You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/getting_started.md
+4-3Lines changed: 4 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -241,11 +241,12 @@ Optional arguments are:
241
241
242
242
- `--no-validate` (**not suggested**): By default, the codebase will perform evaluation at every k iterations during the training. To disable this behavior, use `--no-validate`.
243
243
- `--work-dir ${WORK_DIR}`: Override the working directory specified in the config file.
244
-
- `--resume-from ${CHECKPOINT_FILE}`: Resume from a previous checkpoint file.
244
+
- `--resume-from ${CHECKPOINT_FILE}`: Resume from a previous checkpoint file (to continue the training process).
245
+
- `--load-from ${CHECKPOINT_FILE}`: Load weights from a checkpoint file (to start finetuning for another task).
245
246
246
247
Difference between `resume-from` and `load-from`:
247
-
`resume-from` loads both the model weights and optimizer status, and the iteration number is also inherited from the specified checkpoint. It is usually used for resuming the training process that is interrupted accidentally.
248
-
`load-from`only loads the model weights and the training iteration starts from 0. It is usually used for finetuning.
248
+
- `resume-from` loads both the model weights and optimizer state including the iteration number.
249
+
- `load-from` loads only the model weights, starts the training from iteration 0.
0 commit comments