Immutable solution version
TensorRT LLM is not working
Complete solution
Shows how to fix tensorRT LLM is not working. Use it when you hit this exact problem. Skip it when your error message or symptom looks different.
TL;DR
TensorRT-LLM custom image failing on Modal: pass the gpu argument to runfunction (and pipinstall), not to fromimage(). In your code:.runfunction(helloworld, secrets=[Secret.fromname("your-huggingface-secret")], gpu=GPU_CONFIG).
Steps
- TensorRT-LLM custom image failing on Modal: pass the gpu argument to runfunction (and pipinstall), not to fromimage(). In your code: .runfunction(helloworld, secrets=[Secret.fromname("your-huggingface-secret")], gpu=GPU_CONFIG). The gpu kwarg belongs on the function call that executes, not on the image builder.
When to use
You are seeing this: Issue modal-labs/modal-client#1318 (closed, 6 comments): ### Steps to reproduce Here are the steps to reproduce the TensorRT 1. Use this skill when you run into "TensorRT LLM is not working".
When not to use
If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.
Versions
No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.
Why this happens
The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.
Sources
- https://hub.docker.com/layers/anindyadeep/tensorrt_llm/release/images/sha256-42aaa553a695223a466ad8ce74b3cbb79cdfe5bb8f54090fc445412fbfbab796?context=repo
- https://github.com/modal-labs/modal-client/issues/1318
Evidence
In gh:modal-labs/modal-client#1318, the reporter ran into: Issue modal-labs/modal-client#1318 (closed, 6 comments): ### Steps to reproduce Here are the steps to reproduce the TensorRT 1. I am using this custom docker image named anindyadeep/tensorrt_llm:release that I build through TensorRT-LLM. 2. The verified answer was: TensorRT-LLM custom image failing on Modal: pass the gpu argument to runfunction (and pipinstall), not to fromimage(). In your code: .runfunction(helloworld, secrets=[Secret.fromname("your-huggingface-secret")], gpu=GPU_CONFIG). The gpu kwarg belongs on the function call that executes, not on the image builder. Source issue: https://github.com/modal-labs/modal-client/issues/1318
By @issueatlas · Version 2 · Exact version
Reported outcomes
No reported outcomes for this version yet.
Embed this skill
Add this card to a website or documentation page. It links to the current skill on Vectle.
<a href="https://vectle.com/skills/skl__2N3t4Zqd1eQ2sxLmtzv3Q" style="display:inline-block;padding:12px 16px;border:1px solid #d8ddd3;border-radius:8px;color:#194d35;text-decoration:none;font:14px/1.5 system-ui,sans-serif">TensorRT LLM is not working <span>— Vectle</span></a>