Skip to content
This repository has been archived by the owner on Sep 19, 2023. It is now read-only.

Latest commit

 

History

History
45 lines (37 loc) · 1.18 KB

tips_troubleshooting.md

File metadata and controls

45 lines (37 loc) · 1.18 KB

Solving Errors

Latest version of fastai library

Do git pull of fastai library. Updates may sort out some errors.

git pull

Update Anaconda packages

conda env update
conda update --all 

Delete tmp directory and rerun

CUDA out of memory error

  • interrupt kernel
  • reduce batch size
  • RESTART kernel!

TTA (Test Time Augmentation)

  • forum post
  • "TTA used to return the average of the augmentations as a prediction. Now it returns the set so you can do with them as you please."

Error with this code

log_preds,y = learn.TTA()
probs = np.exp(log_preds)
accuracy(log_preds,y), metrics.log_loss(y, probs)

Adjust with this code

log_preds,y = learn.TTA()
preds = np.mean(np.exp(log_preds),0)

Empty graph with learning rate finder

  • try increasing the batch size

Debugging

Note from Jeremy:
Immediately after you get the error, type %debug in a cell to enter the debugger. Then use the standard python debugger commands to follow your code to see what’s happening.