Fixed TypeError at start of unet or segnet training #52
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Hi @liebharc , thanks for the PR!! The modification looks good to me. Merging this. |
liebharc
added a commit
to liebharc/oemer
that referenced
this pull request
Dec 1, 2023
Fixed TypeError at start of unet or segnet training (BreezeWhite#52)
liebharc
added a commit
to liebharc/oemer
that referenced
this pull request
Dec 1, 2023
* Fixed 'TypeError: Cannot convert 4.999899999999999e-07 to EagerTensor of dtype int64' in training, fixes BreezeWhite#39 https://stackoverflow.com/questions/76511182/tensorflow-custom-learning-rate-scheduler-gives-unexpected-eagertensor-type-erro * --format was deprecated in ruff and replaced wtih --output-format
BreezeWhite
pushed a commit
that referenced
this pull request
Jan 29, 2024
* Fixed typos in comments * IndexError while scanning for a dot should not abort the whole process * Bound check while getting the note label * Added check if label is in the note_type_map * Filter staffs instead of aborting with an exception * Bound check during symbol extraction * Marking notes as invalid instead of aborting with an exception * Bound check * Fixed type error * Fixed TypeError at start of unet or segnet training (#52) * Fixed 'TypeError: Cannot convert 4.999899999999999e-07 to EagerTensor of dtype int64' in training, fixes #39 https://stackoverflow.com/questions/76511182/tensorflow-custom-learning-rate-scheduler-gives-unexpected-eagertensor-type-erro * --format was deprecated in ruff and replaced wtih --output-format * HoughLinesP can return None if no lines are found * Fixed error which happens if no rest bboxes were found * Limited try/except block * Fixed typo * Use fixed versions for the linter dependencies to avoid that results are different for the same source code level on different test runs due to update of the dependencies * Fixed type errors which came up with the recent version of cv2 * Going back to the newest version of ruff and mypy as the type errors were introduced by cv2
liebharc
added a commit
to liebharc/oemer
that referenced
this pull request
Jan 31, 2024
* Fixed typos in comments * IndexError while scanning for a dot should not abort the whole process * Bound check while getting the note label * Added check if label is in the note_type_map * Filter staffs instead of aborting with an exception * Bound check during symbol extraction * Marking notes as invalid instead of aborting with an exception * Bound check * Fixed type error * Fixed TypeError at start of unet or segnet training (BreezeWhite#52) * Fixed 'TypeError: Cannot convert 4.999899999999999e-07 to EagerTensor of dtype int64' in training, fixes BreezeWhite#39 https://stackoverflow.com/questions/76511182/tensorflow-custom-learning-rate-scheduler-gives-unexpected-eagertensor-type-erro * --format was deprecated in ruff and replaced wtih --output-format * HoughLinesP can return None if no lines are found * Fixed error which happens if no rest bboxes were found * Limited try/except block * Fixed typo * Use fixed versions for the linter dependencies to avoid that results are different for the same source code level on different test runs due to update of the dependencies * Fixed type errors which came up with the recent version of cv2 * Going back to the newest version of ruff and mypy as the type errors were introduced by cv2
BreezeWhite
added a commit
that referenced
this pull request
Feb 17, 2024
…ingle entry point to all training steps (#54) * Fixed 'TypeError: Cannot convert 4.999899999999999e-07 to EagerTensor of dtype int64' in training, fixes #39 https://stackoverflow.com/questions/76511182/tensorflow-custom-learning-rate-scheduler-gives-unexpected-eagertensor-type-erro * --format was deprecated in ruff and replaced wtih --output-format * Added a single entry point to train all models * Added convenience wrapper for oemer * Tried to figure out the definitions for the dense dataset and to document them in code There is likely an official definition somewhere but I just couldn't find it. So I looked at example and tried to reconstruct the mapping. Unknown basically means that I just couldn't see the symbol on the picture. * Decreased queue sizes as otherwise the training process crashed with an out of memory exception after it used up about 30GB of memory * Added model outputs to git ignore * Added checks for dataset folders * Using default training params * Added workarounds for removal of np.float * Using dataset definitions * Added type annotations * Added a train_all_rests even if the resulting model is right now not used in oemer * segnet and unet should now pick the correct model * Changed label definitions from what appears to be used in oemer right now * With this commit the resulting arch.json matches the one inside of oemer/ceckpoints/seg_net/arch.json * Avoid that the OMR processes finishes prematurely (#53) * Fixed typos in comments * IndexError while scanning for a dot should not abort the whole process * Bound check while getting the note label * Added check if label is in the note_type_map * Filter staffs instead of aborting with an exception * Bound check during symbol extraction * Marking notes as invalid instead of aborting with an exception * Bound check * Fixed type error * Fixed TypeError at start of unet or segnet training (#52) * Fixed 'TypeError: Cannot convert 4.999899999999999e-07 to EagerTensor of dtype int64' in training, fixes #39 https://stackoverflow.com/questions/76511182/tensorflow-custom-learning-rate-scheduler-gives-unexpected-eagertensor-type-erro * --format was deprecated in ruff and replaced wtih --output-format * HoughLinesP can return None if no lines are found * Fixed error which happens if no rest bboxes were found * Limited try/except block * Fixed typo * Use fixed versions for the linter dependencies to avoid that results are different for the same source code level on different test runs due to update of the dependencies * Fixed type errors which came up with the recent version of cv2 * Going back to the newest version of ruff and mypy as the type errors were introduced by cv2 * Fix install from github command in README --------- Co-authored-by: Yoyo <miyashita2010@tuta.io>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Fixed 'TypeError: Cannot convert 4.999899999999999e-07 to EagerTensor of dtype int64' in training, fixes #39
The solution was taken from this discussion: https://stackoverflow.com/questions/76511182/tensorflow-custom-learning-rate-scheduler-gives-unexpected-eagertensor-type-erro