Improve documentation after trying on a new machine. #4002
Conversation
|
Hi @SandSnip3r! Thank you for your pull request and welcome to our community. Action RequiredIn order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at [email protected]. Thanks! |
github-actions
bot
added
the
documentation
lanluo-nvidia
left a comment
lanluo-nvidia
left a comment
There was a problem hiding this comment.
Left some minor comments
LGTM
|
|
||
| Download and install the CUDA Toolkit from the `NVIDIA Developer website <https://developer.nvidia.com/cuda-downloads>`_. | ||
|
|
||
| **Important:** Check the required CUDA version in the `MODULE.bazel <https://github.com/pytorch/TensorRT/blob/main/MODULE.bazel>`_ file. You must install the exact CUDA toolkit version specified there (for example, at the time of writing, CUDA 13.0 is required). |
There was a problem hiding this comment.
Supported cuda version are cu129 and cu130,
if you have cu129 installed, you can change this line https://github.com/pytorch/TensorRT/blob/main/MODULE.bazel#L39
to your version
There was a problem hiding this comment.
Since multiple versions are supported, should we just change the MODULE.bazel to look at the /usr/local/cuda/ path instead? Then we don't have to mention specific versions in this document.
Or would you just prefer if I add what you commented to the document?
| # Install PyTorch nightly build (check CUDA version in MODULE.bazel) | ||
| # Replace cuXXX with your CUDA version (e.g., cu130 for CUDA 13.0) | ||
| pip install --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cuXXX |
There was a problem hiding this comment.
we can add a explaination, if user is building torch_tensorrt from TOT(torch_tensorrt TOT is aligned with torch TOT),
It is using the torch nightly,
There was a problem hiding this comment.
Added the explanation, please take a look
| # Replace cuXXX with your CUDA version (e.g., cu130 for CUDA 13.0) | ||
| pip install --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cuXXX | ||
| # If you encounter version conflicts during the build, you may need to specify |
There was a problem hiding this comment.
You are building in a time where torch is already upgraded to 2.11.dev, and torch_tensorrt has not yet upgraded to 2.11.dev
(normally this happens only a few days during release upgrade)
I would suggest to remove this section, it might confuse user.
There was a problem hiding this comment.
Removed the section. Thanks for reviewing
2 participants
Description
I ran through the TensorRT-RTX installation/build/run instructions and ran into some issues along the way. I am on a fresh install, so I think I uncovered some installation steps that might've been not noticed on a more mature development machine.
Fixes # (issue)
Type of change
Checklist: