vllm-project/vllm

[torch.compile] E2E correctness testing for fusions

Open

#39,428 opened on Apr 9, 2026

View on GitHub
 (8 comments) (0 reactions) (0 assignees)Python (16,816 forks)batch import
help wantedtorch.compile

Repository metrics

Stars
 (80,034 stars)
PR merge metrics
 (Avg merge 3d 17h) (993 merged PRs in 30d)

Description

E2E tests for fusions (tests/compile/fusions_e2e) have done a great job preventing fusion regressions where model/forward code changes break a custom torch.compile fusion pass. However, we currently have no way of testing correctness for these fusion configurations.

It would be good to investigate an approach where we only run a few layers of a model and compare the outputs. This would be helpful for correctness testing in general, and we could compare the outputs to both a baseline vLLM configuration and the huggingface baseline.

This would likely require some work to fix weight loading for models like DeepSeek when --hf-overrides.num_hidden_layers is overriden.

Contributor guide

Research direction
Investigate how to run only a few layers of a model for E2E correctness testing of torch.compile fusions, compare outputs to baseline vLLM and HuggingFace configurations, and fix weight loading for models like DeepSeek when overriding num hidden layers.
Tech stack
pythonpytorch
Domain
backend
Issue type
Test
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
PythonPyTorchtorch.compile
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
45