mirror of
https://github.com/zml/zml.git
synced 2026-11-18 18:57:29 +00:00
CPU buffers behave differently from Cuda/Rocm/TPU buffers, and in particular they are addressable from main memory. This was annoying and prevented us to add some safety checks that weren't valid for CPU. I'm adding a tag to buffers so that we know which target they refer to. Also there is a bug in PJRT unitialized which doesn't respected the memory. I've opened https://github.com/openxla/xla/pull/31292 but in the meantime I've a workaround in ZML that uses the regular .fromHostBuffer. (I've also a patched version of PJRT plugins in https://github.com/zml/pjrt-artifacts/pull/41 but not merged yet) |
History
|
|
|---|---|---|
| .. | ||
| howtos | ||
| learn | ||
| misc | ||
| tutorials | ||
| README.md | ||
Greetings, Programs!
If you want to run our example models, please head over to Getting Started.
Ready to write some code? Try starting with your first model in ZML, or familiarize yourself with the high-level ZML concepts first.
MENU:
Tutorials
- Getting Started : install ZML, run LLama
- Writing your first model
- Simplifying Dimension Handling - Tagged Tensors!
How to ?
- HuggingFace Authentication
- Port Pytorch models to ZML
- Add Weights Files to your projects
- Cross Compile and Deploy Models on a Server
- Dockerize Models
Learn more...
- ZML Concepts : Tensors, Models, Executables, etc. explained