You can simply overwrite the Trainer loss: Specify Loss for Trainer / TrainingArguments - #2 by nielsr
nielsr
3
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Train with MLM loss augmentation | 0 | 206 | October 10, 2022 | |
| How loss is calculated in MLM training | 0 | 911 | April 1, 2022 | |
| Multi-label token classification | 34 | 8237 | September 6, 2023 | |
| KeyError: 'loss' even though my dataset has labels | 1 | 1014 | November 10, 2021 | |
| Access feature in custom compute_loss method | 0 | 228 | April 15, 2024 |