Q8 Weight Quantization #24
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
Closed
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Outdated suggestions cannot be applied.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 support for 8 bit quantization in the ggml q8_0 format. The main changes are
MetalQuantizedCompilerwhich takes a set of weights and converts downstream matmuls and gathers to QuantizedMatmuls and QuantizedGathers. Each op takes in weights in q8, and does dequantization internally.In the mistral example, there is a MetalQ8Loader which demonstrates how to load from q q8 gguf file, and returns the set of weight nodes.
Mistral Decode Speed on M1 Pro:
Average token generated in 56.18ms - (17.80 tok/s)