ru-dalle
poolformer
ru-dalle | poolformer | |
---|---|---|
50 | 3 | |
1,639 | 1,226 | |
-0.3% | 0.0% | |
0.0 | 0.0 | |
over 1 year ago | over 1 year ago | |
Jupyter Notebook | Jupyter Notebook | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
ru-dalle
- I trained a custom AI model for fakemon outputs. Feel free to use them for inspiration! No credit needed.
-
I trained an AI model to help me design fakebadge concepts. Full album in comments. Please feel free to take these for your own inspiration, too!
It’s a custom trained model, built in rudalle https://github.com/ai-forever/ru-dalle
-
Using AI to draft new ideas for legendaries.
It's a custom model, built from rudalle
-
SD photorealism to the extreme, is MJ really that better?
ru-dalle has had that feature for quite a while, as it was their first inpainting example notebook:
-
2 Google Colab notebooks are available for the large ruDALL-E Kandinsky model (12 billion parameters). The smaller ruDALL-E model has 1.3 billion parameters.
GitHub repo.
-
Colab notebook "pharmapsychotic modified rudalle" lets the user choose which of 4 ruDALL-E models to use
Colab notebook. There are actually 5 models, but I doubt the 12B parameter Kandinsky model is actually available per looking at this code.
-
Tree in a field.
This was made with a mini version of DALL-E: ruDALL-E
-
I trained an AI model to generate images of ancient Roman imperial denarii
Specifically, I fine-tuned ru-DALLE using a dataset consisting of ~1000 images of imperial denarii (ranging from Augustus through Maximinus Thrax) coupled with descriptions of each coin grabbed from OCRE. For example, the obverse description of this coin would be "Head of Augustus, bare, right" and the reverse description would be "Round shield, spear-head, and curved sword".
-
New ruDALL-E 1.3 billion parameter model version 3 has been released with ruDALL-E v1.0.0
One way to use the version 3 model is to use this official Colab notebook linked to in the ruDALL-E GitHub repo. I recommend making the changes mentioned in this post. If you want to use the older version 2 model with this Colab notebook, change 'Malevich' to 'Malevich_v2' in line "dalle = get_rudalle_model('Malevich', pretrained=True, fp16=True, device=device)" (relevant source code).
-
Preview of ruDALL-E v0.5.0 from the developer
# !pip install rudalle==0.0.1rc8 > /dev/null !pip3 install git+https://github.com/sberbank-ai/ru-dalle.git@feature/new_malevich
poolformer
-
Researchers from Sea AI Lab and National University of Singapore Introduce ‘PoolFormer’: A Derived Model from MetaFormer for Computer Vision Tasks
GitHub: https://github.com/sail-sg/poolformer
-
[D] Are Image Transformers Overhyped? "MetaFormer is all you need" explained (5-minute summary by Casual GAN Papers)
arxiv / code
-
[P] Fine-tuning the new PoolFormer (MetaFormer) model on a Kaggle Competitions Dataset
Code for https://arxiv.org/abs/2111.11418 found: https://github.com/sail-sg/poolformer
What are some alternatives?
NeuralTextToImage - Colabs for text prompt steered image generators
pytorch-seq2seq - Tutorials on implementing a few sequence-to-sequence (seq2seq) models with PyTorch and TorchText.
SpecVQGAN - Source code for "Taming Visually Guided Sound Generation" (Oral at the BMVC 2021)
fastai - The fastai deep learning library
HugsVision - HugsVision is a easy to use huggingface wrapper for state-of-the-art computer vision
naver-webtoon-faces - Generative models on NAVER Webtoon faces
TFLiteClassification - TensorFlow Lite Image Classification Python Implementation
ganspace - Discovering Interpretable GAN Controls [NeurIPS 2020]
FunMatch-Distillation - TF2 implementation of knowledge distillation using the "function matching" hypothesis from https://arxiv.org/abs/2106.05237.
FinRL-Meta - FinRL-Meta: Dynamic datasets and market environments for FinRL.
Transformer-in-Transformer - An Implementation of Transformer in Transformer in TensorFlow for image classification, attention inside local patches