Skip to content

skill(infer-hw-adapt, infer-precision-check): expand backend coverage and add shared precision check skill #18

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Caozhou1995 merged 6 commits into from
Jul 13, 2026
Merged

skill(infer-hw-adapt, infer-precision-check): expand backend coverage and add shared precision check skill #18

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
fb6dda0
skill(infer-hw-adapt): expand backend coverage and fix duplicate content
physics31415926 Jun 29, 2026
6cb1758
tmp: modelscope availability check script
physics31415926 Jul 1, 2026
3222e72
delete tmp file
physics31415926 Jul 1, 2026
0ac8d53
Merge remote-tracking branch 'upstream/main' into skill/infer-hw-adap…
physics31415926 Jul 6, 2026
02d6547
skill(infer-precision-check): add shared precision verification skill…
physics31415926 Jul 6, 2026
6348a5a
skill(infer-precision-check): add SUMMARY.md
physics31415926 Jul 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 21 additions & 2 deletions flagscale_agent/skills/infer-env-setup/SKILL.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -304,15 +304,34 @@ ssh <ssh_host> "docker exec <container> bash -c \

Record the version: `memory_write('<backend>_vllm_pinned_version', 'X.Y.Z')`

### Step 2: Install vLLM CPU-only
### Step 2: Install vLLM CPU-only (from source)

Install from source to ensure `VLLM_TARGET_DEVICE=empty` takes effect correctly.
Pre-built PyPI wheels may bundle device-specific compiled extensions that conflict
with the plugin's hardware backend.

```bash
# Clone the pinned vLLM version
ssh <ssh_host> "docker exec <container> bash -c \
'VLLM_TARGET_DEVICE=empty pip install vllm==<pinned_version> \
'cd /workspace/adapt/<backend>-vllm-<version> && \
git clone https://github.com/vllm-project/vllm.git vllm-src && \
cd vllm-src && git checkout v<pinned_version> && \
git log -1 --oneline'"

# Install from source with empty device target
ssh <ssh_host> "docker exec <container> bash -c \
'cd /workspace/adapt/<backend>-vllm-<version>/vllm-src && \
VLLM_TARGET_DEVICE=empty pip install -e . \
--extra-index-url https://download.pytorch.org/whl/cpu \
2>&1 | tee /workspace/adapt-logs/install_vllm.log'"
```

> **Why source install?** `VLLM_TARGET_DEVICE=empty` must be set at *build time* when
> installing from source so that no CUDA/hardware C extensions are compiled. Installing
> a pre-built wheel from PyPI may silently include compiled ops that break on non-NVIDIA
> backends. Source install with this flag produces a pure-Python, device-agnostic vLLM
> that the plugin can override completely.

### Step 3: Clone vllm-plugin-FL (fresh workspace)

```bash
Expand Down
Loading