-
We can wait on the accelerated discovery post until we hear back from Yang.
The intent is to expose XYZ control, joint angle control, gripper control, and camera snapshots. It may make sense to expos…
-
I want to run the gpu to full speed
-
### System Info
```Shell
Custom sdxl training script using fsdp GRAD_SHARD_OP with cpu offset.
After upgrading accelerate from 33.0 to 34.0, after collecting state_dict with accelerator.get_state_…
-
### Root Cause
The root cause is due to recent transformers update [to resolve high CPU usage for large quantized models](https://github.com/huggingface/transformers/pull/33154).
- what the PR…
-
### System Info
```Shell
- `Accelerate` version: 0.34.0
- Platform: Linux-5.10.134-13.al8.x86_64-x86_64-with-glibc2.31
- `accelerate` bash location: /root/miniconda/envs/flux_diffuser/bin/accelerat…
-
Based on practical tests, deploying omost-llama-3-8b on an A100 using torch==2.3.0+cu118, vllm==0.5.0.post1+cu118, and xformers==0.0.26.post1+cu118 works well. if want to speed up the process, can ref…
-
You can disable touch based scrolling (clicking and dragging and releasing and the scrolling continuing depending on the speed of the pointer move at release) with `touch-action:none` but (to my knowl…
-
Hi, that's a really awesome project you have created!
I've been thinking a lot lately on the in-browser virtualization topics, with my dream to run a full-blown linux distro with acceptable performan…
-
I'm currently trying to run the script with the following paperspace machine
![image](https://github.com/user-attachments/assets/01bbcf97-12b3-4dab-9f85-d066398199c5)
When doing so, I get this err…
-
With the latest version, I get the mentioned error message. How can I fix that? See also #2820.
```
You are using the default legacy behaviour of the . This is expected, and simply means that the …