add eval for attention sink #7070
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
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/7070
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit a5420ae with merge base ddec0c7 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
This was referenced Nov 25, 2024
Merged
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
helunwencser
added a commit
that referenced
this pull request
Nov 25, 2024
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) ghstack-source-id: 255393585 Pull Request resolved: #7070
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: <img width="966" alt="Screenshot 2024-11-25 at 2 46 04 PM" src="https://github.com/user-attachments/assets/ba7118f9-b5d7-4de8-b1fa-7d2ba0646515"> [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
helunwencser
added a commit
that referenced
this pull request
Nov 26, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 255425318 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/)
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: <img width="966" alt="Screenshot 2024-11-25 at 2 46 04 PM" src="https://github.com/user-attachments/assets/ba7118f9-b5d7-4de8-b1fa-7d2ba0646515"> [ghstack-poisoned]
helunwencser
added a commit
that referenced
this pull request
Nov 27, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 255585076 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/)
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
helunwencser
added
the
release notes:attention_sink
larryliu0820
approved these changes
Dec 2, 2024
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: <img width="966" alt="Screenshot 2024-11-25 at 2 46 04 PM" src="https://github.com/user-attachments/assets/ba7118f9-b5d7-4de8-b1fa-7d2ba0646515"> [ghstack-poisoned]
helunwencser
added a commit
that referenced
this pull request
Dec 2, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 256083503 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/)
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: <img width="966" alt="Screenshot 2024-11-25 at 2 46 04 PM" src="https://github.com/user-attachments/assets/ba7118f9-b5d7-4de8-b1fa-7d2ba0646515"> [ghstack-poisoned]
helunwencser
added a commit
that referenced
this pull request
Dec 2, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 256087023 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/)
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: <img width="966" alt="Screenshot 2024-11-25 at 2 46 04 PM" src="https://github.com/user-attachments/assets/ba7118f9-b5d7-4de8-b1fa-7d2ba0646515"> [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
helunwencser
added a commit
that referenced
this pull request
Dec 2, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 256095324 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/)
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: <img width="966" alt="Screenshot 2024-11-25 at 2 46 04 PM" src="https://github.com/user-attachments/assets/ba7118f9-b5d7-4de8-b1fa-7d2ba0646515"> [ghstack-poisoned]
helunwencser
added a commit
that referenced
this pull request
Dec 2, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 256108079 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/)
|
This pull request was exported from Phabricator. Differential Revision: D66474732 |
facebook-github-bot
merged commit de6c4cd
into
gh/helunwencser/80/base
41 of 43 checks passed
facebook-github-bot
temporarily deployed
to
cherry-pick-bot
GitHub Actions
Inactive
helunwencser
added a commit
that referenced
this pull request
Dec 3, 2024
Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 256108079 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Co-authored-by: Lunwen He <[email protected]>
helunwencser
added a commit
that referenced
this pull request
Dec 3, 2024
* Transform model to be able to use Attention Sink Pull Request resolved: #6700 This PR adds necessary functions for transforming the model to be able to use Attention Sink. ghstack-source-id: 256108077 @exported-using-ghexport Differential Revision: [D65571289](https://our.internmc.facebook.com/intern/diff/D65571289/) * Update eager runner to support AttentionSink Pull Request resolved: #6921 This PR updates the eager runner to support AttentionSink. It also fixes issues in the `chat_completion` function to properly handle the position id. ghstack-source-id: 256108078 Differential Revision: [D66076486](https://our.internmc.facebook.com/intern/diff/D66076486/) * add eval for attention sink (#7150) Pull Request resolved: #7070 This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled. This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled. ghstack-source-id: 256108079 @exported-using-ghexport Differential Revision: [D66474732](https://our.internmc.facebook.com/intern/diff/D66474732/) Co-authored-by: Lunwen He <[email protected]> --------- Co-authored-by: Lunwen He <[email protected]>
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.
Stack from ghstack (oldest at bottom):
This PR adds the function to evaluate the model's perplexity when AttentionSink is enabled.
This is mostly copied from https://github.com/mit-han-lab/streaming-llm/blob/main/examples/eval_long_ppl.py which is used by the AttentionSink paper to evaluate the model's perplexity when AttentionSink is enabled.
Differential Revision: D66474732
Perplexity measured for llama 3.2 1B and 1B_Instruct model up to 40k tokens with AttentionSink enabled: