Add Softmax kernel in Triton. Use softmax kernel and argmax in Llama generation.py. + Small changes #11
+211
−4
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
catherinelee274
changed the title
adamomainz
requested changes
Sep 24, 2024
benchmarking/benchmark_utils.py
Outdated
| import pandas as pd | ||
|
|
||
|
|
||
| def compare_benchmarks(benchmarks: Dict[str, Dict[str, Any]]) -> Dict[str, Any]: |
There was a problem hiding this comment.
any reason you are deleting this?
Comment on lines
+196
to
+199
| if self.use_triton: | ||
| probs = triton_softmax(logits[:,-1]) | ||
| else: | ||
| probs = self.Math.softmax(logits[:, -1] / temperature, dim=-1) |
There was a problem hiding this comment.
there is a mathOps file in this directory to abstract this away from users. Lets use that instead and allow for proper benchmarking (see decorator on the functions there)
Comment on lines
+203
to
+206
| if self.use_triton: | ||
| next_token = self.triton.language.argmax(logits[:, -1], axis=-1) | ||
| else: | ||
| next_token = self.Math.argmax(logits[:, -1], dim=-1) |
| import pytest | ||
| from kernels.fused_softmax import triton_softmax | ||
|
|
||
| @pytest.mark.parametrize("input_size", [(1024, 1024), (512, 512), (2048, 512)]) |
There was a problem hiding this comment.
Thanks for adding tests!
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Outdated suggestions cannot be applied.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.
dist.destroy_process_group()to remove warning during benchmarkingResults from calling
python3 main.py llama_chat_completion --benchmark --ckpt_dir <model_checkpoint_path> --tokenizer_path <model_tokenizer_path>With No Changes:
With just softmax
With softmax and argmax