mahadev-balla/isl-conditional-diffusion-linear-128

ISL Conditional Diffusion — Linear Noise Schedule

A class-conditioned DDPM trained from scratch to generate 128×128 RGB images of Indian Sign Language hand gestures using a linear noise schedule.

The model uses class conditioning on one of 35 ISL classes and supports classifier-free guidance during sampling.

The complete implementation and experiments are available in the GitHub repository.

Model

• —DDPM with UNet2D architecture
• —128×128 resolution
• —RGB images
• —Class-conditional generation
• —35 ISL classes
• —Classifier-free guidance (CFG)
• —EMA weights

Training

• —Dataset: 42,000 images (1,200 images per class × 35 classes)
• —Noise schedule: linear
• —Batch size: 64
• —Learning rate: 1e-4
• —Mixed precision: fp16
• —Training steps: 65,000
• —EMA decay: 0.9999
• —CFG label dropout: 0.15
• —Data augmentation: enabled

Sampling

• —Default sampler: DDIM
• —Training diffusion timesteps: 1,000
• —Routine inference steps: 100
• —Evaluation inference steps: 50
• —Default guidance scale: 1.0
• —Random seed: 42

Results

At the FID-optimal guidance scale of 1.0:

• —FID: 58.25
• —Semantic accuracy: 98.0%

The model provides strong class control under the reported evaluation setting.