Latent Consistency Model (LCM) LoRA: SSD-1B
Latent Consistency Model (LCM) LoRA was proposed in LCM-LoRA: A universal Stable-Diffusion Acceleration Module by Simian Luo, Yiqin Tan, Suraj Patil, Daniel Gu et al.
It is a distilled consistency adapter for segmind/SSD-1B
that allows
to reduce the number of inference steps to only between 2 - 8 steps.
Model | Params / M |
---|---|
lcm-lora-sdv1-5 | 67.5 |
lcm-lora-ssd-1b | 105 |
lcm-lora-sdxl | 197M |
Usage
LCM-LoRA is supported in π€ Hugging Face Diffusers library from version v0.23.0 onwards. To run the model, first
install the latest version of the Diffusers library as well as peft
, accelerate
and transformers
.
audio dataset from the Hugging Face Hub:
pip install --upgrade pip
pip install --upgrade diffusers transformers accelerate peft
Text-to-Image
Let's load the base model segmind/SSD-1B
first. Next, the scheduler needs to be changed to LCMScheduler
and we can reduce the number of inference steps to just 2 to 8 steps.
Please make sure to either disable guidance_scale
or use values between 1.0 and 2.0.
import torch
from diffusers import LCMScheduler, AutoPipelineForText2Image
model_id = "segmind/SSD-1B"
adapter_id = "latent-consistency/lcm-lora-ssd-1b"
pipe = AutoPipelineForText2Image.from_pretrained(model_id, torch_dtype=torch.float16, variant="fp16")
pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config)
pipe.to("cuda")
# load and fuse lcm lora
pipe.load_lora_weights(adapter_id)
pipe.fuse_lora()
prompt = "Self-portrait oil painting, a beautiful cyborg with golden hair, 8k"
# disable guidance_scale by passing 0
image = pipe(prompt=prompt, num_inference_steps=4, guidance_scale=0).images[0]
Image-to-Image
Works as well! TODO docs
Inpainting
Works as well! TODO docs
ControlNet
Works as well! TODO docs
T2I Adapter
Works as well! TODO docs
Speed Benchmark
TODO
Training
TODO
- Downloads last month
- 3,140
Model tree for latent-consistency/lcm-lora-ssd-1b
Base model
segmind/SSD-1B