search

michaelnny / miniGPT

Try to implement pre-training and fine-tuning GPT-2 model for research and education purpose.
MIT License
8 stars 1 forks source link

Few other runtime issues #3

Open ankur6ue opened 6 months ago

ankur6ue commented 6 months ago
  1. TypeError: compute_metrics() missing 1 required positional argument: 'mask'
  2. logger.info: info doesn't exist
  3. cfg.max_train_iters is not defined in the config file

Thanks, Ankur

ankur6ue commented 6 months ago

Another one: RuntimeError: output with shape [] doesn't match the broadcast shape [1] in: scaler.step(optimizer)