Skip to content

GitLab

Commit f5ca5af6 authored by Patrick von Platen's avatar Patrick von Platen
Browse files

add to readme

parent 2ac19ff1
Show whitespace changes
Inline Side-by-side
Showing with 22 additions and 75 deletions
README.md
View file @ f5ca5af6
...@@ -82,72 +82,22 @@ For more examples see [schedulers](https://github.com/huggingface/diffusers/tree ...@@ -82,72 +82,22 @@ For more examples see [schedulers](https://github.com/huggingface/diffusers/tree
```python ```python
import torch import torch
from diffusers import UNetModel, DDPMScheduler from diffusers import UNetUnconditionalModel, DDIMScheduler
import PIL import PIL.Image
import numpy as np import numpy as np
import tqdm import tqdm
generator = torch.manual_seed(0)
torch_device = "cuda" if torch.cuda.is_available() else "cpu" torch_device = "cuda" if torch.cuda.is_available() else "cpu"
# 1. Load models # 1. Load models
noise_scheduler = DDPMScheduler.from_config("fusing/ddpm-lsun-church", tensor_format="pt") scheduler = DDIMScheduler.from_config("fusing/ddpm-celeba-hq", tensor_format="pt")
unet = UNetModel.from_pretrained("fusing/ddpm-lsun-church").to(torch_device) unet = UNetUnconditionalModel.from_pretrained("fusing/ddpm-celeba-hq", ddpm=True).to(torch_device)
# 2. Sample gaussian noise # 2. Sample gaussian noise
generator = torch.manual_seed(23)
unet.image_size = unet.resolution
image = torch.randn( image = torch.randn(
(1, unet.in_channels, unet.resolution, unet.resolution), (1, unet.in_channels, unet.image_size, unet.image_size),
generator=generator,
)
image = image.to(torch_device)
# 3. Denoise
num_prediction_steps = len(noise_scheduler)
for t in tqdm.tqdm(reversed(range(num_prediction_steps)), total=num_prediction_steps):
# predict noise residual
with torch.no_grad():
residual = unet(image, t)
# predict previous mean of image x_t-1
pred_prev_image = noise_scheduler.step(residual, image, t)
# optionally sample variance
variance = 0
if t > 0:
noise = torch.randn(image.shape, generator=generator).to(image.device)
variance = noise_scheduler.get_variance(t).sqrt() * noise
# set current image to prev_image: x_t -> x_t-1
image = pred_prev_image + variance
# 5. process image to PIL
image_processed = image.cpu().permute(0, 2, 3, 1)
image_processed = (image_processed + 1.0) * 127.5
image_processed = image_processed.numpy().astype(np.uint8)
image_pil = PIL.Image.fromarray(image_processed[0])
# 6. save image
image_pil.save("test.png")
```
#### **Example for Unconditonal Image generation [LDM](https://github.com/CompVis/latent-diffusion):**
```python
import torch
from diffusers import UNetModel, DDIMScheduler
import PIL
import numpy as np
import tqdm
generator = torch.manual_seed(0)
torch_device = "cuda" if torch.cuda.is_available() else "cpu"
# 1. Load models
noise_scheduler = DDIMScheduler.from_config("fusing/ddpm-celeba-hq", tensor_format="pt")
unet = UNetModel.from_pretrained("fusing/ddpm-celeba-hq").to(torch_device)
# 2. Sample gaussian noise
image = torch.randn(
(1, unet.in_channels, unet.resolution, unet.resolution),
generator=generator, generator=generator,
) )
image = image.to(torch_device) image = image.to(torch_device)
...@@ -155,32 +105,29 @@ image = image.to(torch_device) ...@@ -155,32 +105,29 @@ image = image.to(torch_device)
# 3. Denoise # 3. Denoise
num_inference_steps = 50 num_inference_steps = 50
eta = 0.0 # <- deterministic sampling eta = 0.0 # <- deterministic sampling
scheduler.set_timesteps(num_inference_steps)
for t in tqdm.tqdm(reversed(range(num_inference_steps)), total=num_inference_steps): for t in tqdm.tqdm(scheduler.timesteps):
# 1. predict noise residual # 1. predict noise residual
orig_t = len(noise_scheduler) // num_inference_steps * t
with torch.no_grad(): with torch.no_grad():
residual = unet(image, orig_t) residual = unet(image, t)["sample"]
# 2. predict previous mean of image x_t-1 prev_image = scheduler.step(residual, t, image, eta)["prev_sample"]
pred_prev_image = noise_scheduler.step(residual, image, t, num_inference_steps, eta)
# 3. optionally sample variance # 3. set current image to prev_image: x_t -> x_t-1
variance = 0 image = prev_image
if eta > 0:
noise = torch.randn(image.shape, generator=generator).to(image.device)
variance = noise_scheduler.get_variance(t).sqrt() * eta * noise
# 4. set current image to prev_image: x_t -> x_t-1 # 4. process image to PIL
image = pred_prev_image + variance
# 5. process image to PIL
image_processed = image.cpu().permute(0, 2, 3, 1) image_processed = image.cpu().permute(0, 2, 3, 1)
image_processed = (image_processed + 1.0) * 127.5 image_processed = (image_processed + 1.0) * 127.5
image_processed = image_processed.numpy().astype(np.uint8) image_processed = image_processed.numpy().astype(np.uint8)
image_pil = PIL.Image.fromarray(image_processed[0]) image_pil = PIL.Image.fromarray(image_processed[0])
# 6. save image # 5. save image
image_pil.save("test.png") image_pil.save("generated_image.png")
```
#### **Example for Unconditonal Image generation [LDM](https://github.com/CompVis/latent-diffusion):**
```python
``` ```
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment