Fix/update the LDM pipeline and tests (huggingface#1743)

* Fix/update LDM tests

* batched generators

Author: anton-l

src/diffusers/pipelines/latent_diffusion/pipeline_latent_diffusion.py

@@ -128,29 +128,42 @@ def __call__(
 
         # get unconditional embeddings for classifier free guidance
         if guidance_scale != 1.0:
-            uncond_input = self.tokenizer([""] * batch_size, padding="max_length", max_length=77, return_tensors="pt")
+            uncond_input = self.tokenizer(
+                [""] * batch_size, padding="max_length", max_length=77, truncation=True, return_tensors="pt"
+            )
             uncond_embeddings = self.bert(uncond_input.input_ids.to(self.device))[0]
 
         # get prompt text embeddings
-        text_input = self.tokenizer(prompt, padding="max_length", max_length=77, return_tensors="pt")
+        text_input = self.tokenizer(prompt, padding="max_length", max_length=77, truncation=True, return_tensors="pt")
         text_embeddings = self.bert(text_input.input_ids.to(self.device))[0]
 
         # get the initial random noise unless the user supplied it
         latents_shape = (batch_size, self.unet.in_channels, height // 8, width // 8)
+        if isinstance(generator, list) and len(generator) != batch_size:
+            raise ValueError(
+                f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
+                f" size of {batch_size}. Make sure the batch size matches the length of the generators."
+            )
+
         if latents is None:
-            if self.device.type == "mps":
-                # randn does not work reproducibly on mps
-                latents = torch.randn(latents_shape, generator=generator, device="cpu").to(self.device)
+            rand_device = "cpu" if self.device.type == "mps" else self.device
+
+            if isinstance(generator, list):
+                latents_shape = (1,) + latents_shape[1:]
+                latents = [
+                    torch.randn(latents_shape, generator=generator[i], device=rand_device, dtype=text_embeddings.dtype)
+                    for i in range(batch_size)
+                ]
+                latents = torch.cat(latents, dim=0)
             else:
                 latents = torch.randn(
-                    latents_shape,
-                    generator=generator,
-                    device=self.device,
+                    latents_shape, generator=generator, device=rand_device, dtype=text_embeddings.dtype
                 )
+            latents = latents.to(self.device)
         else:
             if latents.shape != latents_shape:
                 raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {latents_shape}")
-            latents = latents.to(self.device)
+        latents = latents.to(self.device)
 
         self.scheduler.set_timesteps(num_inference_steps)
 

tests/pipelines/latent_diffusion/test_latent_diffusion.py

@@ -13,24 +13,29 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import gc
 import unittest
 
 import numpy as np
 import torch
 
 from diffusers import AutoencoderKL, DDIMScheduler, LDMTextToImagePipeline, UNet2DConditionModel
-from diffusers.utils.testing_utils import require_torch, slow, torch_device
+from diffusers.utils.testing_utils import load_numpy, nightly, require_torch_gpu, slow, torch_device
 from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
 
+from ...test_pipelines_common import PipelineTesterMixin
+
 
 torch.backends.cuda.matmul.allow_tf32 = False
 
 
-class LDMTextToImagePipelineFastTests(unittest.TestCase):
-    @property
-    def dummy_cond_unet(self):
+class LDMTextToImagePipelineFastTests(PipelineTesterMixin, unittest.TestCase):
+    pipeline_class = LDMTextToImagePipeline
+    test_cpu_offload = False
+
+    def get_dummy_components(self):
         torch.manual_seed(0)
-        model = UNet2DConditionModel(
+        unet = UNet2DConditionModel(
             block_out_channels=(32, 64),
             layers_per_block=2,
             sample_size=32,
@@ -40,25 +45,24 @@ def dummy_cond_unet(self):
             up_block_types=("CrossAttnUpBlock2D", "UpBlock2D"),
             cross_attention_dim=32,
         )
-        return model
-
-    @property
-    def dummy_vae(self):
+        scheduler = DDIMScheduler(
+            beta_start=0.00085,
+            beta_end=0.012,
+            beta_schedule="scaled_linear",
+            clip_sample=False,
+            set_alpha_to_one=False,
+        )
         torch.manual_seed(0)
-        model = AutoencoderKL(
-            block_out_channels=[32, 64],
+        vae = AutoencoderKL(
+            block_out_channels=(32, 64),
             in_channels=3,
             out_channels=3,
-            down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"],
-            up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"],
+            down_block_types=("DownEncoderBlock2D", "DownEncoderBlock2D"),
+            up_block_types=("UpDecoderBlock2D", "UpDecoderBlock2D"),
             latent_channels=4,
         )
-        return model
-
-    @property
-    def dummy_text_encoder(self):
         torch.manual_seed(0)
-        config = CLIPTextConfig(
+        text_encoder_config = CLIPTextConfig(
             bos_token_id=0,
             eos_token_id=2,
             hidden_size=32,
@@ -69,96 +73,117 @@ def dummy_text_encoder(self):
             pad_token_id=1,
             vocab_size=1000,
         )
-        return CLIPTextModel(config)
-
-    def test_inference_text2img(self):
-        if torch_device != "cpu":
-            return
-
-        unet = self.dummy_cond_unet
-        scheduler = DDIMScheduler()
-        vae = self.dummy_vae
-        bert = self.dummy_text_encoder
+        text_encoder = CLIPTextModel(text_encoder_config)
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
 
-        ldm = LDMTextToImagePipeline(vqvae=vae, bert=bert, tokenizer=tokenizer, unet=unet, scheduler=scheduler)
-        ldm.to(torch_device)
-        ldm.set_progress_bar_config(disable=None)
-
-        prompt = "A painting of a squirrel eating a burger"
-
-        # Warmup pass when using mps (see #372)
-        if torch_device == "mps":
-            generator = torch.manual_seed(0)
-            _ = ldm(
-                [prompt], generator=generator, guidance_scale=6.0, num_inference_steps=1, output_type="numpy"
-            ).images
-
-        device = torch_device if torch_device != "mps" else "cpu"
-        generator = torch.Generator(device=device).manual_seed(0)
-
-        image = ldm(
-            [prompt], generator=generator, guidance_scale=6.0, num_inference_steps=2, output_type="numpy"
-        ).images
-
-        device = torch_device if torch_device != "mps" else "cpu"
-        generator = torch.Generator(device=device).manual_seed(0)
-
-        image_from_tuple = ldm(
-            [prompt],
-            generator=generator,
-            guidance_scale=6.0,
-            num_inference_steps=2,
-            output_type="numpy",
-            return_dict=False,
-        )[0]
+        components = {
+            "unet": unet,
+            "scheduler": scheduler,
+            "vqvae": vae,
+            "bert": text_encoder,
+            "tokenizer": tokenizer,
+        }
+        return components
+
+    def get_dummy_inputs(self, device, seed=0):
+        if str(device).startswith("mps"):
+            generator = torch.manual_seed(seed)
+        else:
+            generator = torch.Generator(device=device).manual_seed(seed)
+        inputs = {
+            "prompt": "A painting of a squirrel eating a burger",
+            "generator": generator,
+            "num_inference_steps": 2,
+            "guidance_scale": 6.0,
+            "output_type": "numpy",
+        }
+        return inputs
 
-        image_slice = image[0, -3:, -3:, -1]
-        image_from_tuple_slice = image_from_tuple[0, -3:, -3:, -1]
-
-        assert image.shape == (1, 16, 16, 3)
-        expected_slice = np.array([0.6806, 0.5454, 0.5638, 0.4893, 0.4656, 0.4257, 0.6248, 0.5217, 0.5498])
-        assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2
-        assert np.abs(image_from_tuple_slice.flatten() - expected_slice).max() < 1e-2
-
-
-@slow
-@require_torch
-class LDMTextToImagePipelineIntegrationTests(unittest.TestCase):
     def test_inference_text2img(self):
-        ldm = LDMTextToImagePipeline.from_pretrained("CompVis/ldm-text2im-large-256")
-        ldm.to(torch_device)
-        ldm.set_progress_bar_config(disable=None)
-
-        prompt = "A painting of a squirrel eating a burger"
-
-        device = torch_device if torch_device != "mps" else "cpu"
-        generator = torch.Generator(device=device).manual_seed(0)
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
 
-        image = ldm(
-            [prompt], generator=generator, guidance_scale=6.0, num_inference_steps=20, output_type="numpy"
-        ).images
+        components = self.get_dummy_components()
+        pipe = LDMTextToImagePipeline(**components)
+        pipe.to(device)
+        pipe.set_progress_bar_config(disable=None)
 
+        inputs = self.get_dummy_inputs(device)
+        image = pipe(**inputs).images
         image_slice = image[0, -3:, -3:, -1]
 
-        assert image.shape == (1, 256, 256, 3)
-        expected_slice = np.array([0.9256, 0.9340, 0.8933, 0.9361, 0.9113, 0.8727, 0.9122, 0.8745, 0.8099])
-        assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2
-
-    def test_inference_text2img_fast(self):
-        ldm = LDMTextToImagePipeline.from_pretrained("CompVis/ldm-text2im-large-256")
-        ldm.to(torch_device)
-        ldm.set_progress_bar_config(disable=None)
-
-        prompt = "A painting of a squirrel eating a burger"
+        assert image.shape == (1, 16, 16, 3)
+        expected_slice = np.array([0.59450, 0.64078, 0.55509, 0.51229, 0.69640, 0.36960, 0.59296, 0.60801, 0.49332])
 
-        device = torch_device if torch_device != "mps" else "cpu"
-        generator = torch.Generator(device=device).manual_seed(0)
+        assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-3
 
-        image = ldm(prompt, generator=generator, num_inference_steps=1, output_type="numpy").images
 
-        image_slice = image[0, -3:, -3:, -1]
+@slow
+@require_torch_gpu
+class LDMTextToImagePipelineSlowTests(unittest.TestCase):
+    def tearDown(self):
+        super().tearDown()
+        gc.collect()
+        torch.cuda.empty_cache()
+
+    def get_inputs(self, device, dtype=torch.float32, seed=0):
+        generator = torch.Generator(device=device).manual_seed(seed)
+        latents = np.random.RandomState(seed).standard_normal((1, 4, 32, 32))
+        latents = torch.from_numpy(latents).to(device=device, dtype=dtype)
+        inputs = {
+            "prompt": "A painting of a squirrel eating a burger",
+            "latents": latents,
+            "generator": generator,
+            "num_inference_steps": 3,
+            "guidance_scale": 6.0,
+            "output_type": "numpy",
+        }
+        return inputs
+
+    def test_ldm_default_ddim(self):
+        pipe = LDMTextToImagePipeline.from_pretrained("CompVis/ldm-text2im-large-256").to(torch_device)
+        pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_inputs(torch_device)
+        image = pipe(**inputs).images
+        image_slice = image[0, -3:, -3:, -1].flatten()
 
         assert image.shape == (1, 256, 256, 3)
-        expected_slice = np.array([0.3163, 0.8670, 0.6465, 0.1865, 0.6291, 0.5139, 0.2824, 0.3723, 0.4344])
-        assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2
+        expected_slice = np.array([0.51825, 0.52850, 0.52543, 0.54258, 0.52304, 0.52569, 0.54363, 0.55276, 0.56878])
+        max_diff = np.abs(expected_slice - image_slice).max()
+        assert max_diff < 1e-3
+
+
+@nightly
+@require_torch_gpu
+class LDMTextToImagePipelineNightlyTests(unittest.TestCase):
+    def tearDown(self):
+        super().tearDown()
+        gc.collect()
+        torch.cuda.empty_cache()
+
+    def get_inputs(self, device, dtype=torch.float32, seed=0):
+        generator = torch.Generator(device=device).manual_seed(seed)
+        latents = np.random.RandomState(seed).standard_normal((1, 4, 32, 32))
+        latents = torch.from_numpy(latents).to(device=device, dtype=dtype)
+        inputs = {
+            "prompt": "A painting of a squirrel eating a burger",
+            "latents": latents,
+            "generator": generator,
+            "num_inference_steps": 50,
+            "guidance_scale": 6.0,
+            "output_type": "numpy",
+        }
+        return inputs
+
+    def test_ldm_default_ddim(self):
+        pipe = LDMTextToImagePipeline.from_pretrained("CompVis/ldm-text2im-large-256").to(torch_device)
+        pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_inputs(torch_device)
+        image = pipe(**inputs).images[0]
+
+        expected_image = load_numpy(
+            "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main/ldm_text2img/ldm_large_256_ddim.npy"
+        )
+        max_diff = np.abs(expected_image - image).max()
+        assert max_diff < 1e-3