search

InternLM / InternLM-XComposer

InternLM-XComposer-2.5: A Versatile Large Vision Language Model Supporting Long-Contextual Input and Output
Apache License 2.0
2.47k stars 153 forks source link

error when lora finetune: Trainer: evaluation requires an eval_dataset. #221

Closed zhuliyi0 closed 5 months ago

zhuliyi0 commented 6 months ago

inside finetune.py, eval_dataset is set to None, and this gives error when eval is called.

I tried to replicate a similar dataset like the training set, and set value to eval_dataset, but with no luck. No very familiar with transformer code.

Need help. Thanks in advance!

8%|██████████▌ | 10/120 [03:04<33:35, 18.32s/it]Traceback (most recent call last): File "/root/InternLM-XComposer/finetune/finetune.py", line 311, in train() File "/root/InternLM-XComposer/finetune/finetune.py", line 301, in train trainer.train() File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 1553, in train return inner_training_loop( File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 1927, in _inner_training_loop self._maybe_log_save_evaluate(tr_loss, model, trial, epoch, ignore_keys_for_eval) File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 2254, in _maybe_log_save_evaluate metrics = self.evaluate(ignore_keys=ignore_keys_for_eval) File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 2964, in evaluate eval_dataloader = self.get_eval_dataloader(eval_dataset) File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 879, in get_eval_dataloader raise ValueError("Trainer: evaluation requires an eval_dataset.") ValueError: Trainer: evaluation requires an eval_dataset.

yuhangzang commented 6 months ago

The current version does not support the evaluation during fine-tuning.

yuhangzang commented 5 months ago

Please feel free to reopen this issue if you have any further questions. Closing this issue due to inactivity.

khyati2396 commented 1 month ago

The current version does not support the evaluation during fine-tuning.

Is this supported now for internLM-XComposer-2.5?? If not then when will it be supported?