ddpm-cifar10-32-finetuned-2000steps-20251204
Fine-tuned DDPM model based on google/ddpm-cifar10-32.
Training Details
- Base Model: google/ddpm-cifar10-32
- Training Scenario: representative_mix (20% clean + 80% corrupted CIFAR-10)
- Corruptions: 4 representative types at severity 3
- Training Steps: finetuned
Files
unet_model.pth: Fine-tuned UNet model weightsmetadata.json: Training metadata and configuration
Usage
Load this checkpoint using the CheckpointManager from the training repository.
