[Img2Img] Fix batch size mismatch prompts vs. init images (huggingfac…

…e#793)

* [Img2Img] Fix batch size mismatch prompts vs. init images

* Remove bogus folder

* fix

* Update src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py

Co-authored-by: Pedro Cuenca <[email protected]>

Co-authored-by: Pedro Cuenca <[email protected]>

src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py

@@ -195,6 +195,7 @@ def __call__(
         """
         if isinstance(prompt, str):
             batch_size = 1
+            prompt = [prompt]
         elif isinstance(prompt, list):
             batch_size = len(prompt)
         else:
@@ -284,8 +285,23 @@ def __call__(
         init_latents = init_latent_dist.sample(generator=generator)
         init_latents = 0.18215 * init_latents
 
-        # expand init_latents for batch_size
-        init_latents = torch.cat([init_latents] * batch_size * num_images_per_prompt, dim=0)
+        if len(prompt) > init_latents.shape[0] and len(prompt) % init_latents.shape[0] == 0:
+            # expand init_latents for batch_size
+            deprecation_message = (
+                f"You have passed {len(prompt)} text prompts (`prompt`), but only {init_latents.shape[0]} initial"
+                " images (`init_image`). Initial images are now duplicating to match the number of text prompts. Note"
+                " that this behavior is deprecated and will be removed in a version 1.0.0. Please make sure to update"
+                " your script to pass as many init images as text prompts to suppress this warning."
+            )
+            deprecate("len(prompt) != len(init_image)", "1.0.0", deprecation_message, standard_warn=False)
+            additional_image_per_prompt = len(prompt) // init_latents.shape[0]
+            init_latents = torch.cat([init_latents] * additional_image_per_prompt * num_images_per_prompt, dim=0)
+        elif len(prompt) > init_latents.shape[0] and len(prompt) % init_latents.shape[0] != 0:
+            raise ValueError(
+                f"Cannot duplicate `init_image` of batch size {init_latents.shape[0]} to {len(prompt)} text prompts."
+            )
+        else:
+            init_latents = torch.cat([init_latents] * num_images_per_prompt, dim=0)
 
         # get the original timestep using init_timestep
         offset = self.scheduler.config.get("steps_offset", 0)

tests/test_pipelines.py

@@ -698,6 +698,48 @@ def test_stable_diffusion_img2img(self):
         assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2
         assert np.abs(image_from_tuple_slice.flatten() - expected_slice).max() < 1e-2
 
+    def test_stable_diffusion_img2img_multiple_init_images(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        unet = self.dummy_cond_unet
+        scheduler = PNDMScheduler(skip_prk_steps=True)
+        vae = self.dummy_vae
+        bert = self.dummy_text_encoder
+        tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
+
+        init_image = self.dummy_image.to(device).repeat(2, 1, 1, 1)
+
+        # make sure here that pndm scheduler skips prk
+        sd_pipe = StableDiffusionImg2ImgPipeline(
+            unet=unet,
+            scheduler=scheduler,
+            vae=vae,
+            text_encoder=bert,
+            tokenizer=tokenizer,
+            safety_checker=self.dummy_safety_checker,
+            feature_extractor=self.dummy_extractor,
+        )
+        sd_pipe = sd_pipe.to(device)
+        sd_pipe.set_progress_bar_config(disable=None)
+
+        prompt = 2 * ["A painting of a squirrel eating a burger"]
+        generator = torch.Generator(device=device).manual_seed(0)
+        output = sd_pipe(
+            prompt,
+            generator=generator,
+            guidance_scale=6.0,
+            num_inference_steps=2,
+            output_type="np",
+            init_image=init_image,
+        )
+
+        image = output.images
+
+        image_slice = image[-1, -3:, -3:, -1]
+
+        assert image.shape == (2, 32, 32, 3)
+        expected_slice = np.array([0.5144, 0.4447, 0.4735, 0.6676, 0.5526, 0.5454, 0.645, 0.5149, 0.4689])
+        assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2
+
     def test_stable_diffusion_img2img_k_lms(self):
         device = "cpu"  # ensure determinism for the device-dependent torch.Generator
         unet = self.dummy_cond_unet