πŸŽ‰ 75% of content is free forever β€” Unlock Premium from $10/mo β†’
CW
Search courses…
πŸ’Ό Servicesℹ️ Aboutβœ‰οΈ ContactView Pricing Plansfrom $10

Consistency Models: Single-Step Generation

Generative AIConsistency Models: Single-Step Generation🟒 Free Lesson

Advertisement

Consistency Models: Single-Step Generation

Module: Generative AI | Difficulty: Advanced

Consistency Function

for all along the ODE trajectory.

Boundary Condition

Self-Consistency Property

for any on the same trajectory.

Training

class ConsistencyModel(nn.Module):
    def __init__(self, model, sigma_min=0.002, sigma_max=80):
        super().__init__()
        self.model = model
        self.sigma_min = sigma_min
        self.sigma_max = sigma_max
    def forward(self, x, sigma):
        c_in = 1 / (sigma**2 + 1)**0.5
        c_out = sigma / (sigma**2 + 1)**0.5
        return x + c_out * self.model(c_in * x, sigma)
MethodStepsFIDIS
Consistency (train)14.125.2
Consistency (distill)13.456.1
Consistency (2 steps)23.127.3
Consistency (4 steps)42.988.1

Need Expert Generative AI Help?

Get personalized tutoring, project support, or professional consulting.

Advertisement