Skip to content

Conversation

@ochougul
Copy link
Contributor

@ochougul ochougul commented Dec 23, 2025

This is counter-intuitive but being done for easy interface and less options to play with for DA mode

…only model, even though this is not intuitive, being done for vllm interface to be easier

Signed-off-by: Onkar Chougule <[email protected]>
@ochougul ochougul changed the title making kv_cache_batch_size equivalent to full_batch_size for prefill_… making kv_cache_batch_size equivalent to full_batch_size for prefill-only model Dec 23, 2025
@quic-hemagnih quic-hemagnih merged commit e5a3497 into release/v1.21.0 Dec 23, 2025
6 checks passed
ochougul added a commit that referenced this pull request Dec 24, 2025
…only model (#687)

This is counter-intuitive but being done for easy interface and less
options to play with for DA mode

---------

Signed-off-by: Onkar Chougule <[email protected]>
ochougul added a commit that referenced this pull request Dec 24, 2025
…only model (#687)

This is counter-intuitive but being done for easy interface and less
options to play with for DA mode

---------

Signed-off-by: Onkar Chougule <[email protected]>
ochougul added a commit that referenced this pull request Dec 24, 2025
…only model (#687)

This is counter-intuitive but being done for easy interface and less
options to play with for DA mode

---------

Signed-off-by: Onkar Chougule <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants