-
### *Project idea 3: Research about deploying LLM with Jina*
| info | details |
| ---------------- | ------------------------------…
-
### This issue is to have a centralized place to list and track work on adding support to new ops for the MPS backend.
[**PyTorch MPS Ops Project**](https://github.com/users/kulinseth/projects/1/vi…
-
Users typically go through 3 phases
- **quickstart** comes with preloaded data (quickstart-batch.sh) (start's zookeeper, controller, server, broker in a single process)
- **standalone (local)**: sta…
-
I would like to start a high level discussion about the priorities and organization of the pangeo CMIP6 archive in the cloud.
I have officially? taken over this effort, and would first and foremost…
-
Is there any plan to make PerfView usable on MacOs?
-
### Context
The amount of custom frontend Javascript has continued to grow. We chose to keep our code components isolated to the degree practicable and use Drupal's native libraries functionality t…
-
In addition to the `balena ssh` command, it would be helpful to be able to have `balena scp`. As far as I understand, this will require a change in the proxy server (`ssh.balena-devices.com:22`) as we…
-
I was really excited to see RFD 114! GPGPU compute on Illumos has been an interest of mine for years. I think the shortest open path on Triton would be to enable PCI-E passthrough to KVM instances run…
-
```
Tools: 5.1.1
Platform: Ubuntu
```
### Steps to reproduce
I was upgrading the `firebase-tools` from `3.17.7` to the latest version `5.1.1` since our CI was unable to deploy the functions d…
-
Doing hardware-accelerated inference in a serverless environment is compelling use case.
However, adding straight up GPU passthrough means that microVM can't oversubscribe memory, and we need to a…