fix: fix CohereForAI/c4ai-command-r-plus #1707
Merged
Conversation
This file contains hidden or 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
drbh
reviewed
Apr 4, 2024
| @@ -43,7 +43,7 @@ def __init__( | |||
| ] | |||
| self.free_block_mask = torch.ones(num_blocks, dtype=torch.int32, device="cpu") | |||
| self.slots = torch.arange( | |||
| 0, num_blocks * self.block_size, dtype=torch.int32 | |||
| 0, num_blocks * self.block_size, dtype=torch.int64 | |||
There was a problem hiding this comment.
quick question, is there a case where num_blocks is really really big? or maybe there very large block_indices sometimes?
just trying to understand the type change
There was a problem hiding this comment.
It's because the vllm kernel now ask for this dtype. I don't know why they changed.
Slots is a very small tensor anyway.
OlivierDehaene
commented
Apr 9, 2024
server/text_generation_server/models/custom_modeling/flash_mistral_modeling.py
Outdated
Show resolved
Hide resolved
OlivierDehaene
force-pushed
the
chore/update_flash
branch
from
e0e96d2 to
26da6bf
Compare
|
Hello @OlivierDehaene, @drbh this pull request slightly changes decoding, and so breaks my integration pipeline (I'm testing multiple inputs on my models and asserting their outputs don't change with do_sample=False).Do you know why this change was needed and if it's going to stay this way? |
Nilabhra
pushed a commit
to TII-AI-Research-Center/text-generation-inference
that referenced
this pull request
May 14, 2024
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Outdated suggestions cannot be applied.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.
@Narsil @drbh this will update flash attention v2 and vllm.
You will need to re-install them.