Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-18 16:54:33 +02:00
Code Issues Packages Projects Releases Wiki Activity

fix(pipelines): restore clip-skip text encoding for transformers 5.6+

Browse Source 
The vendored encode_prompt copies in the PAG, APG, ControlNet-XS and differential
diffusion pipelines dereference the .text_model wrapper that transformers 5.6
removed from CLIPTextModel, so their clip-skip path crashes on SD1.5 and SDXL TE1.
Apply the same getattr(te, 'text_model', te) fix as the core parser.

Mirrors upstream diffusers, which still carries this deref in pipeline
encode_prompt; only the single-file loader was fixed there.
This commit is contained in:
CalamitousFelicitousness
2026-06-25 02:06:05 +01:00
parent aeef359bd7
commit 3dd0cd590c
4 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/differential_diffusion.py
+2 -1
View File
@@ -1546,7 +1546,8 @@ class StableDiffusionDiffImg2ImgPipeline(DiffusionPipeline):
# representations. The `last_hidden_states` that we typically use for
# obtaining the final prompt representations passes through the LayerNorm
# layer.
prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)
# transformers >=5.6 flattened CLIPTextModel; CLIPTextModelWithProjection still nests it under .text_model
prompt_embeds = getattr(self.text_encoder, 'text_model', self.text_encoder).final_layer_norm(prompt_embeds)
if self.text_encoder is not None:
prompt_embeds_dtype = self.text_encoder.dtype