[fix] Update bench command to use forward_prefill and forward_decode (#55) #58

Open
sleepy wants to merge 1 commit from fix/55-bench-command into main
Owner

Summary

Fixes the bench subcommand which was calling model.forward() (which now returns error.UseForwardPrefillDecode).

Changes

  • bench_prefill: Uses model.forward_prefill() instead of model.forward()
  • bench_decode: Uses model.forward_prefill() for prompt, then model.forward_decode() for token generation
  • Added sync_linear_states_to_gpu() after prefill to match engine pattern

Test Results

  • zig build compiles cleanly
## Summary Fixes the `bench` subcommand which was calling `model.forward()` (which now returns `error.UseForwardPrefillDecode`). ## Changes - `bench_prefill`: Uses `model.forward_prefill()` instead of `model.forward()` - `bench_decode`: Uses `model.forward_prefill()` for prompt, then `model.forward_decode()` for token generation - Added `sync_linear_states_to_gpu()` after prefill to match engine pattern ## Test Results - `zig build` compiles cleanly
This pull request can be merged automatically.
This branch is out-of-date with the base branch
You are not authorized to merge this pull request.
View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u origin fix/55-bench-command:fix/55-bench-command
git switch fix/55-bench-command

Merge

Merge the changes and update on Forgejo.

Warning: The "Autodetect manual merge" setting is not enabled for this repository, you will have to mark this pull request as manually merged afterwards.

git switch main
git merge --no-ff fix/55-bench-command
git switch fix/55-bench-command
git rebase main
git switch main
git merge --ff-only fix/55-bench-command
git switch fix/55-bench-command
git rebase main
git switch main
git merge --no-ff fix/55-bench-command
git switch main
git merge --squash fix/55-bench-command
git switch main
git merge --ff-only fix/55-bench-command
git switch main
git merge fix/55-bench-command
git push origin main
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
sleepy/sleepy-llm!58
No description provided.