load_pipeline Logic Inverted for transformer_config

[StellarCheng]

System Info / 系統信息

   if self.transformer_config.get("image_dim", None) is not None:
         pipe = WanPipeline.from_pretrained(
             self.pretrained_model_name_or_path, **components, revision=self.revision, cache_dir=self.cache_dir
         )
     else:
         pipe = WanImageToVideoPipeline.from_pretrained(
            self.pretrained_model_name_or_path, **components, revision=self.revision, cache_dir=self.cache_dir
         )WanPipeline

It appears there's an inversion in the logic of the load_pipeline method in WanModelSpecification. Currently, when self.transformer_config is set to True, the code incorrectly loads the Text-to-Video (T2V) pipeline. Conversely, when self.transformer_config is None, it proceeds to load the Image-to-Video (I2V) pipeline.

Information / 问题信息

• The official example scripts / 官方的示例脚本
• My own modified scripts / 我自己修改的脚本和任务

Reproduction / 复现过程

e

Expected behavior / 期待表现

e

[tolgacangoz]

Hmm, right 👍. Would you like to open a PR to fix it?