Burhanuddin Mustafa Lakdawala

Sure, I'll give it a go.

+1 for this issue. I have a Windows Server 2019 (v1809) machine which has the same issue mentioned above. I tried switching to VirtualBox as well but VirtualBox doesn't seem...

+1 I was going to create a new feature request before I found this one. Can it be reopened?

#2266 has a pretty good workaround that is more targeted [here](https://github.com/AdAway/AdAway/issues/2266#issuecomment-866780924)

Questionary has been really great, especially with limited terminal support in Windows.

@robsalk 1. Find the GPU id of the M6000 (I used `torch.cuda.get_device_name()` to verify). 2. Open `cmd_windows.bat`. 3. Set the `CUDA_VISIBLE_DEVICES` environment variable to the GPU id you found. 4....

`CUDA_VISIBLE_DEVICES` does work for me, but I would've preferred a solution from within the webui (maybe using `device_map`?). But if it's not feasible, I'll keep using the `CUDA_VISIBLE_DEVICES` workaround.

@vivekratr 1. Run `cmd_windows.bat` (or your OS specific executable) in the text-generation-webui folder. 2. Set the CUDA_VISIBLE_DEVICES like so: - Windows: `set CUDA_VISIBLE_DEVICES=0` - Linux and MacOS: `export CUDA_VISIBLE_DEVICES=0` You...

Please don't close this issue I'm facing the same problem: https://stackoverflow.com/questions/77197723/how-do-i-stream-huggingfacepipeline-output-to-a-langchain-dataframe-agent

The command you specified does work temporarily, this is what I used: ```sh poetry run pip install -e ./path_to_diff_package --config-settings editable_mode=strict ```` However the next `poetry install` or `poetry install...