Motif Mega 187M

Created by Léon (13 years old) - Motif AI Labs

Specifications

Spec Value
Parameters 187,730,881
Architecture DiT
Image Size 128x128
Patch Size 8x8
Hidden Size 768
Depth 16
Heads 12

Training

  • Dataset : 64,000 examples
  • Steps : 8,000
  • Final Loss : 0.0015
  • Average Loss : 0.0475

Techniques

  • DiT Architecture
  • AdaLN-Zero Conditioning
  • Pixel Precision Upscaler 4x
  • Prompt Reread
  • Multi-Layer Precision
  • Web Search Enhancement
  • Mega Quality Tags

Links

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support