Model save
Browse files- README.md +75 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +509 -0
README.md
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: zephyr-7b-align-scan-2e-07-0.39-polynomial-1.0
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# zephyr-7b-align-scan-2e-07-0.39-polynomial-1.0
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.6324
|
| 22 |
+
- Rewards/chosen: 0.2521
|
| 23 |
+
- Rewards/rejected: -0.1128
|
| 24 |
+
- Rewards/accuracies: 0.3492
|
| 25 |
+
- Rewards/margins: 0.3648
|
| 26 |
+
- Logps/rejected: -81.4175
|
| 27 |
+
- Logps/chosen: -73.8449
|
| 28 |
+
- Logits/rejected: -2.5055
|
| 29 |
+
- Logits/chosen: -2.5222
|
| 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: 2e-07
|
| 49 |
+
- train_batch_size: 8
|
| 50 |
+
- eval_batch_size: 8
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 4
|
| 54 |
+
- gradient_accumulation_steps: 2
|
| 55 |
+
- total_train_batch_size: 64
|
| 56 |
+
- total_eval_batch_size: 32
|
| 57 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 58 |
+
- lr_scheduler_type: polynomial
|
| 59 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
+
- num_epochs: 1
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 65 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 66 |
+
| 0.6538 | 0.3484 | 100 | 0.6320 | 0.5810 | 0.2964 | 0.3472 | 0.2846 | -80.3683 | -73.0014 | -2.5308 | -2.5469 |
|
| 67 |
+
| 0.6462 | 0.6969 | 200 | 0.6324 | 0.2521 | -0.1128 | 0.3492 | 0.3648 | -81.4175 | -73.8449 | -2.5055 | -2.5222 |
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
### Framework versions
|
| 71 |
+
|
| 72 |
+
- Transformers 4.44.2
|
| 73 |
+
- Pytorch 2.4.0
|
| 74 |
+
- Datasets 2.21.0
|
| 75 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.6390685353960309,
|
| 5 |
+
"train_runtime": 3341.9408,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.488,
|
| 8 |
+
"train_steps_per_second": 0.086
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.2"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.6390685353960309,
|
| 5 |
+
"train_runtime": 3341.9408,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.488,
|
| 8 |
+
"train_steps_per_second": 0.086
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,509 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 287,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.003484320557491289,
|
| 13 |
+
"grad_norm": 225.57155111737745,
|
| 14 |
+
"learning_rate": 6.896551724137931e-09,
|
| 15 |
+
"logits/chosen": -2.5345611572265625,
|
| 16 |
+
"logits/rejected": -2.581700563430786,
|
| 17 |
+
"logps/chosen": -60.002105712890625,
|
| 18 |
+
"logps/rejected": -99.98374938964844,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/rejected": 0.0,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03484320557491289,
|
| 28 |
+
"grad_norm": 216.6261369698797,
|
| 29 |
+
"learning_rate": 6.896551724137931e-08,
|
| 30 |
+
"logits/chosen": -2.563714027404785,
|
| 31 |
+
"logits/rejected": -2.562321901321411,
|
| 32 |
+
"logps/chosen": -59.66276550292969,
|
| 33 |
+
"logps/rejected": -73.39944458007812,
|
| 34 |
+
"loss": 0.6916,
|
| 35 |
+
"rewards/accuracies": 0.2291666716337204,
|
| 36 |
+
"rewards/chosen": -0.001211778144352138,
|
| 37 |
+
"rewards/margins": 0.00601451750844717,
|
| 38 |
+
"rewards/rejected": -0.007226296700537205,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06968641114982578,
|
| 43 |
+
"grad_norm": 262.5210498959883,
|
| 44 |
+
"learning_rate": 1.3793103448275863e-07,
|
| 45 |
+
"logits/chosen": -2.6048154830932617,
|
| 46 |
+
"logits/rejected": -2.5637078285217285,
|
| 47 |
+
"logps/chosen": -104.09293365478516,
|
| 48 |
+
"logps/rejected": -94.90653991699219,
|
| 49 |
+
"loss": 0.6903,
|
| 50 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 51 |
+
"rewards/chosen": 0.013311183080077171,
|
| 52 |
+
"rewards/margins": 0.018149277195334435,
|
| 53 |
+
"rewards/rejected": -0.004838095046579838,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.10452961672473868,
|
| 58 |
+
"grad_norm": 268.8517801211917,
|
| 59 |
+
"learning_rate": 1.9961240310077517e-07,
|
| 60 |
+
"logits/chosen": -2.5926074981689453,
|
| 61 |
+
"logits/rejected": -2.572615623474121,
|
| 62 |
+
"logps/chosen": -82.38002014160156,
|
| 63 |
+
"logps/rejected": -91.52497863769531,
|
| 64 |
+
"loss": 0.6791,
|
| 65 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 66 |
+
"rewards/chosen": 0.04545076563954353,
|
| 67 |
+
"rewards/margins": 0.039520200341939926,
|
| 68 |
+
"rewards/rejected": 0.005930562969297171,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13937282229965156,
|
| 73 |
+
"grad_norm": 209.92212859559183,
|
| 74 |
+
"learning_rate": 1.9573643410852711e-07,
|
| 75 |
+
"logits/chosen": -2.4985804557800293,
|
| 76 |
+
"logits/rejected": -2.496473550796509,
|
| 77 |
+
"logps/chosen": -77.8263931274414,
|
| 78 |
+
"logps/rejected": -73.18889617919922,
|
| 79 |
+
"loss": 0.6608,
|
| 80 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 81 |
+
"rewards/chosen": 0.03588957339525223,
|
| 82 |
+
"rewards/margins": 0.10904321819543839,
|
| 83 |
+
"rewards/rejected": -0.07315365225076675,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.17421602787456447,
|
| 88 |
+
"grad_norm": 197.5547212101371,
|
| 89 |
+
"learning_rate": 1.9186046511627906e-07,
|
| 90 |
+
"logits/chosen": -2.53145170211792,
|
| 91 |
+
"logits/rejected": -2.5351381301879883,
|
| 92 |
+
"logps/chosen": -63.558860778808594,
|
| 93 |
+
"logps/rejected": -75.93846130371094,
|
| 94 |
+
"loss": 0.6646,
|
| 95 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 96 |
+
"rewards/chosen": 0.1483130156993866,
|
| 97 |
+
"rewards/margins": 0.08396634459495544,
|
| 98 |
+
"rewards/rejected": 0.06434665620326996,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.20905923344947736,
|
| 103 |
+
"grad_norm": 191.33111792340372,
|
| 104 |
+
"learning_rate": 1.87984496124031e-07,
|
| 105 |
+
"logits/chosen": -2.4998061656951904,
|
| 106 |
+
"logits/rejected": -2.493762254714966,
|
| 107 |
+
"logps/chosen": -71.91557312011719,
|
| 108 |
+
"logps/rejected": -67.55859375,
|
| 109 |
+
"loss": 0.6549,
|
| 110 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 111 |
+
"rewards/chosen": 0.5034655928611755,
|
| 112 |
+
"rewards/margins": 0.17197737097740173,
|
| 113 |
+
"rewards/rejected": 0.3314882814884186,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.24390243902439024,
|
| 118 |
+
"grad_norm": 204.75139928311083,
|
| 119 |
+
"learning_rate": 1.8410852713178295e-07,
|
| 120 |
+
"logits/chosen": -2.5222249031066895,
|
| 121 |
+
"logits/rejected": -2.5176424980163574,
|
| 122 |
+
"logps/chosen": -62.58455276489258,
|
| 123 |
+
"logps/rejected": -67.23956298828125,
|
| 124 |
+
"loss": 0.6491,
|
| 125 |
+
"rewards/accuracies": 0.28125,
|
| 126 |
+
"rewards/chosen": 0.644144594669342,
|
| 127 |
+
"rewards/margins": 0.18119266629219055,
|
| 128 |
+
"rewards/rejected": 0.4629518389701843,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2787456445993031,
|
| 133 |
+
"grad_norm": 237.92667425371948,
|
| 134 |
+
"learning_rate": 1.8023255813953487e-07,
|
| 135 |
+
"logits/chosen": -2.4706761837005615,
|
| 136 |
+
"logits/rejected": -2.4605259895324707,
|
| 137 |
+
"logps/chosen": -73.8516845703125,
|
| 138 |
+
"logps/rejected": -76.4228286743164,
|
| 139 |
+
"loss": 0.6602,
|
| 140 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 141 |
+
"rewards/chosen": 0.7071425318717957,
|
| 142 |
+
"rewards/margins": 0.2812144160270691,
|
| 143 |
+
"rewards/rejected": 0.4259282052516937,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.313588850174216,
|
| 148 |
+
"grad_norm": 188.49653222846555,
|
| 149 |
+
"learning_rate": 1.7635658914728681e-07,
|
| 150 |
+
"logits/chosen": -2.48022723197937,
|
| 151 |
+
"logits/rejected": -2.494345188140869,
|
| 152 |
+
"logps/chosen": -64.58412170410156,
|
| 153 |
+
"logps/rejected": -69.22706604003906,
|
| 154 |
+
"loss": 0.6454,
|
| 155 |
+
"rewards/accuracies": 0.3125,
|
| 156 |
+
"rewards/chosen": 0.5452820062637329,
|
| 157 |
+
"rewards/margins": 0.1993606686592102,
|
| 158 |
+
"rewards/rejected": 0.34592127799987793,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.34843205574912894,
|
| 163 |
+
"grad_norm": 230.52946384799608,
|
| 164 |
+
"learning_rate": 1.7248062015503873e-07,
|
| 165 |
+
"logits/chosen": -2.460999011993408,
|
| 166 |
+
"logits/rejected": -2.462146282196045,
|
| 167 |
+
"logps/chosen": -73.88617706298828,
|
| 168 |
+
"logps/rejected": -80.31554412841797,
|
| 169 |
+
"loss": 0.6538,
|
| 170 |
+
"rewards/accuracies": 0.3125,
|
| 171 |
+
"rewards/chosen": 0.5401462316513062,
|
| 172 |
+
"rewards/margins": 0.2530631422996521,
|
| 173 |
+
"rewards/rejected": 0.28708305954933167,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.34843205574912894,
|
| 178 |
+
"eval_logits/chosen": -2.5468788146972656,
|
| 179 |
+
"eval_logits/rejected": -2.530789613723755,
|
| 180 |
+
"eval_logps/chosen": -73.00140380859375,
|
| 181 |
+
"eval_logps/rejected": -80.36827850341797,
|
| 182 |
+
"eval_loss": 0.6319671273231506,
|
| 183 |
+
"eval_rewards/accuracies": 0.3472222089767456,
|
| 184 |
+
"eval_rewards/chosen": 0.5810332298278809,
|
| 185 |
+
"eval_rewards/margins": 0.2846057713031769,
|
| 186 |
+
"eval_rewards/rejected": 0.2964273989200592,
|
| 187 |
+
"eval_runtime": 114.037,
|
| 188 |
+
"eval_samples_per_second": 17.538,
|
| 189 |
+
"eval_steps_per_second": 0.552,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.3832752613240418,
|
| 194 |
+
"grad_norm": 251.4434825633798,
|
| 195 |
+
"learning_rate": 1.6860465116279068e-07,
|
| 196 |
+
"logits/chosen": -2.4884157180786133,
|
| 197 |
+
"logits/rejected": -2.44923734664917,
|
| 198 |
+
"logps/chosen": -72.09432220458984,
|
| 199 |
+
"logps/rejected": -62.603981018066406,
|
| 200 |
+
"loss": 0.6295,
|
| 201 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 202 |
+
"rewards/chosen": 0.41952481865882874,
|
| 203 |
+
"rewards/margins": 0.2609493136405945,
|
| 204 |
+
"rewards/rejected": 0.15857550501823425,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.4181184668989547,
|
| 209 |
+
"grad_norm": 183.32886181861778,
|
| 210 |
+
"learning_rate": 1.6472868217054262e-07,
|
| 211 |
+
"logits/chosen": -2.5084633827209473,
|
| 212 |
+
"logits/rejected": -2.4761087894439697,
|
| 213 |
+
"logps/chosen": -76.3115234375,
|
| 214 |
+
"logps/rejected": -66.27922058105469,
|
| 215 |
+
"loss": 0.6149,
|
| 216 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 217 |
+
"rewards/chosen": 0.44851747155189514,
|
| 218 |
+
"rewards/margins": 0.29549649357795715,
|
| 219 |
+
"rewards/rejected": 0.15302105247974396,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4529616724738676,
|
| 224 |
+
"grad_norm": 260.8387325270783,
|
| 225 |
+
"learning_rate": 1.6085271317829457e-07,
|
| 226 |
+
"logits/chosen": -2.5476431846618652,
|
| 227 |
+
"logits/rejected": -2.5282294750213623,
|
| 228 |
+
"logps/chosen": -82.90985870361328,
|
| 229 |
+
"logps/rejected": -87.62440490722656,
|
| 230 |
+
"loss": 0.6315,
|
| 231 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 232 |
+
"rewards/chosen": 0.37865743041038513,
|
| 233 |
+
"rewards/margins": 0.42385053634643555,
|
| 234 |
+
"rewards/rejected": -0.045193083584308624,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.4878048780487805,
|
| 239 |
+
"grad_norm": 234.8662840093771,
|
| 240 |
+
"learning_rate": 1.569767441860465e-07,
|
| 241 |
+
"logits/chosen": -2.445432186126709,
|
| 242 |
+
"logits/rejected": -2.4326701164245605,
|
| 243 |
+
"logps/chosen": -80.27764892578125,
|
| 244 |
+
"logps/rejected": -70.95781707763672,
|
| 245 |
+
"loss": 0.6136,
|
| 246 |
+
"rewards/accuracies": 0.34375,
|
| 247 |
+
"rewards/chosen": 0.3384358584880829,
|
| 248 |
+
"rewards/margins": 0.4139274060726166,
|
| 249 |
+
"rewards/rejected": -0.07549159228801727,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.5226480836236934,
|
| 254 |
+
"grad_norm": 234.49493704296844,
|
| 255 |
+
"learning_rate": 1.5310077519379846e-07,
|
| 256 |
+
"logits/chosen": -2.5063323974609375,
|
| 257 |
+
"logits/rejected": -2.4608452320098877,
|
| 258 |
+
"logps/chosen": -78.8966064453125,
|
| 259 |
+
"logps/rejected": -79.84175872802734,
|
| 260 |
+
"loss": 0.6478,
|
| 261 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 262 |
+
"rewards/chosen": 0.22921499609947205,
|
| 263 |
+
"rewards/margins": 0.34269630908966064,
|
| 264 |
+
"rewards/rejected": -0.1134813204407692,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.5574912891986062,
|
| 269 |
+
"grad_norm": 208.63525471171238,
|
| 270 |
+
"learning_rate": 1.4922480620155038e-07,
|
| 271 |
+
"logits/chosen": -2.462036609649658,
|
| 272 |
+
"logits/rejected": -2.483565092086792,
|
| 273 |
+
"logps/chosen": -64.0514907836914,
|
| 274 |
+
"logps/rejected": -72.12866973876953,
|
| 275 |
+
"loss": 0.6534,
|
| 276 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 277 |
+
"rewards/chosen": 0.18461140990257263,
|
| 278 |
+
"rewards/margins": 0.24535641074180603,
|
| 279 |
+
"rewards/rejected": -0.060745008289813995,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.5923344947735192,
|
| 284 |
+
"grad_norm": 215.9542751971873,
|
| 285 |
+
"learning_rate": 1.4534883720930232e-07,
|
| 286 |
+
"logits/chosen": -2.459441900253296,
|
| 287 |
+
"logits/rejected": -2.444371461868286,
|
| 288 |
+
"logps/chosen": -68.60945129394531,
|
| 289 |
+
"logps/rejected": -77.02355194091797,
|
| 290 |
+
"loss": 0.6049,
|
| 291 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 292 |
+
"rewards/chosen": 0.1395992934703827,
|
| 293 |
+
"rewards/margins": 0.416309118270874,
|
| 294 |
+
"rewards/rejected": -0.27670982480049133,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.627177700348432,
|
| 299 |
+
"grad_norm": 220.77017684857242,
|
| 300 |
+
"learning_rate": 1.4147286821705424e-07,
|
| 301 |
+
"logits/chosen": -2.5126798152923584,
|
| 302 |
+
"logits/rejected": -2.5011754035949707,
|
| 303 |
+
"logps/chosen": -91.29668426513672,
|
| 304 |
+
"logps/rejected": -86.69538879394531,
|
| 305 |
+
"loss": 0.6539,
|
| 306 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 307 |
+
"rewards/chosen": 0.21858568489551544,
|
| 308 |
+
"rewards/margins": 0.29070785641670227,
|
| 309 |
+
"rewards/rejected": -0.07212214171886444,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.662020905923345,
|
| 314 |
+
"grad_norm": 162.410148810253,
|
| 315 |
+
"learning_rate": 1.3759689922480619e-07,
|
| 316 |
+
"logits/chosen": -2.5120596885681152,
|
| 317 |
+
"logits/rejected": -2.502159595489502,
|
| 318 |
+
"logps/chosen": -70.31501770019531,
|
| 319 |
+
"logps/rejected": -81.38325500488281,
|
| 320 |
+
"loss": 0.6211,
|
| 321 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 322 |
+
"rewards/chosen": 0.28721699118614197,
|
| 323 |
+
"rewards/margins": 0.23069223761558533,
|
| 324 |
+
"rewards/rejected": 0.05652473121881485,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.6968641114982579,
|
| 329 |
+
"grad_norm": 291.03325951738674,
|
| 330 |
+
"learning_rate": 1.3372093023255813e-07,
|
| 331 |
+
"logits/chosen": -2.545940399169922,
|
| 332 |
+
"logits/rejected": -2.5534462928771973,
|
| 333 |
+
"logps/chosen": -89.51722717285156,
|
| 334 |
+
"logps/rejected": -91.98088073730469,
|
| 335 |
+
"loss": 0.6462,
|
| 336 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 337 |
+
"rewards/chosen": 0.4442625641822815,
|
| 338 |
+
"rewards/margins": 0.4170360565185547,
|
| 339 |
+
"rewards/rejected": 0.02722644805908203,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.6968641114982579,
|
| 344 |
+
"eval_logits/chosen": -2.522193670272827,
|
| 345 |
+
"eval_logits/rejected": -2.505502939224243,
|
| 346 |
+
"eval_logps/chosen": -73.84490966796875,
|
| 347 |
+
"eval_logps/rejected": -81.41748046875,
|
| 348 |
+
"eval_loss": 0.6323519349098206,
|
| 349 |
+
"eval_rewards/accuracies": 0.3492063581943512,
|
| 350 |
+
"eval_rewards/chosen": 0.2520641088485718,
|
| 351 |
+
"eval_rewards/margins": 0.3648246228694916,
|
| 352 |
+
"eval_rewards/rejected": -0.1127605140209198,
|
| 353 |
+
"eval_runtime": 113.9702,
|
| 354 |
+
"eval_samples_per_second": 17.548,
|
| 355 |
+
"eval_steps_per_second": 0.553,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.7317073170731707,
|
| 360 |
+
"grad_norm": 333.8956625644709,
|
| 361 |
+
"learning_rate": 1.2984496124031008e-07,
|
| 362 |
+
"logits/chosen": -2.528879404067993,
|
| 363 |
+
"logits/rejected": -2.501743793487549,
|
| 364 |
+
"logps/chosen": -68.72173309326172,
|
| 365 |
+
"logps/rejected": -64.01664733886719,
|
| 366 |
+
"loss": 0.6317,
|
| 367 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 368 |
+
"rewards/chosen": 0.2329626977443695,
|
| 369 |
+
"rewards/margins": 0.471733957529068,
|
| 370 |
+
"rewards/rejected": -0.2387712299823761,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.7665505226480837,
|
| 375 |
+
"grad_norm": 176.716286583645,
|
| 376 |
+
"learning_rate": 1.2596899224806202e-07,
|
| 377 |
+
"logits/chosen": -2.5579559803009033,
|
| 378 |
+
"logits/rejected": -2.535482883453369,
|
| 379 |
+
"logps/chosen": -72.86189270019531,
|
| 380 |
+
"logps/rejected": -71.62294006347656,
|
| 381 |
+
"loss": 0.6392,
|
| 382 |
+
"rewards/accuracies": 0.28125,
|
| 383 |
+
"rewards/chosen": 0.23254899680614471,
|
| 384 |
+
"rewards/margins": 0.21685466170310974,
|
| 385 |
+
"rewards/rejected": 0.015694305300712585,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.8013937282229965,
|
| 390 |
+
"grad_norm": 226.51144499862002,
|
| 391 |
+
"learning_rate": 1.2209302325581394e-07,
|
| 392 |
+
"logits/chosen": -2.5561137199401855,
|
| 393 |
+
"logits/rejected": -2.5325522422790527,
|
| 394 |
+
"logps/chosen": -89.1368637084961,
|
| 395 |
+
"logps/rejected": -88.74254608154297,
|
| 396 |
+
"loss": 0.6316,
|
| 397 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 398 |
+
"rewards/chosen": 0.3368709683418274,
|
| 399 |
+
"rewards/margins": 0.5684143900871277,
|
| 400 |
+
"rewards/rejected": -0.23154346644878387,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8362369337979094,
|
| 405 |
+
"grad_norm": 200.36993244283045,
|
| 406 |
+
"learning_rate": 1.1821705426356587e-07,
|
| 407 |
+
"logits/chosen": -2.5511903762817383,
|
| 408 |
+
"logits/rejected": -2.5139944553375244,
|
| 409 |
+
"logps/chosen": -86.218994140625,
|
| 410 |
+
"logps/rejected": -80.59002685546875,
|
| 411 |
+
"loss": 0.6428,
|
| 412 |
+
"rewards/accuracies": 0.34375,
|
| 413 |
+
"rewards/chosen": 0.2038179188966751,
|
| 414 |
+
"rewards/margins": 0.3808489143848419,
|
| 415 |
+
"rewards/rejected": -0.17703098058700562,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8710801393728222,
|
| 420 |
+
"grad_norm": 220.61998054567871,
|
| 421 |
+
"learning_rate": 1.1434108527131783e-07,
|
| 422 |
+
"logits/chosen": -2.566251516342163,
|
| 423 |
+
"logits/rejected": -2.52921199798584,
|
| 424 |
+
"logps/chosen": -95.03993225097656,
|
| 425 |
+
"logps/rejected": -90.61654663085938,
|
| 426 |
+
"loss": 0.5848,
|
| 427 |
+
"rewards/accuracies": 0.39375001192092896,
|
| 428 |
+
"rewards/chosen": 0.15963879227638245,
|
| 429 |
+
"rewards/margins": 0.42108219861984253,
|
| 430 |
+
"rewards/rejected": -0.26144343614578247,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.9059233449477352,
|
| 435 |
+
"grad_norm": 172.2124811551973,
|
| 436 |
+
"learning_rate": 1.1046511627906976e-07,
|
| 437 |
+
"logits/chosen": -2.4705214500427246,
|
| 438 |
+
"logits/rejected": -2.48685622215271,
|
| 439 |
+
"logps/chosen": -58.55452346801758,
|
| 440 |
+
"logps/rejected": -65.9000015258789,
|
| 441 |
+
"loss": 0.6216,
|
| 442 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 443 |
+
"rewards/chosen": 0.19298657774925232,
|
| 444 |
+
"rewards/margins": 0.29895681142807007,
|
| 445 |
+
"rewards/rejected": -0.10597021877765656,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.9407665505226481,
|
| 450 |
+
"grad_norm": 221.01197823196617,
|
| 451 |
+
"learning_rate": 1.0658914728682171e-07,
|
| 452 |
+
"logits/chosen": -2.5705666542053223,
|
| 453 |
+
"logits/rejected": -2.5713424682617188,
|
| 454 |
+
"logps/chosen": -68.17798614501953,
|
| 455 |
+
"logps/rejected": -83.17510223388672,
|
| 456 |
+
"loss": 0.6039,
|
| 457 |
+
"rewards/accuracies": 0.34375,
|
| 458 |
+
"rewards/chosen": 0.2738538086414337,
|
| 459 |
+
"rewards/margins": 0.5247575640678406,
|
| 460 |
+
"rewards/rejected": -0.2509037256240845,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.975609756097561,
|
| 465 |
+
"grad_norm": 201.10787443979413,
|
| 466 |
+
"learning_rate": 1.0271317829457364e-07,
|
| 467 |
+
"logits/chosen": -2.4812636375427246,
|
| 468 |
+
"logits/rejected": -2.4561610221862793,
|
| 469 |
+
"logps/chosen": -67.36260986328125,
|
| 470 |
+
"logps/rejected": -71.26799011230469,
|
| 471 |
+
"loss": 0.6039,
|
| 472 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 473 |
+
"rewards/chosen": 0.18398845195770264,
|
| 474 |
+
"rewards/margins": 0.3530852496623993,
|
| 475 |
+
"rewards/rejected": -0.16909679770469666,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.0,
|
| 480 |
+
"step": 287,
|
| 481 |
+
"total_flos": 0.0,
|
| 482 |
+
"train_loss": 0.6390685353960309,
|
| 483 |
+
"train_runtime": 3341.9408,
|
| 484 |
+
"train_samples_per_second": 5.488,
|
| 485 |
+
"train_steps_per_second": 0.086
|
| 486 |
+
}
|
| 487 |
+
],
|
| 488 |
+
"logging_steps": 10,
|
| 489 |
+
"max_steps": 287,
|
| 490 |
+
"num_input_tokens_seen": 0,
|
| 491 |
+
"num_train_epochs": 1,
|
| 492 |
+
"save_steps": 100,
|
| 493 |
+
"stateful_callbacks": {
|
| 494 |
+
"TrainerControl": {
|
| 495 |
+
"args": {
|
| 496 |
+
"should_epoch_stop": false,
|
| 497 |
+
"should_evaluate": false,
|
| 498 |
+
"should_log": false,
|
| 499 |
+
"should_save": true,
|
| 500 |
+
"should_training_stop": true
|
| 501 |
+
},
|
| 502 |
+
"attributes": {}
|
| 503 |
+
}
|
| 504 |
+
},
|
| 505 |
+
"total_flos": 0.0,
|
| 506 |
+
"train_batch_size": 8,
|
| 507 |
+
"trial_name": null,
|
| 508 |
+
"trial_params": null
|
| 509 |
+
}
|