Rename README.md to lambdalabs/image-mixer
#1
by
erwanfoto
- opened
- README.md +0 -42
- lambdalabs/image-mixer +2 -0
README.md
DELETED
|
@@ -1,42 +0,0 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: openrail
|
| 3 |
-
datasets:
|
| 4 |
-
- ChristophSchuhmann/LAION-5B-EN-Aesthetics-Subset_above_6
|
| 5 |
-
---
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
Image Mixer is a model that lets you combine the concepts, styles, and compositions from multiple images (and text prompts too) and generate new images.
|
| 10 |
-
|
| 11 |
-
It was trained by [Justin Pinkney](https://www.justinpinkney.com) at [Lambda Labs](https://lambdalabs.com/).
|
| 12 |
-
|
| 13 |
-
## Training details
|
| 14 |
-
|
| 15 |
-
This model is a fine tuned version of [Stable Diffusion Image Variations](https://huggingface.co/lambdalabs/sd-image-variations-diffusers)
|
| 16 |
-
it has been trained to accept multiple CLIP embedding concatenated along the sequence dimension (as opposed to 1 in the original model).
|
| 17 |
-
During training up to 5 crops of the training images are taken and CLIP embeddings are extracted, these are concatenated and used as the conditioning for the model.
|
| 18 |
-
At inference time, CLIP embeddings from multiple images can be used to generate images which are influence by multiple inputs.
|
| 19 |
-
|
| 20 |
-
Training was done at 640x640 on a subset of LAION improved aesthetics, using 8xA100 from [Lambda GPU Cloud](https://cloud.lambdalabs.com).
|
| 21 |
-
|
| 22 |
-
_Note text captions were not used during training of the model,
|
| 23 |
-
although input text embeddings works to some extent during inference, the model is primarily designed to accept image embeddings_
|
| 24 |
-
|
| 25 |
-
## Usage
|
| 26 |
-
|
| 27 |
-
The model is available on [huggingface spaces](https://huggingface.co/spaces/lambdalabs/image-mixer-demo) or to run locally do the following:
|
| 28 |
-
|
| 29 |
-
```bash
|
| 30 |
-
git clone https://github.com/justinpinkney/stable-diffusion.git
|
| 31 |
-
cd stable-diffusion
|
| 32 |
-
git checkout 1c8a598f312e54f614d1b9675db0e66382f7e23c
|
| 33 |
-
python -m venv .venv --prompt sd
|
| 34 |
-
. .venv/bin/activate
|
| 35 |
-
pip install -U pip
|
| 36 |
-
pip install -r requirements.txt
|
| 37 |
-
python scripts/gradio_image_mixer.py
|
| 38 |
-
```
|
| 39 |
-
|
| 40 |
-
Then navigate to the gradio demo link printed in the terminal.
|
| 41 |
-
|
| 42 |
-
For details on how to use the model outside the app refer to the [`run` function](https://github.com/justinpinkney/stable-diffusion/blob/c1963a36a4f8ce23784c8247fa1af0e34e02b766/scripts/gradio_image_mixer.py#L79) in `gradio_image_mixer.py` in the [original repo](https://github.com/justinpinkney/stable-diffusion#image-mixer)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
lambdalabs/image-mixer
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
https://www.salvador-dali.org/media/upload/cataleg_pintura/BAJA/1100.jpg
|
| 2 |
+
https://lepetitjournal.com/sites/default/files/Image%202%20-%20Main%20tenant%20un%20miroir%20spherique%20M.%20C.%20Escher%20lithographie%201935%20%20Courtesy%20of%20the%20Palazzo%20Reale.jpg
|