-
https://github.com/EESSI/software-layer/pull/710 introduced a change in `configure_easybuild` to support accelerators but this change has not been incorportated into `EESSI-extend`. It's not entirely …
-
### Feature request
Expand `AcceleratorConfig` and corresponding transformers trainer args to allow transformer users to use full feature set of accelerate through the config arguments supported by…
-
https://github.com/onnx/onnx-mlir/blob/main/docs/BuildOnWindows.md
`FAILED: src/Accelerators/CMakeFiles/…
-
### Describe the Bug with repro steps
1. Go to template page
2. Click on an Accelerator card
3. On the accelerator detail page there is a **close** button in footer that takes you back to the temp…
-
# Describe the bug
Problem in Deployment. Logging this Issue on behalf of Youssef Naggar
# Expected behavior
I am currently working on deploying the "Chat with Your Data" solution accelerat…
-
### Description
After you create a custom routing Global Accelerator, it is a common practice to use the port mappings generated by the GA, for example as seen in this [AWS blog post](https://aws.a…
-
Recently our pipelines have stopped getting created when we create new solutions and DevOps Projects. I have traced back the error to the line below and get a '**The remote server returned an error: (…
-
As I understand, through a tag based system, the plugin can assign different compute targets for each node as follows:
```yaml
# excerpt from vertexai.yml
# see https://kedro-vertexai.readthedo…
-
In `train.py`, I see the lines:
```
model = model().to(accelerator.device)
```
This did not work on my mac, because `model()` requires an argument since it calls the `forward` method.
T…
-
InstructLab has train configuration profiles for NVIDIA A100, L40, and L4, but not for Intel Gaudi. Please add a profile for Intel Gaudi 3 systems.
An Intel Gaudi 3 server comes as an 8-way system.…