https://github.com/aredden/flux-fp8-api

Flux diffusion model implementation using quantized fp8 matmul & remaining layers use faster half precision accumulate, which is ~2x faster on consumer devices.
https://github.com/aredden/flux-fp8-api

diffusion fast-inference flux fp8 pytorch quantization

Last synced: 2 days ago
JSON representation

Flux diffusion model implementation using quantized fp8 matmul & remaining layers use faster half precision accumulate, which is ~2x faster on consumer devices.

• Host: GitHub
• URL: https://github.com/aredden/flux-fp8-api
• Owner: aredden
• License: apache-2.0
• Created: 2024-08-05T20:04:25.000Z (about 2 months ago)
• Default Branch: main
• Last Pushed: 2024-09-13T16:01:24.000Z (9 days ago)
• Last Synced: 2024-09-14T06:56:28.331Z (8 days ago)
• Topics: diffusion, fast-inference, flux, fp8, pytorch, quantization
• Language: Python
• Homepage:
• Size: 124 KB
• Stars: 105
• Watchers: 6
• Forks: 12
• Open Issues: 7
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-flux-ai - Flux FP8 Matmul Implementation with FastAPI