Text-to-Image
Diffusers
TensorBoard
stable-diffusion
stable-diffusion-diffusers
diffusers-training
lora
Instructions to use lamble-lambe/atelie with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use lamble-lambe/atelie with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("fill-in-base-model", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("lamble-lambe/atelie") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
metadata
license: mit
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- diffusers-training
- lora
library_name: diffusers
Ateliê Generativo — Estilo Lambe-Lambe
O modelo foi treinado sobre a base do Stable Diffusion v1.5 no estilo lambe-lambe / retrato vintage, por meio de fine-tuning LoRA (Low-Rank Adaptation), integrado a um pipeline multimodal completo (texto → imagem → áudio).
🌟 Informações Gerais
- Disciplina: Inteligência Artificial Generativa e Modelos Multimodais — UniCEUB
- Professor: Prof. Romes Heriberto
- Desenvolvido por: Diego Nunes de Morais, Eduardo Deodoro de Moraes Florindo, Higo Soares do Lago, Lucio Flavio Vilar de Azevedo, Paulo Victor Torres Martins
- Dataset: Imagens coletadas de acervos com proveniência verificável (Wikimedia Commons, The Met, Flickr), sob licenças de domínio público, CC0, CC-BY ou CC-BY-SA.
Hiperparâmetros
- rank: 4
- learning_rate: 5e-5
- max_train_steps: 1000
- lr_warmup_steps: 100
- lr_scheduler: cosine
🚀 Como Usar
from diffusers import StableDiffusionPipeline
import torch
pipe = StableDiffusionPipeline.from_pretrained(
"stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.float16
).to("cuda")
# revision fixa uma versão específica (ver "Versionamento" abaixo); v1.4.0 = produção
pipe.load_lora_weights("lamble-lambe/atelie", revision="v1.4.0")
img = pipe(
"estilo_lambelambe, a vintage portrait of an elderly man with a hat, detailed face, black and white",
cross_attention_kwargs={"scale": 0.7},
).images[0]
img.save("saida.png")