Instructions to use nebulette/nano-transformers with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use nebulette/nano-transformers with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("nebulette/nano-transformers", device_map="auto") - Notebooks
- Google Colab
- Kaggle
|
Download README.md from nebulette/nano-transformers: direct link, hf CLI and curl.
- Browser
- Download file 952 Bytes
-
https://huggingface.co/nebulette/nano-transformers/resolve/main/README.md
- Command line
-
hf download hf://nebulette/nano-transformers/README.md
-
curl -L -o README.md https://huggingface.co/nebulette/nano-transformers/resolve/main/README.md
952 Bytes
metadata
license: apache-2.0
library_name: transformers
The transformer implementation of the Nano diffusion model.
The licence was kept as Apache 2.0 due to the mixed licence.
from model import Transformer2DModel
from PIL import Image
from pipeline import Pipeline
import torch
from transformers import AutoTokenizer, Gemma3ForCausalLM
# from semantic_vae import load_vae
from vae import VAE
if __name__ == '__main__':
pipeline = Pipeline('diffusion_model.safetensors', 'unsloth/gemma-3-270m-it', 'vae.safetensors', device='cuda')
image = pipeline.generate('1girl')
# Manual initialization.
vae = VAE.from_safetensors('vae.safetensors')
# vae = load_vae('semantic_vae_step_00050000.safetensors') # old VAE
tokenizer = AutoTokenizer.from_pretrained('unsloth/gemma-3-270m-it')
te = Gemma3ForCausalLM.from_pretrained('unsloth/gemma-3-270m-it')
model = Transformer2DModel.from_safetensors('diffusion_model.safetensors')