AI/diffusers
1
0
Fork 0
mirror of https://github.com/huggingface/diffusers.git synced 2026-01-27 17:22:53 +03:00
Code Activity

Fix more bugs in LTX2Pipeline.__call__

Browse Source
This commit is contained in:
Daniel Gu
2025-12-23 10:41:27 +01:00
parent a56cf23483
commit 90edc6abc9
1 changed files with 11 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
src/diffusers/pipelines/ltx2/pipeline_ltx2.py
View File

@@ -888,16 +888,6 @@ class LTX2Pipeline(DiffusionPipeline, FromSingleFileMixin, LTXVideoLoraLoaderMix
self.scheduler.config.get("base_shift", 0.95),
self.scheduler.config.get("max_shift", 2.05),
)
timesteps, num_inference_steps = retrieve_timesteps(
self.scheduler,
num_inference_steps,
device,
timesteps,
sigmas=sigmas,
mu=mu,
)
num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
self._num_timesteps = len(timesteps)
# For now, duplicate the scheduler for use with the audio latents
audio_scheduler = copy.deepcopy(self.scheduler)
_, _ = retrieve_timesteps(
@@ -908,6 +898,16 @@ class LTX2Pipeline(DiffusionPipeline, FromSingleFileMixin, LTXVideoLoraLoaderMix
sigmas=sigmas,
mu=mu,
)
timesteps, num_inference_steps = retrieve_timesteps(
self.scheduler,
num_inference_steps,
device,
timesteps,
sigmas=sigmas,
mu=mu,
)
num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
self._num_timesteps = len(timesteps)
# 6. Prepare micro-conditions
rope_interpolation_scale = (
@@ -937,7 +937,7 @@ class LTX2Pipeline(DiffusionPipeline, FromSingleFileMixin, LTXVideoLoraLoaderMix
hidden_states=latent_model_input,
audio_hidden_states=audio_latent_model_input,
encoder_hidden_states=prompt_embeds,
audio_encoder_hidden_states=audio_latent_model_input,
audio_encoder_hidden_states=audio_prompt_embeds,
timestep=timestep,
encoder_attention_mask=prompt_attention_mask,
audio_encoder_attention_mask=prompt_attention_mask,