Model save
Browse files- README.md +61 -0
- all_results.json +9 -0
- generation_config.json +8 -0
- train_results.json +9 -0
- trainer_state.json +447 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: gemma
|
| 4 |
+
base_model: google/gemma-2-9b-it
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- simpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: simpo-train-small-correct
|
| 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 |
+
# simpo-train-small-correct
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it) on an unknown dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 8e-07
|
| 39 |
+
- train_batch_size: 2
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- gradient_accumulation_steps: 16
|
| 45 |
+
- total_train_batch_size: 128
|
| 46 |
+
- total_eval_batch_size: 16
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 1
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- Transformers 4.44.2
|
| 59 |
+
- Pytorch 2.7.0+cu128
|
| 60 |
+
- Datasets 2.18.0
|
| 61 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.995910949568378,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 4.128190910729178,
|
| 5 |
+
"train_runtime": 2318.2745,
|
| 6 |
+
"train_samples": 17605,
|
| 7 |
+
"train_samples_per_second": 7.594,
|
| 8 |
+
"train_steps_per_second": 0.059
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 2,
|
| 4 |
+
"cache_implementation": "hybrid",
|
| 5 |
+
"eos_token_id": 1,
|
| 6 |
+
"pad_token_id": 0,
|
| 7 |
+
"transformers_version": "4.44.2"
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.995910949568378,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 4.128190910729178,
|
| 5 |
+
"train_runtime": 2318.2745,
|
| 6 |
+
"train_samples": 17605,
|
| 7 |
+
"train_samples_per_second": 7.594,
|
| 8 |
+
"train_steps_per_second": 0.059
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,447 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.995910949568378,
|
| 5 |
+
"eval_steps": 400,
|
| 6 |
+
"global_step": 137,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03634711494775102,
|
| 13 |
+
"grad_norm": 32.58869576990809,
|
| 14 |
+
"learning_rate": 2.857142857142857e-07,
|
| 15 |
+
"logits/chosen": -11.543076515197754,
|
| 16 |
+
"logits/rejected": -11.31420612335205,
|
| 17 |
+
"logps/chosen": -0.5522164106369019,
|
| 18 |
+
"logps/rejected": -0.5735588669776917,
|
| 19 |
+
"loss": 4.805,
|
| 20 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 21 |
+
"rewards/chosen": -5.522164344787598,
|
| 22 |
+
"rewards/margins": 0.21342459321022034,
|
| 23 |
+
"rewards/rejected": -5.735588550567627,
|
| 24 |
+
"step": 5
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.07269422989550205,
|
| 28 |
+
"grad_norm": 34.3946259201971,
|
| 29 |
+
"learning_rate": 5.714285714285714e-07,
|
| 30 |
+
"logits/chosen": -10.83501148223877,
|
| 31 |
+
"logits/rejected": -10.637781143188477,
|
| 32 |
+
"logps/chosen": -0.6140329241752625,
|
| 33 |
+
"logps/rejected": -0.6750808954238892,
|
| 34 |
+
"loss": 4.7363,
|
| 35 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 36 |
+
"rewards/chosen": -6.140329837799072,
|
| 37 |
+
"rewards/margins": 0.6104797124862671,
|
| 38 |
+
"rewards/rejected": -6.7508087158203125,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.10904134484325306,
|
| 43 |
+
"grad_norm": 41.73056692404051,
|
| 44 |
+
"learning_rate": 7.998695344323425e-07,
|
| 45 |
+
"logits/chosen": -10.648634910583496,
|
| 46 |
+
"logits/rejected": -10.217233657836914,
|
| 47 |
+
"logps/chosen": -0.5939905047416687,
|
| 48 |
+
"logps/rejected": -0.6135314702987671,
|
| 49 |
+
"loss": 4.8553,
|
| 50 |
+
"rewards/accuracies": 0.731249988079071,
|
| 51 |
+
"rewards/chosen": -5.939904689788818,
|
| 52 |
+
"rewards/margins": 0.1954106092453003,
|
| 53 |
+
"rewards/rejected": -6.135315418243408,
|
| 54 |
+
"step": 15
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.1453884597910041,
|
| 58 |
+
"grad_norm": 36.02519008680044,
|
| 59 |
+
"learning_rate": 7.953121695121394e-07,
|
| 60 |
+
"logits/chosen": -10.66395378112793,
|
| 61 |
+
"logits/rejected": -10.348527908325195,
|
| 62 |
+
"logps/chosen": -0.5527507066726685,
|
| 63 |
+
"logps/rejected": -0.5806955099105835,
|
| 64 |
+
"loss": 4.6386,
|
| 65 |
+
"rewards/accuracies": 0.699999988079071,
|
| 66 |
+
"rewards/chosen": -5.527507781982422,
|
| 67 |
+
"rewards/margins": 0.27944836020469666,
|
| 68 |
+
"rewards/rejected": -5.806955337524414,
|
| 69 |
+
"step": 20
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.18173557473875512,
|
| 73 |
+
"grad_norm": 40.07097906112277,
|
| 74 |
+
"learning_rate": 7.843163833184991e-07,
|
| 75 |
+
"logits/chosen": -10.465521812438965,
|
| 76 |
+
"logits/rejected": -10.2030668258667,
|
| 77 |
+
"logps/chosen": -0.5359245538711548,
|
| 78 |
+
"logps/rejected": -0.5927993655204773,
|
| 79 |
+
"loss": 4.5869,
|
| 80 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 81 |
+
"rewards/chosen": -5.359244346618652,
|
| 82 |
+
"rewards/margins": 0.5687500834465027,
|
| 83 |
+
"rewards/rejected": -5.927994728088379,
|
| 84 |
+
"step": 25
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.21808268968650613,
|
| 88 |
+
"grad_norm": 34.259554554601195,
|
| 89 |
+
"learning_rate": 7.670612634414511e-07,
|
| 90 |
+
"logits/chosen": -10.547082901000977,
|
| 91 |
+
"logits/rejected": -10.052377700805664,
|
| 92 |
+
"logps/chosen": -0.5360420346260071,
|
| 93 |
+
"logps/rejected": -0.5889440774917603,
|
| 94 |
+
"loss": 4.5463,
|
| 95 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 96 |
+
"rewards/chosen": -5.360420227050781,
|
| 97 |
+
"rewards/margins": 0.5290209054946899,
|
| 98 |
+
"rewards/rejected": -5.889441013336182,
|
| 99 |
+
"step": 30
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.25442980463425713,
|
| 103 |
+
"grad_norm": 30.666733170155783,
|
| 104 |
+
"learning_rate": 7.438278427948805e-07,
|
| 105 |
+
"logits/chosen": -10.336666107177734,
|
| 106 |
+
"logits/rejected": -10.034899711608887,
|
| 107 |
+
"logps/chosen": -0.6049242615699768,
|
| 108 |
+
"logps/rejected": -0.6828798055648804,
|
| 109 |
+
"loss": 4.3807,
|
| 110 |
+
"rewards/accuracies": 0.831250011920929,
|
| 111 |
+
"rewards/chosen": -6.0492424964904785,
|
| 112 |
+
"rewards/margins": 0.7795557379722595,
|
| 113 |
+
"rewards/rejected": -6.828797817230225,
|
| 114 |
+
"step": 35
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.2907769195820082,
|
| 118 |
+
"grad_norm": 37.89264733124544,
|
| 119 |
+
"learning_rate": 7.149945224533862e-07,
|
| 120 |
+
"logits/chosen": -10.575347900390625,
|
| 121 |
+
"logits/rejected": -10.193809509277344,
|
| 122 |
+
"logps/chosen": -0.6202062964439392,
|
| 123 |
+
"logps/rejected": -0.6880999803543091,
|
| 124 |
+
"loss": 4.4737,
|
| 125 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 126 |
+
"rewards/chosen": -6.20206356048584,
|
| 127 |
+
"rewards/margins": 0.678936779499054,
|
| 128 |
+
"rewards/rejected": -6.880999565124512,
|
| 129 |
+
"step": 40
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.3271240345297592,
|
| 133 |
+
"grad_norm": 36.08127980591517,
|
| 134 |
+
"learning_rate": 6.810309086608129e-07,
|
| 135 |
+
"logits/chosen": -11.216996192932129,
|
| 136 |
+
"logits/rejected": -10.75574016571045,
|
| 137 |
+
"logps/chosen": -0.6237494349479675,
|
| 138 |
+
"logps/rejected": -0.687574028968811,
|
| 139 |
+
"loss": 4.4138,
|
| 140 |
+
"rewards/accuracies": 0.731249988079071,
|
| 141 |
+
"rewards/chosen": -6.237493991851807,
|
| 142 |
+
"rewards/margins": 0.6382461190223694,
|
| 143 |
+
"rewards/rejected": -6.875740051269531,
|
| 144 |
+
"step": 45
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.36347114947751025,
|
| 148 |
+
"grad_norm": 34.2641539929766,
|
| 149 |
+
"learning_rate": 6.424901643866552e-07,
|
| 150 |
+
"logits/chosen": -11.69914722442627,
|
| 151 |
+
"logits/rejected": -11.246435165405273,
|
| 152 |
+
"logps/chosen": -0.6562625169754028,
|
| 153 |
+
"logps/rejected": -0.7224141359329224,
|
| 154 |
+
"loss": 4.3367,
|
| 155 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 156 |
+
"rewards/chosen": -6.562624931335449,
|
| 157 |
+
"rewards/margins": 0.6615163087844849,
|
| 158 |
+
"rewards/rejected": -7.2241411209106445,
|
| 159 |
+
"step": 50
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.39981826442526125,
|
| 163 |
+
"grad_norm": 36.53529303407606,
|
| 164 |
+
"learning_rate": 6e-07,
|
| 165 |
+
"logits/chosen": -11.810195922851562,
|
| 166 |
+
"logits/rejected": -11.363134384155273,
|
| 167 |
+
"logps/chosen": -0.7002917528152466,
|
| 168 |
+
"logps/rejected": -0.778505802154541,
|
| 169 |
+
"loss": 4.3706,
|
| 170 |
+
"rewards/accuracies": 0.699999988079071,
|
| 171 |
+
"rewards/chosen": -7.002917289733887,
|
| 172 |
+
"rewards/margins": 0.7821400761604309,
|
| 173 |
+
"rewards/rejected": -7.785058498382568,
|
| 174 |
+
"step": 55
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.43616537937301225,
|
| 178 |
+
"grad_norm": 35.761908817299414,
|
| 179 |
+
"learning_rate": 5.542524497952543e-07,
|
| 180 |
+
"logits/chosen": -11.559731483459473,
|
| 181 |
+
"logits/rejected": -11.183916091918945,
|
| 182 |
+
"logps/chosen": -0.7286056280136108,
|
| 183 |
+
"logps/rejected": -0.8559492230415344,
|
| 184 |
+
"loss": 4.2217,
|
| 185 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 186 |
+
"rewards/chosen": -7.2860565185546875,
|
| 187 |
+
"rewards/margins": 1.273435354232788,
|
| 188 |
+
"rewards/rejected": -8.559491157531738,
|
| 189 |
+
"step": 60
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.4725124943207633,
|
| 193 |
+
"grad_norm": 231.37525001147708,
|
| 194 |
+
"learning_rate": 5.059926008786647e-07,
|
| 195 |
+
"logits/chosen": -11.819048881530762,
|
| 196 |
+
"logits/rejected": -11.407302856445312,
|
| 197 |
+
"logps/chosen": -0.734784722328186,
|
| 198 |
+
"logps/rejected": -0.8366082906723022,
|
| 199 |
+
"loss": 4.1901,
|
| 200 |
+
"rewards/accuracies": 0.75,
|
| 201 |
+
"rewards/chosen": -7.347847938537598,
|
| 202 |
+
"rewards/margins": 1.0182350873947144,
|
| 203 |
+
"rewards/rejected": -8.366083145141602,
|
| 204 |
+
"step": 65
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.5088596092685143,
|
| 208 |
+
"grad_norm": 51.233160095021326,
|
| 209 |
+
"learning_rate": 4.5600645798745166e-07,
|
| 210 |
+
"logits/chosen": -11.959329605102539,
|
| 211 |
+
"logits/rejected": -11.445878028869629,
|
| 212 |
+
"logps/chosen": -0.7511547803878784,
|
| 213 |
+
"logps/rejected": -0.8529723882675171,
|
| 214 |
+
"loss": 4.2912,
|
| 215 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 216 |
+
"rewards/chosen": -7.511547088623047,
|
| 217 |
+
"rewards/margins": 1.0181769132614136,
|
| 218 |
+
"rewards/rejected": -8.52972412109375,
|
| 219 |
+
"step": 70
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.5452067242162654,
|
| 223 |
+
"grad_norm": 59.733803732439824,
|
| 224 |
+
"learning_rate": 4.051081418863895e-07,
|
| 225 |
+
"logits/chosen": -12.409022331237793,
|
| 226 |
+
"logits/rejected": -11.8023042678833,
|
| 227 |
+
"logps/chosen": -0.8404130935668945,
|
| 228 |
+
"logps/rejected": -0.9404484629631042,
|
| 229 |
+
"loss": 4.0174,
|
| 230 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 231 |
+
"rewards/chosen": -8.404130935668945,
|
| 232 |
+
"rewards/margins": 1.0003544092178345,
|
| 233 |
+
"rewards/rejected": -9.404484748840332,
|
| 234 |
+
"step": 75
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.5815538391640164,
|
| 238 |
+
"grad_norm": 60.192585986207376,
|
| 239 |
+
"learning_rate": 3.541266298406398e-07,
|
| 240 |
+
"logits/chosen": -12.791674613952637,
|
| 241 |
+
"logits/rejected": -12.519620895385742,
|
| 242 |
+
"logps/chosen": -0.8719121217727661,
|
| 243 |
+
"logps/rejected": -1.016026258468628,
|
| 244 |
+
"loss": 4.1563,
|
| 245 |
+
"rewards/accuracies": 0.737500011920929,
|
| 246 |
+
"rewards/chosen": -8.719121932983398,
|
| 247 |
+
"rewards/margins": 1.4411423206329346,
|
| 248 |
+
"rewards/rejected": -10.160263061523438,
|
| 249 |
+
"step": 80
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.6179009541117674,
|
| 253 |
+
"grad_norm": 63.01387669368681,
|
| 254 |
+
"learning_rate": 3.0389225412181565e-07,
|
| 255 |
+
"logits/chosen": -13.761955261230469,
|
| 256 |
+
"logits/rejected": -13.322006225585938,
|
| 257 |
+
"logps/chosen": -0.8667750358581543,
|
| 258 |
+
"logps/rejected": -1.0224027633666992,
|
| 259 |
+
"loss": 4.0181,
|
| 260 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 261 |
+
"rewards/chosen": -8.667750358581543,
|
| 262 |
+
"rewards/margins": 1.5562773942947388,
|
| 263 |
+
"rewards/rejected": -10.224027633666992,
|
| 264 |
+
"step": 85
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.6542480690595184,
|
| 268 |
+
"grad_norm": 58.79093252453101,
|
| 269 |
+
"learning_rate": 2.5522317844515273e-07,
|
| 270 |
+
"logits/chosen": -13.937658309936523,
|
| 271 |
+
"logits/rejected": -13.547172546386719,
|
| 272 |
+
"logps/chosen": -0.9299012422561646,
|
| 273 |
+
"logps/rejected": -1.087494134902954,
|
| 274 |
+
"loss": 3.8751,
|
| 275 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 276 |
+
"rewards/chosen": -9.29901123046875,
|
| 277 |
+
"rewards/margins": 1.5759302377700806,
|
| 278 |
+
"rewards/rejected": -10.874940872192383,
|
| 279 |
+
"step": 90
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.6905951840072694,
|
| 283 |
+
"grad_norm": 90.56773515906556,
|
| 284 |
+
"learning_rate": 2.0891207259509476e-07,
|
| 285 |
+
"logits/chosen": -14.276998519897461,
|
| 286 |
+
"logits/rejected": -13.726056098937988,
|
| 287 |
+
"logps/chosen": -0.9327268600463867,
|
| 288 |
+
"logps/rejected": -1.1215860843658447,
|
| 289 |
+
"loss": 3.8327,
|
| 290 |
+
"rewards/accuracies": 0.793749988079071,
|
| 291 |
+
"rewards/chosen": -9.327268600463867,
|
| 292 |
+
"rewards/margins": 1.8885917663574219,
|
| 293 |
+
"rewards/rejected": -11.215860366821289,
|
| 294 |
+
"step": 95
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.7269422989550205,
|
| 298 |
+
"grad_norm": 68.83137467247752,
|
| 299 |
+
"learning_rate": 1.6571320226872206e-07,
|
| 300 |
+
"logits/chosen": -15.408729553222656,
|
| 301 |
+
"logits/rejected": -14.846705436706543,
|
| 302 |
+
"logps/chosen": -0.994869589805603,
|
| 303 |
+
"logps/rejected": -1.1596581935882568,
|
| 304 |
+
"loss": 3.8063,
|
| 305 |
+
"rewards/accuracies": 0.762499988079071,
|
| 306 |
+
"rewards/chosen": -9.948695182800293,
|
| 307 |
+
"rewards/margins": 1.64788818359375,
|
| 308 |
+
"rewards/rejected": -11.596583366394043,
|
| 309 |
+
"step": 100
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.7632894139027715,
|
| 313 |
+
"grad_norm": 81.97650507265567,
|
| 314 |
+
"learning_rate": 1.2633014440382787e-07,
|
| 315 |
+
"logits/chosen": -15.801300048828125,
|
| 316 |
+
"logits/rejected": -15.365028381347656,
|
| 317 |
+
"logps/chosen": -1.0137146711349487,
|
| 318 |
+
"logps/rejected": -1.173123836517334,
|
| 319 |
+
"loss": 3.7038,
|
| 320 |
+
"rewards/accuracies": 0.768750011920929,
|
| 321 |
+
"rewards/chosen": -10.137146949768066,
|
| 322 |
+
"rewards/margins": 1.5940909385681152,
|
| 323 |
+
"rewards/rejected": -11.731237411499023,
|
| 324 |
+
"step": 105
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.7996365288505225,
|
| 328 |
+
"grad_norm": 76.7946168478378,
|
| 329 |
+
"learning_rate": 9.14043280712228e-08,
|
| 330 |
+
"logits/chosen": -16.20096206665039,
|
| 331 |
+
"logits/rejected": -15.696843147277832,
|
| 332 |
+
"logps/chosen": -1.1373329162597656,
|
| 333 |
+
"logps/rejected": -1.354536771774292,
|
| 334 |
+
"loss": 3.6719,
|
| 335 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 336 |
+
"rewards/chosen": -11.373329162597656,
|
| 337 |
+
"rewards/margins": 2.1720378398895264,
|
| 338 |
+
"rewards/rejected": -13.545367240905762,
|
| 339 |
+
"step": 110
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.8359836437982735,
|
| 343 |
+
"grad_norm": 85.18089495737438,
|
| 344 |
+
"learning_rate": 6.150458756494239e-08,
|
| 345 |
+
"logits/chosen": -16.4575138092041,
|
| 346 |
+
"logits/rejected": -16.271873474121094,
|
| 347 |
+
"logps/chosen": -1.0763649940490723,
|
| 348 |
+
"logps/rejected": -1.265742540359497,
|
| 349 |
+
"loss": 3.621,
|
| 350 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 351 |
+
"rewards/chosen": -10.763648986816406,
|
| 352 |
+
"rewards/margins": 1.8937755823135376,
|
| 353 |
+
"rewards/rejected": -12.657424926757812,
|
| 354 |
+
"step": 115
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.8723307587460245,
|
| 358 |
+
"grad_norm": 80.36432752317246,
|
| 359 |
+
"learning_rate": 3.711789783843522e-08,
|
| 360 |
+
"logits/chosen": -16.593286514282227,
|
| 361 |
+
"logits/rejected": -16.234304428100586,
|
| 362 |
+
"logps/chosen": -1.041198492050171,
|
| 363 |
+
"logps/rejected": -1.273951768875122,
|
| 364 |
+
"loss": 3.481,
|
| 365 |
+
"rewards/accuracies": 0.800000011920929,
|
| 366 |
+
"rewards/chosen": -10.411985397338867,
|
| 367 |
+
"rewards/margins": 2.327531337738037,
|
| 368 |
+
"rewards/rejected": -12.739517211914062,
|
| 369 |
+
"step": 120
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.9086778736937755,
|
| 373 |
+
"grad_norm": 81.72943922815173,
|
| 374 |
+
"learning_rate": 1.8641443178027784e-08,
|
| 375 |
+
"logits/chosen": -16.43305206298828,
|
| 376 |
+
"logits/rejected": -16.25759506225586,
|
| 377 |
+
"logps/chosen": -1.142892599105835,
|
| 378 |
+
"logps/rejected": -1.3480346202850342,
|
| 379 |
+
"loss": 3.6377,
|
| 380 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 381 |
+
"rewards/chosen": -11.428924560546875,
|
| 382 |
+
"rewards/margins": 2.051419734954834,
|
| 383 |
+
"rewards/rejected": -13.480344772338867,
|
| 384 |
+
"step": 125
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.9450249886415266,
|
| 388 |
+
"grad_norm": 74.32925754994847,
|
| 389 |
+
"learning_rate": 6.376148290617145e-09,
|
| 390 |
+
"logits/chosen": -16.8632869720459,
|
| 391 |
+
"logits/rejected": -16.60898208618164,
|
| 392 |
+
"logps/chosen": -1.1211764812469482,
|
| 393 |
+
"logps/rejected": -1.3164139986038208,
|
| 394 |
+
"loss": 3.4601,
|
| 395 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 396 |
+
"rewards/chosen": -11.21176528930664,
|
| 397 |
+
"rewards/margins": 1.9523746967315674,
|
| 398 |
+
"rewards/rejected": -13.164140701293945,
|
| 399 |
+
"step": 130
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.9813721035892776,
|
| 403 |
+
"grad_norm": 82.1235499731324,
|
| 404 |
+
"learning_rate": 5.217771643080127e-10,
|
| 405 |
+
"logits/chosen": -16.644031524658203,
|
| 406 |
+
"logits/rejected": -16.64177131652832,
|
| 407 |
+
"logps/chosen": -1.1249492168426514,
|
| 408 |
+
"logps/rejected": -1.297629714012146,
|
| 409 |
+
"loss": 3.5925,
|
| 410 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 411 |
+
"rewards/chosen": -11.249493598937988,
|
| 412 |
+
"rewards/margins": 1.726805329322815,
|
| 413 |
+
"rewards/rejected": -12.976297378540039,
|
| 414 |
+
"step": 135
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.995910949568378,
|
| 418 |
+
"step": 137,
|
| 419 |
+
"total_flos": 0.0,
|
| 420 |
+
"train_loss": 4.128190910729178,
|
| 421 |
+
"train_runtime": 2318.2745,
|
| 422 |
+
"train_samples_per_second": 7.594,
|
| 423 |
+
"train_steps_per_second": 0.059
|
| 424 |
+
}
|
| 425 |
+
],
|
| 426 |
+
"logging_steps": 5,
|
| 427 |
+
"max_steps": 137,
|
| 428 |
+
"num_input_tokens_seen": 0,
|
| 429 |
+
"num_train_epochs": 1,
|
| 430 |
+
"save_steps": 1000000,
|
| 431 |
+
"stateful_callbacks": {
|
| 432 |
+
"TrainerControl": {
|
| 433 |
+
"args": {
|
| 434 |
+
"should_epoch_stop": false,
|
| 435 |
+
"should_evaluate": false,
|
| 436 |
+
"should_log": false,
|
| 437 |
+
"should_save": true,
|
| 438 |
+
"should_training_stop": true
|
| 439 |
+
},
|
| 440 |
+
"attributes": {}
|
| 441 |
+
}
|
| 442 |
+
},
|
| 443 |
+
"total_flos": 0.0,
|
| 444 |
+
"train_batch_size": 2,
|
| 445 |
+
"trial_name": null,
|
| 446 |
+
"trial_params": null
|
| 447 |
+
}
|