Skip to content

Releases: mosaicml/diffusion

v0.1.3 - CLIP + T5

21 Nov 20:06
ba8ca02
Compare
Choose a tag to compare

What's Changed

  • Add docker images for diffusion repo by @j316chuck in #174
  • Add option for default prompts/negative prompts in eval by @coryMosaicML in #148
  • Add image generator to generate images for use with geneval by @coryMosaicML in #172
  • Add composer model class for running with precomputed CLIP and T5 text latents by @coryMosaicML in #171
  • Update latent precomputation script with batching by @jazcollins in #170
  • Update dependencies for use with torch 2.4.1 and composer 0.25.0 by @coryMosaicML in #176
  • Add option for stream weights in the image caption latents dataloader by @coryMosaicML in #178
  • Update evaluation and inference code to handle other precisions and models by @coryMosaicML in #179

New Contributors

Full Changelog: v0.1.2...v0.1.3

v0.1.2

05 Sep 21:52
c1f953f
Compare
Choose a tag to compare

What's Changed

New Contributors

Full Changelog: v0.1.1...v0.1.2

v0.1.1

18 Jun 21:52
93a5469
Compare
Choose a tag to compare

Minor bug fix related to max_pad_tokens at generate time. Other Noise Schedule related features and options.

What's Changed

Full Changelog: v0.1.0...v0.1.1

v0.1.0

22 Apr 20:25
41b13bc
Compare
Choose a tag to compare

What's Changed

New Contributors

Full Changelog: v0.0.1...v0.1.0

v0.0.1

13 Nov 22:53
3122b81
Compare
Choose a tag to compare
Add callback to catch NaNs in the train loss (#97)