[SD3 Inference] T5 Token limit (#8506)

* max_sequence_length for the T5

* updated img2img

* apply suggestions

---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: YiYi Xu <yixu310@gmail.com>

src/diffusers/pipelines/stable_diffusion_3/pipeline_stable_diffusion_3.py

@@ -205,6 +205,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
         self,
         prompt: Union[str, List[str]] = None,
         num_images_per_prompt: int = 1,
+        max_sequence_length: int = 256,
         device: Optional[torch.device] = None,
         dtype: Optional[torch.dtype] = None,
     ):
@@ -224,7 +225,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
         text_inputs = self.tokenizer_3(
             prompt,
             padding="max_length",
-            max_length=self.tokenizer_max_length,
+            max_length=max_sequence_length,
             truncation=True,
             add_special_tokens=True,
             return_tensors="pt",
@@ -235,8 +236,8 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
         if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(text_input_ids, untruncated_ids):
             removed_text = self.tokenizer_3.batch_decode(untruncated_ids[:, self.tokenizer_max_length - 1 : -1])
             logger.warning(
-                "The following part of your input was truncated because CLIP can only handle sequences up to"
-                f" {self.tokenizer_max_length} tokens: {removed_text}"
+                "The following part of your input was truncated because `max_sequence_length` is set to "
+                f" {max_sequence_length} tokens: {removed_text}"
             )
 
         prompt_embeds = self.text_encoder_3(text_input_ids.to(device))[0]
@@ -323,6 +324,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
         pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
         negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
         clip_skip: Optional[int] = None,
+        max_sequence_length: int = 256,
     ):
         r"""
 
@@ -403,6 +405,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
             t5_prompt_embed = self._get_t5_prompt_embeds(
                 prompt=prompt_3,
                 num_images_per_prompt=num_images_per_prompt,
+                max_sequence_length=max_sequence_length,
                 device=device,
             )
 
@@ -456,7 +459,10 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
             negative_clip_prompt_embeds = torch.cat([negative_prompt_embed, negative_prompt_2_embed], dim=-1)
 
             t5_negative_prompt_embed = self._get_t5_prompt_embeds(
-                prompt=negative_prompt_3, num_images_per_prompt=num_images_per_prompt, device=device
+                prompt=negative_prompt_3,
+                num_images_per_prompt=num_images_per_prompt,
+                max_sequence_length=max_sequence_length,
+                device=device,
             )
 
             negative_clip_prompt_embeds = torch.nn.functional.pad(
@@ -486,6 +492,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
         pooled_prompt_embeds=None,
         negative_pooled_prompt_embeds=None,
         callback_on_step_end_tensor_inputs=None,
+        max_sequence_length=None,
     ):
         if height % 8 != 0 or width % 8 != 0:
             raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")
@@ -557,6 +564,9 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
                 "If `negative_prompt_embeds` are provided, `negative_pooled_prompt_embeds` also have to be passed. Make sure to generate `negative_pooled_prompt_embeds` from the same text encoder that was used to generate `negative_prompt_embeds`."
             )
 
+        if max_sequence_length is not None and max_sequence_length > 512:
+            raise ValueError(f"`max_sequence_length` cannot be greater than 512 but is {max_sequence_length}")
+
     def prepare_latents(
         self,
         batch_size,
@@ -643,6 +653,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
         clip_skip: Optional[int] = None,
         callback_on_step_end: Optional[Callable[[int, int, Dict], None]] = None,
         callback_on_step_end_tensor_inputs: List[str] = ["latents"],
+        max_sequence_length: int = 256,
     ):
         r"""
         Function invoked when calling the pipeline for generation.
@@ -726,6 +737,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
                 The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list
                 will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the
                 `._callback_tensor_inputs` attribute of your pipeline class.
+            max_sequence_length (`int` defaults to 256): Maximum sequence length to use with the `prompt`.
 
         Examples:
 
@@ -753,6 +765,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
             pooled_prompt_embeds=pooled_prompt_embeds,
             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,
         )
 
         self._guidance_scale = guidance_scale
@@ -790,6 +803,7 @@ class StableDiffusion3Pipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingle
             device=device,
             clip_skip=self.clip_skip,
             num_images_per_prompt=num_images_per_prompt,
+            max_sequence_length=max_sequence_length,
         )
 
         if self.do_classifier_free_guidance:

src/diffusers/pipelines/stable_diffusion_3/pipeline_stable_diffusion_3_img2img.py

@@ -220,6 +220,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
         self,
         prompt: Union[str, List[str]] = None,
         num_images_per_prompt: int = 1,
+        max_sequence_length: int = 256,
         device: Optional[torch.device] = None,
         dtype: Optional[torch.dtype] = None,
     ):
@@ -239,7 +240,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
         text_inputs = self.tokenizer_3(
             prompt,
             padding="max_length",
-            max_length=self.tokenizer_max_length,
+            max_length=max_sequence_length,
             truncation=True,
             add_special_tokens=True,
             return_tensors="pt",
@@ -250,8 +251,8 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
         if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(text_input_ids, untruncated_ids):
             removed_text = self.tokenizer_3.batch_decode(untruncated_ids[:, self.tokenizer_max_length - 1 : -1])
             logger.warning(
-                "The following part of your input was truncated because CLIP can only handle sequences up to"
-                f" {self.tokenizer_max_length} tokens: {removed_text}"
+                "The following part of your input was truncated because `max_sequence_length` is set to "
+                f" {max_sequence_length} tokens: {removed_text}"
             )
 
         prompt_embeds = self.text_encoder_3(text_input_ids.to(device))[0]
@@ -340,6 +341,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
         pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
         negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
         clip_skip: Optional[int] = None,
+        max_sequence_length: int = 256,
     ):
         r"""
 
@@ -420,6 +422,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
             t5_prompt_embed = self._get_t5_prompt_embeds(
                 prompt=prompt_3,
                 num_images_per_prompt=num_images_per_prompt,
+                max_sequence_length=max_sequence_length,
                 device=device,
             )
 
@@ -473,7 +476,10 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
             negative_clip_prompt_embeds = torch.cat([negative_prompt_embed, negative_prompt_2_embed], dim=-1)
 
             t5_negative_prompt_embed = self._get_t5_prompt_embeds(
-                prompt=negative_prompt_3, num_images_per_prompt=num_images_per_prompt, device=device
+                prompt=negative_prompt_3,
+                num_images_per_prompt=num_images_per_prompt,
+                max_sequence_length=max_sequence_length,
+                device=device,
             )
 
             negative_clip_prompt_embeds = torch.nn.functional.pad(
@@ -502,6 +508,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
         pooled_prompt_embeds=None,
         negative_pooled_prompt_embeds=None,
         callback_on_step_end_tensor_inputs=None,
+        max_sequence_length=None,
     ):
         if strength < 0 or strength > 1:
             raise ValueError(f"The value of strength should in [0.0, 1.0] but is {strength}")
@@ -573,6 +580,9 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
                 "If `negative_prompt_embeds` are provided, `negative_pooled_prompt_embeds` also have to be passed. Make sure to generate `negative_pooled_prompt_embeds` from the same text encoder that was used to generate `negative_prompt_embeds`."
             )
 
+        if max_sequence_length is not None and max_sequence_length > 512:
+            raise ValueError(f"`max_sequence_length` cannot be greater than 512 but is {max_sequence_length}")
+
     def get_timesteps(self, num_inference_steps, strength, device):
         # get the original timestep using init_timestep
         init_timestep = min(num_inference_steps * strength, num_inference_steps)
@@ -684,6 +694,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
         clip_skip: Optional[int] = None,
         callback_on_step_end: Optional[Callable[[int, int, Dict], None]] = None,
         callback_on_step_end_tensor_inputs: List[str] = ["latents"],
+        max_sequence_length: int = 256,
     ):
         r"""
         Function invoked when calling the pipeline for generation.
@@ -763,6 +774,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
                 The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list
                 will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the
                 `._callback_tensor_inputs` attribute of your pipeline class.
+            max_sequence_length (`int` defaults to 256): Maximum sequence length to use with the `prompt`.
 
         Examples:
 
@@ -786,6 +798,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
             pooled_prompt_embeds=pooled_prompt_embeds,
             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,
         )
 
         self._guidance_scale = guidance_scale
@@ -822,6 +835,7 @@ class StableDiffusion3Img2ImgPipeline(DiffusionPipeline):
             device=device,
             clip_skip=self.clip_skip,
             num_images_per_prompt=num_images_per_prompt,
+            max_sequence_length=max_sequence_length,
         )
 
         if self.do_classifier_free_guidance: