Enabling transformer and T5 to be quantized with different types #173
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
…uding qint8, qfloat8_e4m3fn (qfloat8 is the alias) and qfloat8_e5m2 by specifying quantization_type_transformer or quantization_type_t5 in the model config section of a config yaml file.
|
A bit of a back story why this PR exists. I have the following spec: Windows And I found that ai-toolkit is not able to generate pre-training samples and/or training due to qfloat8 quantization. The generation creates black (blank) images for me due to encountering invalid values. the training encounters inf values. The solution that I have found so far is to switch the quantization from qfloat8 to qint8 for both transformer and T5. At this point it is not clear which of the modules above are causing the qfloat8 quantization to fail. |
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Outdated suggestions cannot be applied.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.
Enabling transformer and T5 to be quantized with different types including qint8, qfloat8_e4m3fn (qfloat8 is the alias) and qfloat8_e5m2 by specifying quantization_type_transformer or quantization_type_t5 in the model config section of a config yaml file.