Update pipeline.py

pipeline.py

@@ -592,6 +592,7 @@ class FluxWithCFGPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleFile
             negative_pooled_prompt_embeds=negative_pooled_prompt_embeds,
             callback_on_step_end_tensor_inputs=callback_on_step_end_tensor_inputs,
             max_sequence_length=max_sequence_length,
+            lora_scale=lora_scale
         )
 
         self._guidance_scale = guidance_scale
@@ -615,33 +616,6 @@ class FluxWithCFGPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleFile
             self.joint_attention_kwargs.get("scale", None) if self.joint_attention_kwargs is not None else None
         )
 
-        # 3. Encode prompt
-        lora_scale = (
-            self.joint_attention_kwargs.get("scale", None) if self.joint_attention_kwargs is not None else None
-        )
-        (
-            prompt_embeds,
-            negative_prompt_embeds,
-            pooled_prompt_embeds,
-            negative_pooled_prompt_embeds,
-            text_ids,
-        ) = self.encode_prompt(
-            prompt=prompt,
-            prompt_2=prompt_2,
-            negative_prompt=negative_prompt,
-            negative_prompt_2=negative_prompt_2,
-            do_classifier_free_guidance=self.do_classifier_free_guidance,
-            prompt_embeds=prompt_embeds,
-            negative_prompt_embeds=negative_prompt_embeds,
-            pooled_prompt_embeds=pooled_prompt_embeds,
-            negative_pooled_prompt_embeds=negative_pooled_prompt_embeds,
-            device=device,
-            clip_skip=self.clip_skip,
-            num_images_per_prompt=num_images_per_prompt,
-            max_sequence_length=max_sequence_length,
-            lora_scale=lora_scale,
-        )
-
         if self.do_classifier_free_guidance:
             prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds], dim=0)
             pooled_prompt_embeds = torch.cat([negative_pooled_prompt_embeds, pooled_prompt_embeds], dim=0)