Dario

> please have a look and merge if this is fine - this should update this PR at the same time. Definitely good and it works, I just tested it!...

Nevermind for the upstream, already opened two PRs for the GODEL's [base](https://huggingface.co/microsoft/GODEL-v1_1-base-seq2seq) and [large](https://huggingface.co/microsoft/GODEL-v1_1-large-seq2seq) models: - https://huggingface.co/microsoft/GODEL-v1_1-large-seq2seq/discussions/3 - https://huggingface.co/microsoft/GODEL-v1_1-base-seq2seq/discussions/6

@guillaume-be tests added! The only weird thing that I see here is the output from GODEL (you can see in the tests), for some strange reason, every generated texts starts...

@guillaume-be should we have to ping Microsoft someway in order to get PRs merged?

Just opened an issue [here](https://github.com/microsoft/GODEL/issues/50) 💪🏻

Same here, how we can do this, maybe programmatically? I think that a function like `webview.zoom(percentage);` would be awesome, how I can do that? there are bindings?

@Blakeinstein this is a different thing, what I need is to zoom the Webkit renderer, is something different, this is surely possible but not exposed by this lib.

@raphtlw done [here](https://github.com/Boscop/web-view/pull/235) too :-)

> Based on your details and links, I was able to get the code to run using the fp16 weights for unet (available [here](https://huggingface.co/CompVis/stable-diffusion-v1-4/tree/fp16/unet)) and making a couple changes to...

Thanks @LaurentMazare! Tried just now, converted from the original weight but now I get the out of memory issue: ``` CUDA out of memory. Tried to allocate 3.16 GiB (GPU...