Fix shortfin gibberish tokens and add cpu_llm_server_test.py #351
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
renxida
force-pushed
the
make_llm_count
branch
2 times, most recently
from
789abe8
to
59bbe42
Compare
renxida
force-pushed
the
make_llm_count
branch
from
59bbe42
to
56e5b2a
Compare
|
removed failing integration tests & will put them in another pr |
|
To test these changes, run: https://gist.github.com/renxida/d40e14b1c9b4606279d19899bcce5e9a |
kumardeepakamd
approved these changes
Oct 29, 2024
| while token_int != 128001: | ||
| # TODO: Use correct eot token from config. | ||
| # while token_int != 128001: | ||
| for i in range(15): |
There was a problem hiding this comment.
needs a proper fix to iterate over a variable number of tokens. okay to land and iterate.
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Outdated suggestions cannot be applied.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.
Rob and I found some weirdness in paged_llm_v1 that was inconsistent with how shortfin worked.
We hacked a solution for now but we really need to fix paged_llm_v1 by cleaning up the seq_lens and start_positions and adding some CLEAR documentation as to EXACTLY what they do.
Currently we're able to make the llm count.
Right now the context length is effectively 16 tokens. By adding cache allocation, we'll soon be able to actually talk with the LLM.