[Dance Diffusion] FP16 (#980)

* add in fp16

* up

src/diffusers/models/unet_1d.py

@@ -149,7 +149,7 @@ class UNet1DModel(ModelMixin, ConfigMixin):
             timestep = timestep[None]
 
         timestep_embed = self.time_proj(timestep)[..., None]
-        timestep_embed = timestep_embed.repeat([1, 1, sample.shape[2]])
+        timestep_embed = timestep_embed.repeat([1, 1, sample.shape[2]]).to(sample.dtype)
 
         # 2. down
         down_block_res_samples = ()

src/diffusers/pipelines/dance_diffusion/pipeline_dance_diffusion.py

@@ -91,10 +91,14 @@ class DanceDiffusionPipeline(DiffusionPipeline):
             )
         sample_size = int(sample_size)
 
-        audio = torch.randn((batch_size, self.unet.in_channels, sample_size), generator=generator, device=self.device)
+        dtype = next(iter(self.unet.parameters())).dtype
+        audio = torch.randn(
+            (batch_size, self.unet.in_channels, sample_size), generator=generator, device=self.device, dtype=dtype
+        )
 
         # set step values
         self.scheduler.set_timesteps(num_inference_steps, device=audio.device)
+        self.scheduler.timesteps = self.scheduler.timesteps.to(dtype)
 
         for t in self.progress_bar(self.scheduler.timesteps):
             # 1. predict noise model_output
@@ -103,7 +107,7 @@ class DanceDiffusionPipeline(DiffusionPipeline):
             # 2. compute previous image: x_t -> t_t-1
             audio = self.scheduler.step(model_output, t, audio).prev_sample
 
-        audio = audio.clamp(-1, 1).cpu().numpy()
+        audio = audio.clamp(-1, 1).float().cpu().numpy()
 
         audio = audio[:, :, :original_sample_size]
 

tests/pipelines/dance_diffusion/test_dance_diffusion.py

@@ -99,3 +99,20 @@ class PipelineIntegrationTests(unittest.TestCase):
         assert audio.shape == (1, 2, pipe.unet.sample_size)
         expected_slice = np.array([-0.1576, -0.1526, -0.127, -0.2699, -0.2762, -0.2487])
         assert np.abs(audio_slice.flatten() - expected_slice).max() < 1e-2
+
+    def test_dance_diffusion_fp16(self):
+        device = torch_device
+
+        pipe = DanceDiffusionPipeline.from_pretrained("harmonai/maestro-150k", torch_dtype=torch.float16)
+        pipe = pipe.to(device)
+        pipe.set_progress_bar_config(disable=None)
+
+        generator = torch.Generator(device=device).manual_seed(0)
+        output = pipe(generator=generator, num_inference_steps=100, sample_length_in_s=4.096)
+        audio = output.audios
+
+        audio_slice = audio[0, -3:, -3:]
+
+        assert audio.shape == (1, 2, pipe.unet.sample_size)
+        expected_slice = np.array([-0.1693, -0.1698, -0.1447, -0.3044, -0.3203, -0.2937])
+        assert np.abs(audio_slice.flatten() - expected_slice).max() < 1e-2