[Feature request] increment epoch number when restoring model from checkpoint
Open
- Dominant language
- Python
- Stars
- 452
- Forks
- 76
- PR merge metrics
- No merged PRs in 30d
Description
Currently, the epoch counter of the training script starts from 0, even when a model is restored from a checkpoint (e.g. epoch-2).
Contributor guide
Research direction
Start at the training script's checkpoint-restore path and inspect how the epoch counter is initialized. Verify that restoring a checkpoint such as epoch-2 continues with the corresponding epoch number instead of restarting from 0.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100