Comments (4)
Hi @jasonrayshd ,
For cosine noise schedule, it may suffer terrible numerical issues for t near to T. In my previous implementations in DPM-Solver paper, I changed the start time from t_start
=1.0 to t_start
=0.9946 (by comparing the lambda, i.e. half-log-SNR). You can also try other t_start
but do not use t_start
=1.0. I will further support the cosine schedule in a more elegant way.
from dpm-solver.
Got it. Thank you for your great suggestions! Looking forward to your excellent work in the future!
from dpm-solver.
@LuChengTHU @jasonrayshd i have tried cosine scheduler, and the result is better than ddim&ddpm, but i don't know the reason about it, and i'm not sure about is observation could be seen in High dimension data.
here is the settings:
`model = self.denoise_fn
model_kwargs = {}
guidance_scale = w
noise_schedule = NoiseScheduleVP(schedule='discrete', betas=self.betas)
model_fn = model_wrapper(
model,
noise_schedule,
model_type="noise", # or "x_start" or "v" or "score"
model_kwargs=model_kwargs,
guidance_type="classifier-free",
condition=condition,
unconditional_condition=unconditional_condition,
guidance_scale=guidance_scale,
)
dpm_solver = DPM_Solver(model_fn, noise_schedule, algorithm_type="dpmsolver++",
correcting_x0_fn="dynamic_thresholding")
x_sample = dpm_solver.sample(
x_T,
steps=20,
order=2,
skip_type="logSNR",
method="multistep",
denoise_to_zero=True
)
`
from dpm-solver.
Hi guys, I've fixed the numerical issue in the cosine beta schedule; please try the newest file for dpmsolver and see details in this function.
You can also try this script with the ImageNet64 (improved-DDPM checkpoint) example
, which is a cosine schedule.
from dpm-solver.
Related Issues (20)
- Question regarding likelihood evaluation HOT 5
- Noisy results with "order == 1" (trying to replicate DDIM resutls) HOT 3
- discrete-time and continuous-time diffusion models HOT 5
- how to set parameters when sampling HOT 2
- sampling in a deterministic way like DDIM? HOT 2
- an error happens in "x0 = (x - sigma_t * noise) / alpha_t" HOT 2
- Control net HOT 5
- timestep_respacing HOT 2
- How to use dpm for guided diffision model(ddpm) HOT 4
- Is this error at single step 1st-order code? HOT 2
- Wonder you may integrate the dpm-solver to guided diffusion repo: HOT 1
- With all due respect which part of your model is this going to be added to, the training or the testing part? HOT 1
- Results of Imagenet256 using DPM++ not reproducible HOT 7
- Can DPM-Solver support self-conditioning? HOT 2
- No randomness is added during the sampling process? HOT 2
- Worse in high guidance scale HOT 4
- The results are blurry images with vanilla ddpm HOT 2
- problem with order 2 CIFAR-10 (VP deep continuous-time model ) HOT 1
- Multiple Guidance in Sampling? HOT 2
- How to use an accelerator to quickly predict noise reduction in sequence data (3D tensor)?
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from dpm-solver.