updated info.md
Browse files
info.md
CHANGED
|
@@ -44,13 +44,15 @@
|
|
| 44 |
|
| 45 |
📚 **キャラの中国語名は何ですか？ここにご覧ください：[ウマ娘ビリビリWiki](https://wiki.biligame.com/umamusume/%E8%B5%9B%E9%A9%AC%E5%A8%98%E4%B8%80%E8%A7%88).** 📚
|
| 46 |
|
|
|
|
|
|
|
| 47 |
## Training Details - For those who may be interested
|
| 48 |
|
| 49 |
🎈 **This work switches [cl-tohoku/bert-base-japanese-v3](https://huggingface.co/cl-tohoku/bert-base-japanese-v3) to [ku-nlp/deberta-v2-base-japanese](https://huggingface.co/ku-nlp/deberta-v2-base-japanese) expecting potentially better performance, and, just for fun.** 🥰
|
| 50 |
|
| 51 |
❤ Thanks to **SUSTech Center for Computational Science and Engineering**. ❤ This model is trained on A100 (40GB) x 2 with **batch size 32** in total.
|
| 52 |
|
| 53 |
-
💪 This model has been trained for **
|
| 54 |
|
| 55 |
📕 This work uses linear with warmup **(7.5% of total steps)** LR scheduler with ` max_lr=1e-4`. 📕
|
| 56 |
|
|
|
|
| 44 |
|
| 45 |
📚 **キャラの中国語名は何ですか？ここにご覧ください：[ウマ娘ビリビリWiki](https://wiki.biligame.com/umamusume/%E8%B5%9B%E9%A9%AC%E5%A8%98%E4%B8%80%E8%A7%88).** 📚
|
| 46 |
|
| 47 |
+
---------------
|
| 48 |
+
|
| 49 |
## Training Details - For those who may be interested
|
| 50 |
|
| 51 |
🎈 **This work switches [cl-tohoku/bert-base-japanese-v3](https://huggingface.co/cl-tohoku/bert-base-japanese-v3) to [ku-nlp/deberta-v2-base-japanese](https://huggingface.co/ku-nlp/deberta-v2-base-japanese) expecting potentially better performance, and, just for fun.** 🥰
|
| 52 |
|
| 53 |
❤ Thanks to **SUSTech Center for Computational Science and Engineering**. ❤ This model is trained on A100 (40GB) x 2 with **batch size 32** in total.
|
| 54 |
|
| 55 |
+
💪 This model has been trained for **3 cycles, 270K steps (=180 epoch)** . 💪
|
| 56 |
|
| 57 |
📕 This work uses linear with warmup **(7.5% of total steps)** LR scheduler with ` max_lr=1e-4`. 📕
|
| 58 |
|