-
Notifications
You must be signed in to change notification settings - Fork 148
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Exllama kernels support #313
Exllama kernels support #313
Conversation
This is a great integration if it results in higher inference speed with the same accuracy. Have you benchmarked perplexity/speed? However, there are a few things that are not great:
|
You might also be able to use awq_ext.dequantize_weights_cuda(qweight, scales, qzeros, split_k_iters, 0, 0, False) |
I just finished PPL comparison, almost exactly the same.
I noticed however that when generating the output ids are generally the same, but there is a small difference in the logits. |
For 1, yes I think we can do that, we'd also like to add ROCm wheels, and I'm not sure how that can be done exactly @fxmarty |
Updates from internal discussion:
Results:
|
I made a new release https://github.com/casper-hansen/AutoAWQ_kernels/releases/tag/v0.0.2 that includes the ExLlama kernels. I also renamed to |
I benchmarked GEMM vs ExLlamaV2 on a single RTX 4090. Results in end-to-end testing with
Prefilling speedup is probably something you can achieve with AWQ kernels as well - the strategy is to dequantize and run FP16 matmul since it's faster. GEMM (AWQ kernel)
ExLlamaV2
|
# What does this PR do? <!-- Congratulations! You've made it this far! You're not quite done yet though. Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution. Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change. Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost. --> <!-- Remove if not applicable --> This PR adds the possibility to run AWQ models with Exllama/GPTQ kernels, specifically for ROCm devices that support Exllama kernels but not AWQ's GEMM. This is done by : - un-packing, reordering and re-packing AWQ weights when `--quantize gptq` but the model's `quant_method=awq`. - avoiding overflows when adding 1 to zeros in exllama and triton. Ref: casper-hansen/AutoAWQ#313 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @ @OlivierDehaene OR @Narsil --> --------- Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
impressive. was this benchmark comparing AWQ vs. EXL2? |
This was comparing the AWQ GEMM kernel vs EXL2 kernel in AutoAWQ, so not directly against ExLlamaV2 repository. |
# What does this PR do? <!-- Congratulations! You've made it this far! You're not quite done yet though. Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution. Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change. Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost. --> <!-- Remove if not applicable --> This PR adds the possibility to run AWQ models with Exllama/GPTQ kernels, specifically for ROCm devices that support Exllama kernels but not AWQ's GEMM. This is done by : - un-packing, reordering and re-packing AWQ weights when `--quantize gptq` but the model's `quant_method=awq`. - avoiding overflows when adding 1 to zeros in exllama and triton. Ref: casper-hansen/AutoAWQ#313 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @ @OlivierDehaene OR @Narsil --> --------- Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
# What does this PR do? <!-- Congratulations! You've made it this far! You're not quite done yet though. Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution. Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change. Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost. --> <!-- Remove if not applicable --> This PR adds the possibility to run AWQ models with Exllama/GPTQ kernels, specifically for ROCm devices that support Exllama kernels but not AWQ's GEMM. This is done by : - un-packing, reordering and re-packing AWQ weights when `--quantize gptq` but the model's `quant_method=awq`. - avoiding overflows when adding 1 to zeros in exllama and triton. Ref: casper-hansen/AutoAWQ#313 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @ @OlivierDehaene OR @Narsil --> --------- Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
This PR adds a new layer
WQLinear_Exllama
/WQLinear_ExllamaV2
to perform inference using exllama kernels (requires installing AutoGPTQ).