-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | ubuntu 22.04
Queue | TigerUbuntu
Baseline | [026f615bdecc40a6627fdbbdd6d578d35a8da1e7](https://github.com/dotnet/runtime/commit/026…
-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | ubuntu 22.04
Queue | TigerUbuntu
Baseline | [e4473e06b50d898193b5a260b4521977cd0be34c](https://github.com/dotnet/runtime/commit/e44…
-
### Checks
- [X] I've already read https://github.com/actions/actions-runner-controller/blob/master/TROUBLESHOOTING.md and I'm sure my issue is not covered in the troubleshooting guide.
- [X] I'm not…
-
I often get OOM before the VAE decode, but if I queue again, with latents from sampling still in memory, it works. After OOM, Comfy unloads all models. I think this clears up enough vram for VAE decod…
-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | ubuntu 22.04
Queue | TigerUbuntu
Baseline | [8734660d8907a58cc32cc1df9ca9efc02edc31c8](https://github.com/dotnet/runtime/co…
-
### Feature Idea
when you run a workflow, then add an additional save image input then rerun it, comfy will naturally start running at the level of the newly added nodes (provided nothing has changed…
-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | Windows 10.0.22631
Queue | ViperWindows
Baseline | [fe3dcc541a9fbf01a8f9807a956c77b50bb38e81](https://github.com/dotnet/runtime/com…
-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | ubuntu 22.04
Queue | TigerUbuntu
Baseline | [abde3f9498f09b5016a7a0d7d2e1b81ce5c1b614](https://github.com/dotnet/runtime/co…
-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | Windows 10.0.22621
Queue | TigerWindows
Baseline | [2864ec446610af711889f22c4bf88904d57b6bf3](https://github.com/dotnet/runtime/com…
-
### Run Information
Name | Value
-- | --
Architecture | x86
OS | Windows 10.0.22621
Queue | TigerWindows
Baseline | [e8e3e148fd9c56182cf5676b47f2697779ec5c6e](https://github.com/dotnet/run…