-
Notifications
You must be signed in to change notification settings - Fork 292
Add Claude MD file #2311
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.
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
base: main
Are you sure you want to change the base?
Add Claude MD file #2311
Conversation
stack-info: PR: #2311, branch: drisspg/stack/66
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/2311
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 2525ed3 with merge base 9cd5851 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Some of this would be nice documentation for humans too I think, particularly architecture overview / core component descriptions, as well as the useful commands for building, linting, testing. Maybe we can include some of it in the contributor guide? |
Humans can read this file if they want 😂 |
|
||
## Development Commands | ||
|
||
### Installation & Build |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
does this install all the pre-reqs? I found that torch wasnt one of those?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
stamping but this might be good to fix
- `uintx/floatx/` - Unsigned integer and floating-point quantized tensors | ||
|
||
**torchao/float8/** - High-performance float8 training | ||
- Delivers up to 1.5x speedup on 512 GPU clusters |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Compared to what?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
fwiw I just ran CC's /setup
command to generate but yeah this likely doens't help the model
**torchao/csrc/** - Custom CUDA/CPU kernels | ||
- CUTLASS-based implementations for maximum performance | ||
- ROCm support for AMD GPUs | ||
- CPU kernels with AVX512 optimizations |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@metascroy maybe add details on experimental folder including mps kernels.
Stacked PRs:
Add Claude MD file