Model save
Browse files- README.md +58 -0
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +475 -0
README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: qwen-2.5-7b-instruct-sft-qlora-countdown-search-1k
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- sft
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for qwen-2.5-7b-instruct-sft-qlora-countdown-search-1k
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="chloeli/qwen-2.5-7b-instruct-sft-qlora-countdown-search-1k", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/chloeli/huggingface/runs/bvttf6gb)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with SFT.
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.15.2
|
| 38 |
+
- Transformers: 4.49.0
|
| 39 |
+
- Pytorch: 2.6.0
|
| 40 |
+
- Datasets: 3.3.2
|
| 41 |
+
- Tokenizers: 0.21.0
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
Cite TRL as:
|
| 48 |
+
|
| 49 |
+
```bibtex
|
| 50 |
+
@misc{vonwerra2022trl,
|
| 51 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 52 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
|
| 53 |
+
year = 2020,
|
| 54 |
+
journal = {GitHub repository},
|
| 55 |
+
publisher = {GitHub},
|
| 56 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 57 |
+
}
|
| 58 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 8.539757562822656e+16,
|
| 3 |
+
"train_loss": 0.07947162318229675,
|
| 4 |
+
"train_runtime": 16205.2206,
|
| 5 |
+
"train_samples": 1000,
|
| 6 |
+
"train_samples_per_second": 0.062,
|
| 7 |
+
"train_steps_per_second": 0.008
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 8.539757562822656e+16,
|
| 3 |
+
"train_loss": 0.07947162318229675,
|
| 4 |
+
"train_runtime": 16205.2206,
|
| 5 |
+
"train_samples": 1000,
|
| 6 |
+
"train_samples_per_second": 0.062,
|
| 7 |
+
"train_steps_per_second": 0.008
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,475 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 5,
|
| 6 |
+
"global_step": 125,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.008,
|
| 13 |
+
"grad_norm": 0.2523050010204315,
|
| 14 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 15 |
+
"loss": 0.4523,
|
| 16 |
+
"mean_token_accuracy": 0.9231184720993042,
|
| 17 |
+
"step": 1
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.04,
|
| 21 |
+
"grad_norm": 0.24141842126846313,
|
| 22 |
+
"learning_rate": 7.692307692307693e-05,
|
| 23 |
+
"loss": 0.4396,
|
| 24 |
+
"mean_token_accuracy": 0.9300202205777168,
|
| 25 |
+
"step": 5
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.04,
|
| 29 |
+
"eval_loss": 0.38802939653396606,
|
| 30 |
+
"eval_mean_token_accuracy": 0.931274386882782,
|
| 31 |
+
"eval_runtime": 572.2225,
|
| 32 |
+
"eval_samples_per_second": 1.748,
|
| 33 |
+
"eval_steps_per_second": 0.218,
|
| 34 |
+
"step": 5
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 0.08,
|
| 38 |
+
"grad_norm": 0.16969740390777588,
|
| 39 |
+
"learning_rate": 0.00015384615384615385,
|
| 40 |
+
"loss": 0.3426,
|
| 41 |
+
"mean_token_accuracy": 0.93111452460289,
|
| 42 |
+
"step": 10
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.08,
|
| 46 |
+
"eval_loss": 0.24519848823547363,
|
| 47 |
+
"eval_mean_token_accuracy": 0.943490716934204,
|
| 48 |
+
"eval_runtime": 572.2716,
|
| 49 |
+
"eval_samples_per_second": 1.747,
|
| 50 |
+
"eval_steps_per_second": 0.218,
|
| 51 |
+
"step": 10
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.12,
|
| 55 |
+
"grad_norm": 0.15904344618320465,
|
| 56 |
+
"learning_rate": 0.00019984268150178167,
|
| 57 |
+
"loss": 0.1922,
|
| 58 |
+
"mean_token_accuracy": 0.9507227122783661,
|
| 59 |
+
"step": 15
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.12,
|
| 63 |
+
"eval_loss": 0.12690386176109314,
|
| 64 |
+
"eval_mean_token_accuracy": 0.962551230430603,
|
| 65 |
+
"eval_runtime": 572.2269,
|
| 66 |
+
"eval_samples_per_second": 1.748,
|
| 67 |
+
"eval_steps_per_second": 0.218,
|
| 68 |
+
"step": 15
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.16,
|
| 72 |
+
"grad_norm": 0.12561501562595367,
|
| 73 |
+
"learning_rate": 0.00019807852804032305,
|
| 74 |
+
"loss": 0.094,
|
| 75 |
+
"mean_token_accuracy": 0.9698278903961182,
|
| 76 |
+
"step": 20
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.16,
|
| 80 |
+
"eval_loss": 0.07218066602945328,
|
| 81 |
+
"eval_mean_token_accuracy": 0.9754553689956665,
|
| 82 |
+
"eval_runtime": 572.3734,
|
| 83 |
+
"eval_samples_per_second": 1.747,
|
| 84 |
+
"eval_steps_per_second": 0.218,
|
| 85 |
+
"step": 20
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.2,
|
| 89 |
+
"grad_norm": 0.09696544706821442,
|
| 90 |
+
"learning_rate": 0.00019438833303083678,
|
| 91 |
+
"loss": 0.0578,
|
| 92 |
+
"mean_token_accuracy": 0.9804269909858704,
|
| 93 |
+
"step": 25
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.2,
|
| 97 |
+
"eval_loss": 0.06172483041882515,
|
| 98 |
+
"eval_mean_token_accuracy": 0.9775349216461182,
|
| 99 |
+
"eval_runtime": 571.3615,
|
| 100 |
+
"eval_samples_per_second": 1.75,
|
| 101 |
+
"eval_steps_per_second": 0.219,
|
| 102 |
+
"step": 25
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.24,
|
| 106 |
+
"grad_norm": 0.07436518371105194,
|
| 107 |
+
"learning_rate": 0.00018884456359788724,
|
| 108 |
+
"loss": 0.0474,
|
| 109 |
+
"mean_token_accuracy": 0.9817552506923676,
|
| 110 |
+
"step": 30
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.24,
|
| 114 |
+
"eval_loss": 0.051653869450092316,
|
| 115 |
+
"eval_mean_token_accuracy": 0.97948202419281,
|
| 116 |
+
"eval_runtime": 572.6759,
|
| 117 |
+
"eval_samples_per_second": 1.746,
|
| 118 |
+
"eval_steps_per_second": 0.218,
|
| 119 |
+
"step": 30
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.28,
|
| 123 |
+
"grad_norm": 0.058379314839839935,
|
| 124 |
+
"learning_rate": 0.00018155608689592604,
|
| 125 |
+
"loss": 0.0473,
|
| 126 |
+
"mean_token_accuracy": 0.9801956593990326,
|
| 127 |
+
"step": 35
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.28,
|
| 131 |
+
"eval_loss": 0.04774034023284912,
|
| 132 |
+
"eval_mean_token_accuracy": 0.9803278684616089,
|
| 133 |
+
"eval_runtime": 572.9678,
|
| 134 |
+
"eval_samples_per_second": 1.745,
|
| 135 |
+
"eval_steps_per_second": 0.218,
|
| 136 |
+
"step": 35
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.32,
|
| 140 |
+
"grad_norm": 0.05066407099366188,
|
| 141 |
+
"learning_rate": 0.0001726660322034027,
|
| 142 |
+
"loss": 0.0471,
|
| 143 |
+
"mean_token_accuracy": 0.9803119540214539,
|
| 144 |
+
"step": 40
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.32,
|
| 148 |
+
"eval_loss": 0.04579387605190277,
|
| 149 |
+
"eval_mean_token_accuracy": 0.980647020816803,
|
| 150 |
+
"eval_runtime": 572.7685,
|
| 151 |
+
"eval_samples_per_second": 1.746,
|
| 152 |
+
"eval_steps_per_second": 0.218,
|
| 153 |
+
"step": 40
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.36,
|
| 157 |
+
"grad_norm": 0.05025569722056389,
|
| 158 |
+
"learning_rate": 0.00016234898018587337,
|
| 159 |
+
"loss": 0.0423,
|
| 160 |
+
"mean_token_accuracy": 0.9822601974010468,
|
| 161 |
+
"step": 45
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.36,
|
| 165 |
+
"eval_loss": 0.04449079558253288,
|
| 166 |
+
"eval_mean_token_accuracy": 0.981070707321167,
|
| 167 |
+
"eval_runtime": 572.4781,
|
| 168 |
+
"eval_samples_per_second": 1.747,
|
| 169 |
+
"eval_steps_per_second": 0.218,
|
| 170 |
+
"step": 45
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.4,
|
| 174 |
+
"grad_norm": 0.03607247769832611,
|
| 175 |
+
"learning_rate": 0.00015080753452465296,
|
| 176 |
+
"loss": 0.0465,
|
| 177 |
+
"mean_token_accuracy": 0.9798099577426911,
|
| 178 |
+
"step": 50
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.4,
|
| 182 |
+
"eval_loss": 0.04439844563603401,
|
| 183 |
+
"eval_mean_token_accuracy": 0.9811892700195313,
|
| 184 |
+
"eval_runtime": 572.6901,
|
| 185 |
+
"eval_samples_per_second": 1.746,
|
| 186 |
+
"eval_steps_per_second": 0.218,
|
| 187 |
+
"step": 50
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.44,
|
| 191 |
+
"grad_norm": 0.037045057862997055,
|
| 192 |
+
"learning_rate": 0.000138268343236509,
|
| 193 |
+
"loss": 0.0431,
|
| 194 |
+
"mean_token_accuracy": 0.9816874027252197,
|
| 195 |
+
"step": 55
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.44,
|
| 199 |
+
"eval_loss": 0.04384024068713188,
|
| 200 |
+
"eval_mean_token_accuracy": 0.9812552394866944,
|
| 201 |
+
"eval_runtime": 572.4389,
|
| 202 |
+
"eval_samples_per_second": 1.747,
|
| 203 |
+
"eval_steps_per_second": 0.218,
|
| 204 |
+
"step": 55
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.48,
|
| 208 |
+
"grad_norm": 0.03762717917561531,
|
| 209 |
+
"learning_rate": 0.0001249776478167227,
|
| 210 |
+
"loss": 0.0437,
|
| 211 |
+
"mean_token_accuracy": 0.9811290085315705,
|
| 212 |
+
"step": 60
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.48,
|
| 216 |
+
"eval_loss": 0.04316685348749161,
|
| 217 |
+
"eval_mean_token_accuracy": 0.9816457834243775,
|
| 218 |
+
"eval_runtime": 572.5081,
|
| 219 |
+
"eval_samples_per_second": 1.747,
|
| 220 |
+
"eval_steps_per_second": 0.218,
|
| 221 |
+
"step": 60
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.52,
|
| 225 |
+
"grad_norm": 0.02974274381995201,
|
| 226 |
+
"learning_rate": 0.00011119644761033078,
|
| 227 |
+
"loss": 0.0412,
|
| 228 |
+
"mean_token_accuracy": 0.9832498252391815,
|
| 229 |
+
"step": 65
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.52,
|
| 233 |
+
"eval_loss": 0.04281109571456909,
|
| 234 |
+
"eval_mean_token_accuracy": 0.9816416425704956,
|
| 235 |
+
"eval_runtime": 572.3877,
|
| 236 |
+
"eval_samples_per_second": 1.747,
|
| 237 |
+
"eval_steps_per_second": 0.218,
|
| 238 |
+
"step": 65
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 0.56,
|
| 242 |
+
"grad_norm": 0.03527842089533806,
|
| 243 |
+
"learning_rate": 9.719537437241312e-05,
|
| 244 |
+
"loss": 0.0422,
|
| 245 |
+
"mean_token_accuracy": 0.9818739414215087,
|
| 246 |
+
"step": 70
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 0.56,
|
| 250 |
+
"eval_loss": 0.042156100273132324,
|
| 251 |
+
"eval_mean_token_accuracy": 0.9818385190963745,
|
| 252 |
+
"eval_runtime": 572.8793,
|
| 253 |
+
"eval_samples_per_second": 1.746,
|
| 254 |
+
"eval_steps_per_second": 0.218,
|
| 255 |
+
"step": 70
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.6,
|
| 259 |
+
"grad_norm": 0.03159947320818901,
|
| 260 |
+
"learning_rate": 8.324937766952638e-05,
|
| 261 |
+
"loss": 0.0415,
|
| 262 |
+
"mean_token_accuracy": 0.9812052190303803,
|
| 263 |
+
"step": 75
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.6,
|
| 267 |
+
"eval_loss": 0.04201458767056465,
|
| 268 |
+
"eval_mean_token_accuracy": 0.9818556265830993,
|
| 269 |
+
"eval_runtime": 572.8981,
|
| 270 |
+
"eval_samples_per_second": 1.746,
|
| 271 |
+
"eval_steps_per_second": 0.218,
|
| 272 |
+
"step": 75
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.64,
|
| 276 |
+
"grad_norm": 0.05335124209523201,
|
| 277 |
+
"learning_rate": 6.963232548903853e-05,
|
| 278 |
+
"loss": 0.0434,
|
| 279 |
+
"mean_token_accuracy": 0.981024295091629,
|
| 280 |
+
"step": 80
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.64,
|
| 284 |
+
"eval_loss": 0.0419701524078846,
|
| 285 |
+
"eval_mean_token_accuracy": 0.9820275835990906,
|
| 286 |
+
"eval_runtime": 572.5542,
|
| 287 |
+
"eval_samples_per_second": 1.747,
|
| 288 |
+
"eval_steps_per_second": 0.218,
|
| 289 |
+
"step": 80
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.68,
|
| 293 |
+
"grad_norm": 0.03178861737251282,
|
| 294 |
+
"learning_rate": 5.6611626088244194e-05,
|
| 295 |
+
"loss": 0.0446,
|
| 296 |
+
"mean_token_accuracy": 0.9806674957275391,
|
| 297 |
+
"step": 85
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.68,
|
| 301 |
+
"eval_loss": 0.04141935706138611,
|
| 302 |
+
"eval_mean_token_accuracy": 0.9820602812767029,
|
| 303 |
+
"eval_runtime": 571.1387,
|
| 304 |
+
"eval_samples_per_second": 1.751,
|
| 305 |
+
"eval_steps_per_second": 0.219,
|
| 306 |
+
"step": 85
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 0.72,
|
| 310 |
+
"grad_norm": 0.040076956152915955,
|
| 311 |
+
"learning_rate": 4.444297669803981e-05,
|
| 312 |
+
"loss": 0.04,
|
| 313 |
+
"mean_token_accuracy": 0.9825624287128448,
|
| 314 |
+
"step": 90
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.72,
|
| 318 |
+
"eval_loss": 0.04136450216174126,
|
| 319 |
+
"eval_mean_token_accuracy": 0.9821350569725037,
|
| 320 |
+
"eval_runtime": 571.1079,
|
| 321 |
+
"eval_samples_per_second": 1.751,
|
| 322 |
+
"eval_steps_per_second": 0.219,
|
| 323 |
+
"step": 90
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.76,
|
| 327 |
+
"grad_norm": 0.037734489887952805,
|
| 328 |
+
"learning_rate": 3.336534220479961e-05,
|
| 329 |
+
"loss": 0.0443,
|
| 330 |
+
"mean_token_accuracy": 0.9814210355281829,
|
| 331 |
+
"step": 95
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.76,
|
| 335 |
+
"eval_loss": 0.041187919676303864,
|
| 336 |
+
"eval_mean_token_accuracy": 0.9821804766654968,
|
| 337 |
+
"eval_runtime": 572.5641,
|
| 338 |
+
"eval_samples_per_second": 1.747,
|
| 339 |
+
"eval_steps_per_second": 0.218,
|
| 340 |
+
"step": 95
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.8,
|
| 344 |
+
"grad_norm": 0.04483392462134361,
|
| 345 |
+
"learning_rate": 2.3596262417839255e-05,
|
| 346 |
+
"loss": 0.0394,
|
| 347 |
+
"mean_token_accuracy": 0.9827538132667542,
|
| 348 |
+
"step": 100
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 0.8,
|
| 352 |
+
"eval_loss": 0.041005875915288925,
|
| 353 |
+
"eval_mean_token_accuracy": 0.9823568396568298,
|
| 354 |
+
"eval_runtime": 572.6853,
|
| 355 |
+
"eval_samples_per_second": 1.746,
|
| 356 |
+
"eval_steps_per_second": 0.218,
|
| 357 |
+
"step": 100
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.84,
|
| 361 |
+
"grad_norm": 0.035200804471969604,
|
| 362 |
+
"learning_rate": 1.5327580077171587e-05,
|
| 363 |
+
"loss": 0.0406,
|
| 364 |
+
"mean_token_accuracy": 0.9826211035251617,
|
| 365 |
+
"step": 105
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.84,
|
| 369 |
+
"eval_loss": 0.040836114436388016,
|
| 370 |
+
"eval_mean_token_accuracy": 0.9824077115058899,
|
| 371 |
+
"eval_runtime": 572.7135,
|
| 372 |
+
"eval_samples_per_second": 1.746,
|
| 373 |
+
"eval_steps_per_second": 0.218,
|
| 374 |
+
"step": 105
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.88,
|
| 378 |
+
"grad_norm": 0.033276792615652084,
|
| 379 |
+
"learning_rate": 8.72167349386811e-06,
|
| 380 |
+
"loss": 0.0418,
|
| 381 |
+
"mean_token_accuracy": 0.9821678340435028,
|
| 382 |
+
"step": 110
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.88,
|
| 386 |
+
"eval_loss": 0.040718287229537964,
|
| 387 |
+
"eval_mean_token_accuracy": 0.9824112377166748,
|
| 388 |
+
"eval_runtime": 571.9537,
|
| 389 |
+
"eval_samples_per_second": 1.748,
|
| 390 |
+
"eval_steps_per_second": 0.219,
|
| 391 |
+
"step": 110
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.92,
|
| 395 |
+
"grad_norm": 0.03913981840014458,
|
| 396 |
+
"learning_rate": 3.908267805490051e-06,
|
| 397 |
+
"loss": 0.0413,
|
| 398 |
+
"mean_token_accuracy": 0.9817866742610931,
|
| 399 |
+
"step": 115
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.92,
|
| 403 |
+
"eval_loss": 0.040665969252586365,
|
| 404 |
+
"eval_mean_token_accuracy": 0.9824027805328369,
|
| 405 |
+
"eval_runtime": 572.0692,
|
| 406 |
+
"eval_samples_per_second": 1.748,
|
| 407 |
+
"eval_steps_per_second": 0.219,
|
| 408 |
+
"step": 115
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.96,
|
| 412 |
+
"grad_norm": 0.053303398191928864,
|
| 413 |
+
"learning_rate": 9.818874663554357e-07,
|
| 414 |
+
"loss": 0.0391,
|
| 415 |
+
"mean_token_accuracy": 0.9823383331298828,
|
| 416 |
+
"step": 120
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.96,
|
| 420 |
+
"eval_loss": 0.04064611718058586,
|
| 421 |
+
"eval_mean_token_accuracy": 0.9824066462516785,
|
| 422 |
+
"eval_runtime": 571.9701,
|
| 423 |
+
"eval_samples_per_second": 1.748,
|
| 424 |
+
"eval_steps_per_second": 0.219,
|
| 425 |
+
"step": 120
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.0,
|
| 429 |
+
"grad_norm": 0.031917620450258255,
|
| 430 |
+
"learning_rate": 0.0,
|
| 431 |
+
"loss": 0.0414,
|
| 432 |
+
"mean_token_accuracy": 0.9825141310691834,
|
| 433 |
+
"step": 125
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 1.0,
|
| 437 |
+
"eval_loss": 0.04065406322479248,
|
| 438 |
+
"eval_mean_token_accuracy": 0.9823896341323852,
|
| 439 |
+
"eval_runtime": 571.6781,
|
| 440 |
+
"eval_samples_per_second": 1.749,
|
| 441 |
+
"eval_steps_per_second": 0.219,
|
| 442 |
+
"step": 125
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.0,
|
| 446 |
+
"step": 125,
|
| 447 |
+
"total_flos": 8.539757562822656e+16,
|
| 448 |
+
"train_loss": 0.07947162318229675,
|
| 449 |
+
"train_runtime": 16205.2206,
|
| 450 |
+
"train_samples_per_second": 0.062,
|
| 451 |
+
"train_steps_per_second": 0.008
|
| 452 |
+
}
|
| 453 |
+
],
|
| 454 |
+
"logging_steps": 5,
|
| 455 |
+
"max_steps": 125,
|
| 456 |
+
"num_input_tokens_seen": 0,
|
| 457 |
+
"num_train_epochs": 1,
|
| 458 |
+
"save_steps": 100,
|
| 459 |
+
"stateful_callbacks": {
|
| 460 |
+
"TrainerControl": {
|
| 461 |
+
"args": {
|
| 462 |
+
"should_epoch_stop": false,
|
| 463 |
+
"should_evaluate": false,
|
| 464 |
+
"should_log": false,
|
| 465 |
+
"should_save": true,
|
| 466 |
+
"should_training_stop": true
|
| 467 |
+
},
|
| 468 |
+
"attributes": {}
|
| 469 |
+
}
|
| 470 |
+
},
|
| 471 |
+
"total_flos": 8.539757562822656e+16,
|
| 472 |
+
"train_batch_size": 4,
|
| 473 |
+
"trial_name": null,
|
| 474 |
+
"trial_params": null
|
| 475 |
+
}
|