[ADD] vgg16 tils model #112
Merged
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
Setting mean=sample and std=sample indicates that each image should be normalized to have mean 0 and variance 1.
This reverts commit 3fdcbb3.
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Outdated suggestions cannot be applied.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.
this pr adds the vgg16 tils model from https://github.com/ShahiraAbousamra/til_classification
the original model was implemented in tensorflow slim and the weights were converted to a pytorch state dict using the script https://github.com/SBU-BMI/wsinfer/blob/main/scripts/convert_tf_to_pytorch_til_vgg16.py
the tensorflow model file was downloaded from Box (https://stonybrookmedicine.app.box.com/file/905508899913?s=r0p2h2oh53zw52o6fm6qrqjtcwd3p0ga). the weights files were those with the prefix
tcga_vgg16_mix_new3.ckpt.i am not 100% sure that the image normalization is correct. at the moment, images are normalized to range [-1, 1]. this appears to give correct output. normalizing each image to mean 0 and variance 1 gave poor results so it is probably not correct. using original image values [0, 255] also gave poor results. the original code does reference normalizing to [-1, 1] but it is not fully clear that this was used for this model.
fixes #110