xhinker/sd_embed

Issues

• Error with SD3.5

  #26 opened 2 months ago by blx0102
  0

• get_weighted_text_embeddings_flux1 yields an error for too many tokens for the model

  #11 opened 5 months ago by jonathanbiard
  4

• ROCm support

  #25 opened 2 months ago by renich
  0

• Excuse me, I wanted to ask if there is support for textual inversion?,

  #23 opened 2 months ago by Eduardishion
  3

• in IterComp，how to use sd_embed？

  #24 opened 2 months ago by xddun
  1

• Does not work for long prompt (SD 1.5)

  #22 opened 2 months ago by larinius
  1

• Stable Cascade support, reference implementation for consideration

  #18 opened 4 months ago by Teriks
  3

• Support for pipe.enable_model_cpu_offload() and pipe.enable_sequential_cpu_offload()

  #2 opened 6 months ago by Teriks
  8

• No support for CPU

  #16 opened 4 months ago by davidrippel
  4

• Flux.1 and OSX with M1

  #14 opened 4 months ago by odoremieux
  5

• Not able to run the flux example: "Couldn't swap QLinear.weight"

  #13 opened 4 months ago by MisterTea
  1

• get_weighted_text_embeddings_flux1 device selection meta tensor copy

  #12 opened 5 months ago by Teriks
  2

• Add Flux Support

  #9 opened 5 months ago by MohamedAliRashad
  9

• Error when using pre-SD3 diffusers versions

  #7 opened 5 months ago by rileybolen
  1

• `get_weighted_text_embeddings_sdxl` Fails with Empty Prompt

  #5 opened 5 months ago by thangnd-zenai
  2

• SD3, use_t5_encoder=False, embeddings size missmatch

  #1 opened 6 months ago by Daniel-SicSo-Edinburgh
  1