Legacy checkpoint read, fix checkpoint re-training issue #434
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Related issues
InferenceRPUConfig
out_scaling
factorsDescription
Checkpoint load modified the parameter in the tile reconstruction, which however caused a mismatch in the
module._parameters
dictionary, which meant thatmodel.parameters()
picked the wrong references.This caused the model to not train after checkpoint loading, in case when
out_scaling_alpha
were set to be trainedNow the parameters are re-registered correctly for each load.
Details
out_scaling_alpha
loading