studio / README.md
calcuis's picture
Update README.md
7e72ef2 verified
---
tags:
- gguf-connector
widget:
- text: a girl holding a cat
output:
url: https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9f.webp
- text: remove the buildings in background
output:
url: https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9h.webp
- text: a girl holding a cat and a raccoon standing behind them
output:
url: https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9i.webp
---
# studio
- run it with `gguf-connector`; simply execute the command below in console/terminal
```
ggc w9
```
>
>GGUF/Safetensors available. Please select:
>
>1. fastvlm-0.5b-f16.gguf (act as recognizor [1.52GB](https://huggingface.co/gguf-org/fastvlm-gguf/blob/main/fastvlm-0.5b-f16.gguf))
>2. sd3.5-2b-lite-mxfp4_moe.gguf (act as generator [2.86GB](https://huggingface.co/calcuis/sd3.5-lite-gguf/blob/main/sd3.5-2b-lite-mxfp4_moe.gguf))
>3. sketch-s9-20b-fp4.safetensors (act as transformer; for blackwell card [11.9GB](https://huggingface.co/calcuis/sketch/blob/main/sketch-s9-20b-fp4.safetensors))
>4. sketch-s9-20b-int4.safetensors (act as transformer; for non-blackwell card [11.5GB](https://huggingface.co/calcuis/sketch/blob/main/sketch-s9-20b-int4.safetensors))
>
>Enter your choice (1 to 4): _
>
- all in one inference
## Discriminator
![screenshot](https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9a.png)
- two pictures as input for comparison; similarity metrics as output
![screenshot](https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9b.png)
- the evaluator is squeeze model inside LPIPS
## Describer
![screenshot](https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9c.png)
- recognize image; customize output, i.e., description, locate a specific subject/object
## Generator
![screenshot](https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9d.png)
- generate picture with custom prompt(s)
## Transformer
![screenshot](https://raw.githubusercontent.com/calcuis/gguf-pack/master/w9e.png)
- transform, merge, combine, or edit picture(s)
<Gallery />