Update handler.py

handler.py

@@ -1,70 +1,57 @@
-from typing import Dict, List, Any
+from typing import cast, Union
 
+import PIL.Image
 import torch
 
 from diffusers import AutoencoderKLHunyuanVideo
 from diffusers.video_processor import VideoProcessor
-from safetensors.torch import _tobytes
-
-
-def unpack_tensor(tensor_data: bytes, shape: List[int], dtype: str) -> torch.Tensor:
-    tensor = tensor_data
-    DTYPE_MAP = {
-        "float16": torch.float16,
-        "float32": torch.float32,
-        "bfloat16": torch.bfloat16,
-    }
-    torch_dtype = DTYPE_MAP.get(dtype)
-    tensor = torch.frombuffer(bytearray(tensor), dtype=torch_dtype).reshape(shape)
-    return tensor
-
+from diffusers.utils import export_to_video
 
 class EndpointHandler:
     def __init__(self, path=""):
         self.device = "cuda"
         self.dtype = torch.float16
-        self.vae = (
-            AutoencoderKLHunyuanVideo.from_pretrained(
-                path, subfolder="vae", torch_dtype=self.dtype
-            )
-            .to(self.device, self.dtype)
-            .eval()
-        )
+        self.vae = cast(AutoencoderKLHunyuanVideo, AutoencoderKLHunyuanVideo.from_pretrained(path, torch_dtype=self.dtype).to(self.device, self.dtype).eval())
 
-        self.vae_scale_factor_spatial = self.vae.spatial_compression_ratio
+        self.vae_scale_factor = self.vae_scale_factor_spatial = self.vae.spatial_compression_ratio
         self.video_processor = VideoProcessor(
             vae_scale_factor=self.vae_scale_factor_spatial
         )
 
     @torch.no_grad()
-    def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
+    def __call__(self, data) -> Union[torch.Tensor, bytes]:
         """
         Args:
             data (:obj:):
                 includes the input data and the parameters for the inference.
         """
-        tensor_data = data["inputs"]
-        parameters = data.get("parameters", {})
-        if "shape" not in parameters:
-            raise ValueError("Expected `shape` in parameters.")
-        if "dtype" not in parameters:
-            raise ValueError("Expected `dtype` in parameters.")
-
-        shape = parameters.get("shape")
-        dtype = parameters.get("dtype")
-
-        tensor = unpack_tensor(tensor_data, shape, dtype)
+        tensor = cast(torch.Tensor, data["inputs"])
+        parameters = cast(dict, data.get("parameters", {}))
+        do_scaling = cast(bool, parameters.get("do_scaling", True))
+        output_type = cast(str, parameters.get("output_type", "pil"))
+        partial_postprocess = cast(bool, parameters.get("partial_postprocess", False))
+        if partial_postprocess and output_type != "pt":
+            output_type = "pt"
 
         tensor = tensor.to(self.device, self.dtype)
 
-        tensor = tensor / self.vae.config.scaling_factor
+        if do_scaling:
+            tensor = tensor / self.vae.config.scaling_factor
 
         with torch.no_grad():
-            frames = self.vae.decode(tensor, return_dict=False)[0]
-
-        frames = frames[0].permute(1, 0, 2, 3)
-        frames = torch.stack([(frame * 0.5 + 0.5).clamp(0, 1) for frame in frames])
-        frames = frames.permute(0, 2, 3, 1).contiguous().float()
-        frames = (frames * 255).round().to(torch.uint8)
-
-        return _tobytes(frames, "frames")
+            frames = cast(torch.Tensor, self.vae.decode(tensor, return_dict=False)[0])
+
+        if partial_postprocess:
+            frames = frames[0].permute(1, 0, 2, 3)
+            frames = torch.stack([(frame * 0.5 + 0.5).clamp(0, 1) for frame in frames])
+            frames = frames.permute(0, 2, 3, 1).contiguous().float()
+            frames = (frames * 255).round().to(torch.uint8)
+        elif output_type == "pil" or output_type == "pt":
+            frames = cast(torch.Tensor, self.video_processor.postprocess_video(frames, output_type="pt")[0])
+        elif output_type == "mp4":
+            frames = torch.Tensor, self.video_processor.postprocess_video(frames, output_type="pil")[0]
+            path = export_to_video(frames, fps=15)
+            with open(path, "rb") as f:
+                frames = f.read()
+
+        return frames