 1713 |
+
"learning_rate": 1.2194660537632423e-06,
|
| 1714 |
+
"logits/chosen": -0.9429014325141907,
|
| 1715 |
+
"logits/rejected": -0.9538652300834656,
|
| 1716 |
+
"logps/chosen": -279.4613342285156,
|
| 1717 |
+
"logps/rejected": -323.7087097167969,
|
| 1718 |
+
"loss": 0.0703,
|
| 1719 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1720 |
+
"rewards/chosen": 0.02060026489198208,
|
| 1721 |
+
"rewards/margins": 0.03607643395662308,
|
| 1722 |
+
"rewards/rejected": -0.015476171858608723,
|
| 1723 |
+
"step": 1100
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 0.7,
|
| 1727 |
+
"eval_logits/chosen": -0.9601659178733826,
|
| 1728 |
+
"eval_logits/rejected": -0.9299748539924622,
|
| 1729 |
+
"eval_logps/chosen": -373.62261962890625,
|
| 1730 |
+
"eval_logps/rejected": -383.30255126953125,
|
| 1731 |
+
"eval_loss": 0.03885647654533386,
|
| 1732 |
+
"eval_rewards/accuracies": 0.6000000238418579,
|
| 1733 |
+
"eval_rewards/chosen": 0.022682595998048782,
|
| 1734 |
+
"eval_rewards/margins": 0.03874973580241203,
|
| 1735 |
+
"eval_rewards/rejected": -0.016067136079072952,
|
| 1736 |
+
"eval_runtime": 539.6341,
|
| 1737 |
+
"eval_samples_per_second": 3.706,
|
| 1738 |
+
"eval_steps_per_second": 0.927,
|
| 1739 |
+
"step": 1100
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 0.71,
|
| 1743 |
+
"learning_rate": 1.1717797959573262e-06,
|
| 1744 |
+
"logits/chosen": -1.003103494644165,
|
| 1745 |
+
"logits/rejected": -0.9326059222221375,
|
| 1746 |
+
"logps/chosen": -297.8578186035156,
|
| 1747 |
+
"logps/rejected": -297.1626892089844,
|
| 1748 |
+
"loss": 0.0771,
|
| 1749 |
+
"rewards/accuracies": 0.5,
|
| 1750 |
+
"rewards/chosen": 0.017790749669075012,
|
| 1751 |
+
"rewards/margins": 0.035782478749752045,
|
| 1752 |
+
"rewards/rejected": -0.017991727218031883,
|
| 1753 |
+
"step": 1110
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 0.72,
|
| 1757 |
+
"learning_rate": 1.1247575855571251e-06,
|
| 1758 |
+
"logits/chosen": -0.9976641535758972,
|
| 1759 |
+
"logits/rejected": -0.9637433886528015,
|
| 1760 |
+
"logps/chosen": -306.8631896972656,
|
| 1761 |
+
"logps/rejected": -350.14324951171875,
|
| 1762 |
+
"loss": 0.081,
|
| 1763 |
+
"rewards/accuracies": 0.543749988079071,
|
| 1764 |
+
"rewards/chosen": 0.029013922438025475,
|
| 1765 |
+
"rewards/margins": 0.05382751673460007,
|
| 1766 |
+
"rewards/rejected": -0.02481359988451004,
|
| 1767 |
+
"step": 1120
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.72,
|
| 1771 |
+
"learning_rate": 1.07842293144824e-06,
|
| 1772 |
+
"logits/chosen": -0.9757798314094543,
|
| 1773 |
+
"logits/rejected": -0.9444389343261719,
|
| 1774 |
+
"logps/chosen": -312.77191162109375,
|
| 1775 |
+
"logps/rejected": -317.60406494140625,
|
| 1776 |
+
"loss": 0.0816,
|
| 1777 |
+
"rewards/accuracies": 0.5,
|
| 1778 |
+
"rewards/chosen": 0.015693334862589836,
|
| 1779 |
+
"rewards/margins": 0.029305079951882362,
|
| 1780 |
+
"rewards/rejected": -0.013611746951937675,
|
| 1781 |
+
"step": 1130
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 0.73,
|
| 1785 |
+
"learning_rate": 1.0327989987705781e-06,
|
| 1786 |
+
"logits/chosen": -1.0041489601135254,
|
| 1787 |
+
"logits/rejected": -1.006225824356079,
|
| 1788 |
+
"logps/chosen": -291.49542236328125,
|
| 1789 |
+
"logps/rejected": -287.52978515625,
|
| 1790 |
+
"loss": 0.0888,
|
| 1791 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 1792 |
+
"rewards/chosen": 0.01858539506793022,
|
| 1793 |
+
"rewards/margins": 0.029126202687621117,
|
| 1794 |
+
"rewards/rejected": -0.01054080855101347,
|
| 1795 |
+
"step": 1140
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 0.74,
|
| 1799 |
+
"learning_rate": 9.879085973368805e-07,
|
| 1800 |
+
"logits/chosen": -1.0146872997283936,
|
| 1801 |
+
"logits/rejected": -0.9577957987785339,
|
| 1802 |
+
"logps/chosen": -322.76556396484375,
|
| 1803 |
+
"logps/rejected": -336.08636474609375,
|
| 1804 |
+
"loss": 0.0665,
|
| 1805 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1806 |
+
"rewards/chosen": 0.023015085607767105,
|
| 1807 |
+
"rewards/margins": 0.05666361376643181,
|
| 1808 |
+
"rewards/rejected": -0.0336485281586647,
|
| 1809 |
+
"step": 1150
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.74,
|
| 1813 |
+
"learning_rate": 9.437741702288908e-07,
|
| 1814 |
+
"logits/chosen": -1.0026918649673462,
|
| 1815 |
+
"logits/rejected": -0.9933171272277832,
|
| 1816 |
+
"logps/chosen": -299.72308349609375,
|
| 1817 |
+
"logps/rejected": -311.6734924316406,
|
| 1818 |
+
"loss": 0.0794,
|
| 1819 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 1820 |
+
"rewards/chosen": 0.01335589587688446,
|
| 1821 |
+
"rewards/margins": 0.025558674708008766,
|
| 1822 |
+
"rewards/rejected": -0.012202778831124306,
|
| 1823 |
+
"step": 1160
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 0.75,
|
| 1827 |
+
"learning_rate": 9.004177825768751e-07,
|
| 1828 |
+
"logits/chosen": -1.0369510650634766,
|
| 1829 |
+
"logits/rejected": -0.9392277002334595,
|
| 1830 |
+
"logps/chosen": -267.6405334472656,
|
| 1831 |
+
"logps/rejected": -298.1479797363281,
|
| 1832 |
+
"loss": 0.0757,
|
| 1833 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1834 |
+
"rewards/chosen": 0.012007731013000011,
|
| 1835 |
+
"rewards/margins": 0.024681296199560165,
|
| 1836 |
+
"rewards/rejected": -0.012673566117882729,
|
| 1837 |
+
"step": 1170
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 0.76,
|
| 1841 |
+
"learning_rate": 8.578611105280987e-07,
|
| 1842 |
+
"logits/chosen": -1.0083134174346924,
|
| 1843 |
+
"logits/rejected": -0.992781937122345,
|
| 1844 |
+
"logps/chosen": -316.5964660644531,
|
| 1845 |
+
"logps/rejected": -365.96661376953125,
|
| 1846 |
+
"loss": 0.0643,
|
| 1847 |
+
"rewards/accuracies": 0.5625,
|
| 1848 |
+
"rewards/chosen": 0.02335866168141365,
|
| 1849 |
+
"rewards/margins": 0.050595641136169434,
|
| 1850 |
+
"rewards/rejected": -0.027236973866820335,
|
| 1851 |
+
"step": 1180
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.76,
|
| 1855 |
+
"learning_rate": 8.161254304097715e-07,
|
| 1856 |
+
"logits/chosen": -1.026871919631958,
|
| 1857 |
+
"logits/rejected": -0.9791032075881958,
|
| 1858 |
+
"logps/chosen": -277.67547607421875,
|
| 1859 |
+
"logps/rejected": -294.22369384765625,
|
| 1860 |
+
"loss": 0.0853,
|
| 1861 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 1862 |
+
"rewards/chosen": 0.02239377610385418,
|
| 1863 |
+
"rewards/margins": 0.026370327919721603,
|
| 1864 |
+
"rewards/rejected": -0.003976552281528711,
|
| 1865 |
+
"step": 1190
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 0.77,
|
| 1869 |
+
"learning_rate": 7.752316080918934e-07,
|
| 1870 |
+
"logits/chosen": -0.942740797996521,
|
| 1871 |
+
"logits/rejected": -0.9156352877616882,
|
| 1872 |
+
"logps/chosen": -290.0126953125,
|
| 1873 |
+
"logps/rejected": -303.8479919433594,
|
| 1874 |
+
"loss": 0.0746,
|
| 1875 |
+
"rewards/accuracies": 0.4375,
|
| 1876 |
+
"rewards/chosen": 0.021792907267808914,
|
| 1877 |
+
"rewards/margins": 0.02630285918712616,
|
| 1878 |
+
"rewards/rejected": -0.004509954713284969,
|
| 1879 |
+
"step": 1200
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 0.77,
|
| 1883 |
+
"eval_logits/chosen": -0.930611789226532,
|
| 1884 |
+
"eval_logits/rejected": -0.8943980932235718,
|
| 1885 |
+
"eval_logps/chosen": -373.7153015136719,
|
| 1886 |
+
"eval_logps/rejected": -385.1601257324219,
|
| 1887 |
+
"eval_loss": 0.03884938731789589,
|
| 1888 |
+
"eval_rewards/accuracies": 0.6050000190734863,
|
| 1889 |
+
"eval_rewards/chosen": 0.022589918226003647,
|
| 1890 |
+
"eval_rewards/margins": 0.04051463305950165,
|
| 1891 |
+
"eval_rewards/rejected": -0.017924712970852852,
|
| 1892 |
+
"eval_runtime": 539.5465,
|
| 1893 |
+
"eval_samples_per_second": 3.707,
|
| 1894 |
+
"eval_steps_per_second": 0.927,
|
| 1895 |
+
"step": 1200
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.77,
|
| 1899 |
+
"learning_rate": 7.352000885553012e-07,
|
| 1900 |
+
"logits/chosen": -0.9893749356269836,
|
| 1901 |
+
"logits/rejected": -0.9994084239006042,
|
| 1902 |
+
"logps/chosen": -307.1074523925781,
|
| 1903 |
+
"logps/rejected": -315.70770263671875,
|
| 1904 |
+
"loss": 0.0767,
|
| 1905 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 1906 |
+
"rewards/chosen": 0.016450155526399612,
|
| 1907 |
+
"rewards/margins": 0.02466857247054577,
|
| 1908 |
+
"rewards/rejected": -0.008218420669436455,
|
| 1909 |
+
"step": 1210
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.78,
|
| 1913 |
+
"learning_rate": 6.960508856701464e-07,
|
| 1914 |
+
"logits/chosen": -1.0321322679519653,
|
| 1915 |
+
"logits/rejected": -0.9207181930541992,
|
| 1916 |
+
"logps/chosen": -291.80029296875,
|
| 1917 |
+
"logps/rejected": -298.3025817871094,
|
| 1918 |
+
"loss": 0.0945,
|
| 1919 |
+
"rewards/accuracies": 0.46875,
|
| 1920 |
+
"rewards/chosen": 0.015555900521576405,
|
| 1921 |
+
"rewards/margins": 0.02910393476486206,
|
| 1922 |
+
"rewards/rejected": -0.013548034243285656,
|
| 1923 |
+
"step": 1220
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.79,
|
| 1927 |
+
"learning_rate": 6.578035721899029e-07,
|
| 1928 |
+
"logits/chosen": -0.9535512924194336,
|
| 1929 |
+
"logits/rejected": -1.008029580116272,
|
| 1930 |
+
"logps/chosen": -324.25152587890625,
|
| 1931 |
+
"logps/rejected": -346.2152099609375,
|
| 1932 |
+
"loss": 0.0693,
|
| 1933 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 1934 |
+
"rewards/chosen": 0.012142010033130646,
|
| 1935 |
+
"rewards/margins": 0.03695748746395111,
|
| 1936 |
+
"rewards/rejected": -0.024815475568175316,
|
| 1937 |
+
"step": 1230
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.79,
|
| 1941 |
+
"learning_rate": 6.204772699659126e-07,
|
| 1942 |
+
"logits/chosen": -1.0189663171768188,
|
| 1943 |
+
"logits/rejected": -1.051673173904419,
|
| 1944 |
+
"logps/chosen": -268.8255920410156,
|
| 1945 |
+
"logps/rejected": -264.9363708496094,
|
| 1946 |
+
"loss": 0.0811,
|
| 1947 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1948 |
+
"rewards/chosen": 0.022226419299840927,
|
| 1949 |
+
"rewards/margins": 0.02963913045823574,
|
| 1950 |
+
"rewards/rejected": -0.007412709295749664,
|
| 1951 |
+
"step": 1240
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.8,
|
| 1955 |
+
"learning_rate": 5.840906403873648e-07,
|
| 1956 |
+
"logits/chosen": -0.9550544023513794,
|
| 1957 |
+
"logits/rejected": -0.9686701893806458,
|
| 1958 |
+
"logps/chosen": -311.6850280761719,
|
| 1959 |
+
"logps/rejected": -331.0824279785156,
|
| 1960 |
+
"loss": 0.0602,
|
| 1961 |
+
"rewards/accuracies": 0.518750011920929,
|
| 1962 |
+
"rewards/chosen": 0.01663174293935299,
|
| 1963 |
+
"rewards/margins": 0.03210796043276787,
|
| 1964 |
+
"rewards/rejected": -0.015476214699447155,
|
| 1965 |
+
"step": 1250
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.81,
|
| 1969 |
+
"learning_rate": 5.486618750514813e-07,
|
| 1970 |
+
"logits/chosen": -0.9893242716789246,
|
| 1971 |
+
"logits/rejected": -0.9823307991027832,
|
| 1972 |
+
"logps/chosen": -319.13421630859375,
|
| 1973 |
+
"logps/rejected": -335.4979553222656,
|
| 1974 |
+
"loss": 0.0904,
|
| 1975 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 1976 |
+
"rewards/chosen": 0.019062984734773636,
|
| 1977 |
+
"rewards/margins": 0.03206023946404457,
|
| 1978 |
+
"rewards/rejected": -0.012997254729270935,
|
| 1979 |
+
"step": 1260
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.81,
|
| 1983 |
+
"learning_rate": 5.142086866685783e-07,
|
| 1984 |
+
"logits/chosen": -0.9724270701408386,
|
| 1985 |
+
"logits/rejected": -0.9491437673568726,
|
| 1986 |
+
"logps/chosen": -279.63824462890625,
|
| 1987 |
+
"logps/rejected": -299.930908203125,
|
| 1988 |
+
"loss": 0.0748,
|
| 1989 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 1990 |
+
"rewards/chosen": 0.01930154860019684,
|
| 1991 |
+
"rewards/margins": 0.022973302751779556,
|
| 1992 |
+
"rewards/rejected": -0.0036717529874294996,
|
| 1993 |
+
"step": 1270
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.82,
|
| 1997 |
+
"learning_rate": 4.807483002065439e-07,
|
| 1998 |
+
"logits/chosen": -0.9794095754623413,
|
| 1999 |
+
"logits/rejected": -0.9791723489761353,
|
| 2000 |
+
"logps/chosen": -263.2690124511719,
|
| 2001 |
+
"logps/rejected": -264.7826232910156,
|
| 2002 |
+
"loss": 0.0906,
|
| 2003 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 2004 |
+
"rewards/chosen": 0.020124919712543488,
|
| 2005 |
+
"rewards/margins": 0.028638970106840134,
|
| 2006 |
+
"rewards/rejected": -0.008514048531651497,
|
| 2007 |
+
"step": 1280
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.83,
|
| 2011 |
+
"learning_rate": 4.4829744427917153e-07,
|
| 2012 |
+
"logits/chosen": -1.0250556468963623,
|
| 2013 |
+
"logits/rejected": -0.9449254870414734,
|
| 2014 |
+
"logps/chosen": -262.35614013671875,
|
| 2015 |
+
"logps/rejected": -292.6833190917969,
|
| 2016 |
+
"loss": 0.0725,
|
| 2017 |
+
"rewards/accuracies": 0.543749988079071,
|
| 2018 |
+
"rewards/chosen": 0.029375359416007996,
|
| 2019 |
+
"rewards/margins": 0.04200378805398941,
|
| 2020 |
+
"rewards/rejected": -0.012628423981368542,
|
| 2021 |
+
"step": 1290
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.83,
|
| 2025 |
+
"learning_rate": 4.168723427826382e-07,
|
| 2026 |
+
"logits/chosen": -0.964804470539093,
|
| 2027 |
+
"logits/rejected": -0.9295471906661987,
|
| 2028 |
+
"logps/chosen": -272.17938232421875,
|
| 2029 |
+
"logps/rejected": -284.2236022949219,
|
| 2030 |
+
"loss": 0.0925,
|
| 2031 |
+
"rewards/accuracies": 0.46875,
|
| 2032 |
+
"rewards/chosen": 0.01671653613448143,
|
| 2033 |
+
"rewards/margins": 0.0240671094506979,
|
| 2034 |
+
"rewards/rejected": -0.007350574247539043,
|
| 2035 |
+
"step": 1300
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.83,
|
| 2039 |
+
"eval_logits/chosen": -0.949418306350708,
|
| 2040 |
+
"eval_logits/rejected": -0.9171380400657654,
|
| 2041 |
+
"eval_logps/chosen": -370.0339660644531,
|
| 2042 |
+
"eval_logps/rejected": -380.30718994140625,
|
| 2043 |
+
"eval_loss": 0.03874335065484047,
|
| 2044 |
+
"eval_rewards/accuracies": 0.6029999852180481,
|
| 2045 |
+
"eval_rewards/chosen": 0.02627129666507244,
|
| 2046 |
+
"eval_rewards/margins": 0.039343029260635376,
|
| 2047 |
+
"eval_rewards/rejected": -0.013071730732917786,
|
| 2048 |
+
"eval_runtime": 539.7909,
|
| 2049 |
+
"eval_samples_per_second": 3.705,
|
| 2050 |
+
"eval_steps_per_second": 0.926,
|
| 2051 |
+
"step": 1300
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 0.84,
|
| 2055 |
+
"learning_rate": 3.864887067843251e-07,
|
| 2056 |
+
"logits/chosen": -0.9890697598457336,
|
| 2057 |
+
"logits/rejected": -0.8923286199569702,
|
| 2058 |
+
"logps/chosen": -310.97125244140625,
|
| 2059 |
+
"logps/rejected": -309.845703125,
|
| 2060 |
+
"loss": 0.0789,
|
| 2061 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 2062 |
+
"rewards/chosen": 0.019753634929656982,
|
| 2063 |
+
"rewards/margins": 0.031969424337148666,
|
| 2064 |
+
"rewards/rejected": -0.01221578847616911,
|
| 2065 |
+
"step": 1310
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 0.84,
|
| 2069 |
+
"learning_rate": 3.5716172666802637e-07,
|
| 2070 |
+
"logits/chosen": -1.0090614557266235,
|
| 2071 |
+
"logits/rejected": -0.9583064317703247,
|
| 2072 |
+
"logps/chosen": -266.8111267089844,
|
| 2073 |
+
"logps/rejected": -302.0957946777344,
|
| 2074 |
+
"loss": 0.0764,
|
| 2075 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 2076 |
+
"rewards/chosen": 0.023147406056523323,
|
| 2077 |
+
"rewards/margins": 0.04169207811355591,
|
| 2078 |
+
"rewards/rejected": -0.018544670194387436,
|
| 2079 |
+
"step": 1320
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.85,
|
| 2083 |
+
"learning_rate": 3.289060645394704e-07,
|
| 2084 |
+
"logits/chosen": -0.9466385841369629,
|
| 2085 |
+
"logits/rejected": -0.9388877749443054,
|
| 2086 |
+
"logps/chosen": -291.6334533691406,
|
| 2087 |
+
"logps/rejected": -310.3324279785156,
|
| 2088 |
+
"loss": 0.094,
|
| 2089 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 2090 |
+
"rewards/chosen": 0.021957775577902794,
|
| 2091 |
+
"rewards/margins": 0.026331758126616478,
|
| 2092 |
+
"rewards/rejected": -0.0043739816173911095,
|
| 2093 |
+
"step": 1330
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 0.86,
|
| 2097 |
+
"learning_rate": 3.0173584689596246e-07,
|
| 2098 |
+
"logits/chosen": -0.9820948839187622,
|
| 2099 |
+
"logits/rejected": -0.9631234407424927,
|
| 2100 |
+
"logps/chosen": -290.67987060546875,
|
| 2101 |
+
"logps/rejected": -281.71405029296875,
|
| 2102 |
+
"loss": 0.0816,
|
| 2103 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 2104 |
+
"rewards/chosen": 0.026712050661444664,
|
| 2105 |
+
"rewards/margins": 0.026149820536375046,
|
| 2106 |
+
"rewards/rejected": 0.0005622319877147675,
|
| 2107 |
+
"step": 1340
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 0.86,
|
| 2111 |
+
"learning_rate": 2.756646575638025e-07,
|
| 2112 |
+
"logits/chosen": -0.9123773574829102,
|
| 2113 |
+
"logits/rejected": -0.9757976531982422,
|
| 2114 |
+
"logps/chosen": -270.1868896484375,
|
| 2115 |
+
"logps/rejected": -303.7170715332031,
|
| 2116 |
+
"loss": 0.0753,
|
| 2117 |
+
"rewards/accuracies": 0.46875,
|
| 2118 |
+
"rewards/chosen": 0.025742124766111374,
|
| 2119 |
+
"rewards/margins": 0.034964997321367264,
|
| 2120 |
+
"rewards/rejected": -0.009222874417901039,
|
| 2121 |
+
"step": 1350
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 0.87,
|
| 2125 |
+
"learning_rate": 2.507055309070111e-07,
|
| 2126 |
+
"logits/chosen": -0.9779159426689148,
|
| 2127 |
+
"logits/rejected": -0.9737545847892761,
|
| 2128 |
+
"logps/chosen": -285.8861083984375,
|
| 2129 |
+
"logps/rejected": -280.3493347167969,
|
| 2130 |
+
"loss": 0.0912,
|
| 2131 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 2132 |
+
"rewards/chosen": 0.021084221079945564,
|
| 2133 |
+
"rewards/margins": 0.029098382219672203,
|
| 2134 |
+
"rewards/rejected": -0.008014162071049213,
|
| 2135 |
+
"step": 1360
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 0.88,
|
| 2139 |
+
"learning_rate": 2.2687094531076565e-07,
|
| 2140 |
+
"logits/chosen": -1.0015113353729248,
|
| 2141 |
+
"logits/rejected": -0.9118536114692688,
|
| 2142 |
+
"logps/chosen": -292.0755920410156,
|
| 2143 |
+
"logps/rejected": -336.7613830566406,
|
| 2144 |
+
"loss": 0.0668,
|
| 2145 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 2146 |
+
"rewards/chosen": 0.023335173726081848,
|
| 2147 |
+
"rewards/margins": 0.042082760483026505,
|
| 2148 |
+
"rewards/rejected": -0.018747588619589806,
|
| 2149 |
+
"step": 1370
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 0.88,
|
| 2153 |
+
"learning_rate": 2.0417281694279424e-07,
|
| 2154 |
+
"logits/chosen": -0.9744800329208374,
|
| 2155 |
+
"logits/rejected": -0.8867910504341125,
|
| 2156 |
+
"logps/chosen": -322.03643798828125,
|
| 2157 |
+
"logps/rejected": -347.53192138671875,
|
| 2158 |
+
"loss": 0.0651,
|
| 2159 |
+
"rewards/accuracies": 0.53125,
|
| 2160 |
+
"rewards/chosen": 0.016095371916890144,
|
| 2161 |
+
"rewards/margins": 0.03282975032925606,
|
| 2162 |
+
"rewards/rejected": -0.016734374687075615,
|
| 2163 |
+
"step": 1380
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.89,
|
| 2167 |
+
"learning_rate": 1.8262249379585484e-07,
|
| 2168 |
+
"logits/chosen": -0.8901314735412598,
|
| 2169 |
+
"logits/rejected": -0.9079896807670593,
|
| 2170 |
+
"logps/chosen": -318.6395568847656,
|
| 2171 |
+
"logps/rejected": -306.5535583496094,
|
| 2172 |
+
"loss": 0.0696,
|
| 2173 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 2174 |
+
"rewards/chosen": 0.02375376783311367,
|
| 2175 |
+
"rewards/margins": 0.02338986098766327,
|
| 2176 |
+
"rewards/rejected": 0.00036390620516613126,
|
| 2177 |
+
"step": 1390
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 0.9,
|
| 2181 |
+
"learning_rate": 1.6223075001427667e-07,
|
| 2182 |
+
"logits/chosen": -0.9552518725395203,
|
| 2183 |
+
"logits/rejected": -0.9494439959526062,
|
| 2184 |
+
"logps/chosen": -281.0766906738281,
|
| 2185 |
+
"logps/rejected": -289.63055419921875,
|
| 2186 |
+
"loss": 0.0863,
|
| 2187 |
+
"rewards/accuracies": 0.40625,
|
| 2188 |
+
"rewards/chosen": 0.010858943685889244,
|
| 2189 |
+
"rewards/margins": 0.015300577506422997,
|
| 2190 |
+
"rewards/rejected": -0.004441632889211178,
|
| 2191 |
+
"step": 1400
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 0.9,
|
| 2195 |
+
"eval_logits/chosen": -0.9446871280670166,
|
| 2196 |
+
"eval_logits/rejected": -0.9120718240737915,
|
| 2197 |
+
"eval_logps/chosen": -369.4450378417969,
|
| 2198 |
+
"eval_logps/rejected": -379.5608215332031,
|
| 2199 |
+
"eval_loss": 0.038691576570272446,
|
| 2200 |
+
"eval_rewards/accuracies": 0.6054999828338623,
|
| 2201 |
+
"eval_rewards/chosen": 0.02686023712158203,
|
| 2202 |
+
"eval_rewards/margins": 0.03918563574552536,
|
| 2203 |
+
"eval_rewards/rejected": -0.01232539676129818,
|
| 2204 |
+
"eval_runtime": 539.9174,
|
| 2205 |
+
"eval_samples_per_second": 3.704,
|
| 2206 |
+
"eval_steps_per_second": 0.926,
|
| 2207 |
+
"step": 1400
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.9,
|
| 2211 |
+
"learning_rate": 1.4300778050739317e-07,
|
| 2212 |
+
"logits/chosen": -0.9774508476257324,
|
| 2213 |
+
"logits/rejected": -0.9408830404281616,
|
| 2214 |
+
"logps/chosen": -298.72247314453125,
|
| 2215 |
+
"logps/rejected": -294.75408935546875,
|
| 2216 |
+
"loss": 0.0665,
|
| 2217 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 2218 |
+
"rewards/chosen": 0.01832416281104088,
|
| 2219 |
+
"rewards/margins": 0.014284146018326283,
|
| 2220 |
+
"rewards/rejected": 0.004040017258375883,
|
| 2221 |
+
"step": 1410
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.91,
|
| 2225 |
+
"learning_rate": 1.2496319585257183e-07,
|
| 2226 |
+
"logits/chosen": -0.9892775416374207,
|
| 2227 |
+
"logits/rejected": -0.9716461300849915,
|
| 2228 |
+
"logps/chosen": -323.1170349121094,
|
| 2229 |
+
"logps/rejected": -303.7762451171875,
|
| 2230 |
+
"loss": 0.0854,
|
| 2231 |
+
"rewards/accuracies": 0.518750011920929,
|
| 2232 |
+
"rewards/chosen": 0.02715294435620308,
|
| 2233 |
+
"rewards/margins": 0.03064989112317562,
|
| 2234 |
+
"rewards/rejected": -0.0034969463013112545,
|
| 2235 |
+
"step": 1420
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.92,
|
| 2239 |
+
"learning_rate": 1.0810601749037669e-07,
|
| 2240 |
+
"logits/chosen": -0.9020091891288757,
|
| 2241 |
+
"logits/rejected": -0.9708755612373352,
|
| 2242 |
+
"logps/chosen": -289.08258056640625,
|
| 2243 |
+
"logps/rejected": -339.64703369140625,
|
| 2244 |
+
"loss": 0.0713,
|
| 2245 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 2246 |
+
"rewards/chosen": 0.023048024624586105,
|
| 2247 |
+
"rewards/margins": 0.04538671672344208,
|
| 2248 |
+
"rewards/rejected": -0.022338688373565674,
|
| 2249 |
+
"step": 1430
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.92,
|
| 2253 |
+
"learning_rate": 9.244467321427585e-08,
|
| 2254 |
+
"logits/chosen": -1.0033928155899048,
|
| 2255 |
+
"logits/rejected": -1.0245609283447266,
|
| 2256 |
+
"logps/chosen": -276.730712890625,
|
| 2257 |
+
"logps/rejected": -303.7331237792969,
|
| 2258 |
+
"loss": 0.0785,
|
| 2259 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2260 |
+
"rewards/chosen": 0.033762019127607346,
|
| 2261 |
+
"rewards/margins": 0.05797597020864487,
|
| 2262 |
+
"rewards/rejected": -0.024213949218392372,
|
| 2263 |
+
"step": 1440
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.93,
|
| 2267 |
+
"learning_rate": 7.798699295714002e-08,
|
| 2268 |
+
"logits/chosen": -0.9668358564376831,
|
| 2269 |
+
"logits/rejected": -0.9086298942565918,
|
| 2270 |
+
"logps/chosen": -320.6669006347656,
|
| 2271 |
+
"logps/rejected": -296.5704345703125,
|
| 2272 |
+
"loss": 0.0777,
|
| 2273 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 2274 |
+
"rewards/chosen": 0.023420458659529686,
|
| 2275 |
+
"rewards/margins": 0.03309481590986252,
|
| 2276 |
+
"rewards/rejected": -0.009674356319010258,
|
| 2277 |
+
"step": 1450
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.93,
|
| 2281 |
+
"learning_rate": 6.474020487664934e-08,
|
| 2282 |
+
"logits/chosen": -0.9561643600463867,
|
| 2283 |
+
"logits/rejected": -0.9422094225883484,
|
| 2284 |
+
"logps/chosen": -276.42156982421875,
|
| 2285 |
+
"logps/rejected": -291.9980163574219,
|
| 2286 |
+
"loss": 0.076,
|
| 2287 |
+
"rewards/accuracies": 0.53125,
|
| 2288 |
+
"rewards/chosen": 0.023049993440508842,
|
| 2289 |
+
"rewards/margins": 0.038113731890916824,
|
| 2290 |
+
"rewards/rejected": -0.01506374217569828,
|
| 2291 |
+
"step": 1460
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.94,
|
| 2295 |
+
"learning_rate": 5.271093174155223e-08,
|
| 2296 |
+
"logits/chosen": -1.0463026762008667,
|
| 2297 |
+
"logits/rejected": -0.9269806742668152,
|
| 2298 |
+
"logps/chosen": -271.7001037597656,
|
| 2299 |
+
"logps/rejected": -273.11419677734375,
|
| 2300 |
+
"loss": 0.0773,
|
| 2301 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 2302 |
+
"rewards/chosen": 0.02771763503551483,
|
| 2303 |
+
"rewards/margins": 0.02654975652694702,
|
| 2304 |
+
"rewards/rejected": 0.0011678790906444192,
|
| 2305 |
+
"step": 1470
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.95,
|
| 2309 |
+
"learning_rate": 4.190518762059587e-08,
|
| 2310 |
+
"logits/chosen": -0.9642139673233032,
|
| 2311 |
+
"logits/rejected": -0.9358107447624207,
|
| 2312 |
+
"logps/chosen": -301.32647705078125,
|
| 2313 |
+
"logps/rejected": -310.1565856933594,
|
| 2314 |
+
"loss": 0.0955,
|
| 2315 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 2316 |
+
"rewards/chosen": 0.019302543252706528,
|
| 2317 |
+
"rewards/margins": 0.034524787217378616,
|
| 2318 |
+
"rewards/rejected": -0.015222239308059216,
|
| 2319 |
+
"step": 1480
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.95,
|
| 2323 |
+
"learning_rate": 3.232837487577384e-08,
|
| 2324 |
+
"logits/chosen": -0.9852960705757141,
|
| 2325 |
+
"logits/rejected": -0.9471799731254578,
|
| 2326 |
+
"logps/chosen": -306.68548583984375,
|
| 2327 |
+
"logps/rejected": -294.8326416015625,
|
| 2328 |
+
"loss": 0.0895,
|
| 2329 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 2330 |
+
"rewards/chosen": 0.018351960927248,
|
| 2331 |
+
"rewards/margins": 0.04016681760549545,
|
| 2332 |
+
"rewards/rejected": -0.0218148622661829,
|
| 2333 |
+
"step": 1490
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.96,
|
| 2337 |
+
"learning_rate": 2.39852814614025e-08,
|
| 2338 |
+
"logits/chosen": -0.9789448976516724,
|
| 2339 |
+
"logits/rejected": -0.9438329935073853,
|
| 2340 |
+
"logps/chosen": -281.9150085449219,
|
| 2341 |
+
"logps/rejected": -267.4187316894531,
|
| 2342 |
+
"loss": 0.0904,
|
| 2343 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 2344 |
+
"rewards/chosen": 0.019616561010479927,
|
| 2345 |
+
"rewards/margins": 0.023770466446876526,
|
| 2346 |
+
"rewards/rejected": -0.004153906367719173,
|
| 2347 |
+
"step": 1500
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.96,
|
| 2351 |
+
"eval_logits/chosen": -0.9535864591598511,
|
| 2352 |
+
"eval_logits/rejected": -0.9203009009361267,
|
| 2353 |
+
"eval_logps/chosen": -369.494384765625,
|
| 2354 |
+
"eval_logps/rejected": -379.5999755859375,
|
| 2355 |
+
"eval_loss": 0.03863796219229698,
|
| 2356 |
+
"eval_rewards/accuracies": 0.6044999957084656,
|
| 2357 |
+
"eval_rewards/chosen": 0.026810916140675545,
|
| 2358 |
+
"eval_rewards/margins": 0.03917544335126877,
|
| 2359 |
+
"eval_rewards/rejected": -0.012364527210593224,
|
| 2360 |
+
"eval_runtime": 539.9867,
|
| 2361 |
+
"eval_samples_per_second": 3.704,
|
| 2362 |
+
"eval_steps_per_second": 0.926,
|
| 2363 |
+
"step": 1500
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 0.97,
|
| 2367 |
+
"learning_rate": 1.6880078530367716e-08,
|
| 2368 |
+
"logits/chosen": -0.9464886784553528,
|
| 2369 |
+
"logits/rejected": -0.9203370213508606,
|
| 2370 |
+
"logps/chosen": -291.42730712890625,
|
| 2371 |
+
"logps/rejected": -302.47357177734375,
|
| 2372 |
+
"loss": 0.0968,
|
| 2373 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 2374 |
+
"rewards/chosen": 0.018558986485004425,
|
| 2375 |
+
"rewards/margins": 0.03005158342421055,
|
| 2376 |
+
"rewards/rejected": -0.011492597870528698,
|
| 2377 |
+
"step": 1510
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 0.97,
|
| 2381 |
+
"learning_rate": 1.1016318348746058e-08,
|
| 2382 |
+
"logits/chosen": -0.9404022097587585,
|
| 2383 |
+
"logits/rejected": -0.9021556973457336,
|
| 2384 |
+
"logps/chosen": -266.0870666503906,
|
| 2385 |
+
"logps/rejected": -277.4389343261719,
|
| 2386 |
+
"loss": 0.1015,
|
| 2387 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 2388 |
+
"rewards/chosen": 0.021859928965568542,
|
| 2389 |
+
"rewards/margins": 0.028949573636054993,
|
| 2390 |
+
"rewards/rejected": -0.007089647464454174,
|
| 2391 |
+
"step": 1520
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.98,
|
| 2395 |
+
"learning_rate": 6.396932519840693e-09,
|
| 2396 |
+
"logits/chosen": -0.9894587397575378,
|
| 2397 |
+
"logits/rejected": -1.043691873550415,
|
| 2398 |
+
"logps/chosen": -284.725830078125,
|
| 2399 |
+
"logps/rejected": -302.0658874511719,
|
| 2400 |
+
"loss": 0.0885,
|
| 2401 |
+
"rewards/accuracies": 0.4375,
|
| 2402 |
+
"rewards/chosen": 0.01785745844244957,
|
| 2403 |
+
"rewards/margins": 0.0238940566778183,
|
| 2404 |
+
"rewards/rejected": -0.0060366010293364525,
|
| 2405 |
+
"step": 1530
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.99,
|
| 2409 |
+
"learning_rate": 3.024230518515192e-09,
|
| 2410 |
+
"logits/chosen": -1.0197975635528564,
|
| 2411 |
+
"logits/rejected": -0.9348329305648804,
|
| 2412 |
+
"logps/chosen": -242.8181915283203,
|
| 2413 |
+
"logps/rejected": -304.18841552734375,
|
| 2414 |
+
"loss": 0.115,
|
| 2415 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 2416 |
+
"rewards/chosen": 0.027777481824159622,
|
| 2417 |
+
"rewards/margins": 0.04096692427992821,
|
| 2418 |
+
"rewards/rejected": -0.013189440593123436,
|
| 2419 |
+
"step": 1540
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 0.99,
|
| 2423 |
+
"learning_rate": 8.998985365679669e-10,
|
| 2424 |
+
"logits/chosen": -0.918280303478241,
|
| 2425 |
+
"logits/rejected": -0.9589518308639526,
|
| 2426 |
+
"logps/chosen": -293.98236083984375,
|
| 2427 |
+
"logps/rejected": -301.8656921386719,
|
| 2428 |
+
"loss": 0.0665,
|
| 2429 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 2430 |
+
"rewards/chosen": 0.025340059772133827,
|
| 2431 |
+
"rewards/margins": 0.033479273319244385,
|
| 2432 |
+
"rewards/rejected": -0.008139212615787983,
|
| 2433 |
+
"step": 1550
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 1.0,
|
| 2437 |
+
"learning_rate": 2.499863971494598e-11,
|
| 2438 |
+
"logits/chosen": -0.9552096128463745,
|
| 2439 |
+
"logits/rejected": -0.8985775113105774,
|
| 2440 |
+
"logps/chosen": -317.04779052734375,
|
| 2441 |
+
"logps/rejected": -310.51031494140625,
|
| 2442 |
+
"loss": 0.0696,
|
| 2443 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 2444 |
+
"rewards/chosen": 0.025858771055936813,
|
| 2445 |
+
"rewards/margins": 0.03246783837676048,
|
| 2446 |
+
"rewards/rejected": -0.006609070114791393,
|
| 2447 |
+
"step": 1560
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 1.0,
|
| 2451 |
+
"step": 1562,
|
| 2452 |
+
"total_flos": 0.0,
|
| 2453 |
+
"train_loss": 0.08419492204701015,
|
| 2454 |
+
"train_runtime": 22113.8812,
|
| 2455 |
+
"train_samples_per_second": 1.131,
|
| 2456 |
+
"train_steps_per_second": 0.071
|
| 2457 |
+
}
|
| 2458 |
+
],
|
| 2459 |
+
"logging_steps": 10,
|
| 2460 |
+
"max_steps": 1562,
|
| 2461 |
+
"num_input_tokens_seen": 0,
|
| 2462 |
+
"num_train_epochs": 1,
|
| 2463 |
+
"save_steps": 100,
|
| 2464 |
+
"total_flos": 0.0,
|
| 2465 |
+
"train_batch_size": 4,
|
| 2466 |
+
"trial_name": null,
|
| 2467 |
+
"trial_params": null
|
| 2468 |
+
}
|