Merge branch 'main' of https://huggingface.co/loubnabnl/santacoder-code-to-text into main
Browse files
README.md
CHANGED
|
@@ -1,3 +1,28 @@
|
|
| 1 |
---
|
| 2 |
license: openrail
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: openrail
|
| 3 |
+
datasets:
|
| 4 |
+
- codeparrot/github-jupyter-code-to-text
|
| 5 |
+
library_name: transformers
|
| 6 |
+
tags:
|
| 7 |
+
- code
|
| 8 |
---
|
| 9 |
+
|
| 10 |
+
# Santacoder code-to-text
|
| 11 |
+
|
| 12 |
+
This model is a fine-tuned version of [bigcode/santacoder](https://huggingface.co/bigcode/santacoder) on
|
| 13 |
+
[copdeparrot/gitub-jupyter-code-to-text](https://huggingface.co/datasets/codeparrot/github-jupyter-code-to-text).
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
## Training procedure
|
| 17 |
+
|
| 18 |
+
The model was trained on 4 A100 for 3h40min with the following hyperparameters were used during training on 4 A100:
|
| 19 |
+
- learning_rate: 5e-05
|
| 20 |
+
- train_batch_size: 2
|
| 21 |
+
- eval_batch_size: 2
|
| 22 |
+
- seed: 42
|
| 23 |
+
- gradient_accumulation_steps: 4
|
| 24 |
+
- total_train_batch_size: 4
|
| 25 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 26 |
+
- lr_scheduler_type: cosine
|
| 27 |
+
- lr_scheduler_warmup_steps: 100
|
| 28 |
+
- training_steps: 800
|