[FIX] Updated readme.md and tracer.py - #117
gujialiang123 wants to merge 12 commits into
Conversation
|
#106 Fixed in this PR. |
mark14wu
changed the title
mark14wu
changed the title
readme.md and tracer.py
|
|
||
| ```sh | ||
| pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu121 | ||
| pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu |
There was a problem hiding this comment.
This is not 100% correct. We don't rely on either torch-cpu or torch-gpu
There was a problem hiding this comment.
Better to just refer people to the pytorch installation page
There was a problem hiding this comment.
And say it's up to them to install any torch version they like. We support it anyway.
There was a problem hiding this comment.
Done! Please check it!
There was a problem hiding this comment.
Oh, what I meant is much simpler.
Just refer to this page https://pytorch.org/get-started/locally/
There was a problem hiding this comment.
Without putting any concrete instructions
|
@Jokeren Hi Keren, could you add Jialiang as a maintainer of triton-viz? |
|
@gujialiang123 I sent you two invitations. Please accept |
@Jokeren Thanks! I have accepted them! |
@mark14wu Done! |
|
|
||
| ```sh | ||
| pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu121 | ||
| pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu |
|
|
||
|
|
||
| def collect_grid(): | ||
| # 动态获取当前的 launches，尽最大可能拿到"非空"的那份 |
There was a problem hiding this comment.
@gujialiang123 also you don't have to upload any ipynb files
There was a problem hiding this comment.
@gujialiang123 also you don't have to upload any ipynb files
Sorry, I got the wrong branch... I'll fix these problems later, sorry.
|
Abandoned. |
3 participants
1 Fixed a problem in tracer.py where the register_for_loop_callback function was missing a None return value.
2 Updated the installation process in readme. You can directly install the CPU version of Torch, so there is no need to uninstall pytorch-triton. the previous installation steps do not work on my computer.