Add a custom handler for Inference Endpoints

I've added a custom `EndpointHandler` implementation used by Inference Endpoints. I've also added the needed `requirements.txt` - without it, necessary packages won't be installed automatically. I've also did not add any transforms (apart from the `ToTensor()`) which you said are needed .

.gitignore

@@ -0,0 +1,2 @@
+.venv
+.idea

handler.py

@@ -0,0 +1,54 @@
+import torch
+from torchvision import transforms
+from huggingface_hub import hf_hub_download
+import json
+import io
+import base64
+from PIL import Image
+from omegaconf import OmegaConf
+
+from model import Generator
+
+
+class EndpointHandler:
+
+    def __init__(self, path=''):
+        self.transform = transforms.Compose([
+            transforms.ToTensor()
+        ])
+
+        repo_id = "Kiwinicki/sat2map-generator"
+        generator_path = hf_hub_download(repo_id=repo_id, filename="generator.pth")
+        config_path = hf_hub_download(repo_id=repo_id, filename="config.json")
+        model_path = hf_hub_download(repo_id=repo_id, filename="model.py")
+
+        with open(config_path, "r") as f:
+            config_dict = json.load(f)
+        cfg = OmegaConf.create(config_dict)
+
+        self.generator = Generator(cfg)
+        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+        self.generator.load_state_dict(torch.load(generator_path, map_location=self.device))
+        self.generator.eval()
+
+
+    def __call__(self, data: dict[str, any]) -> dict[str, str]:
+        base64_image = data.get('inputs')
+        input_tensor = self._decode_base64_image(base64_image)
+        # print('Input tensor shape: ' + str(input_tensor.shape))
+        output_tensor = self.generator(input_tensor.to(self.device))
+        output_tensor = output_tensor.squeeze(0)
+        output_image = transforms.ToPILImage()(output_tensor)
+        output_image = output_image.convert('RGB')
+        output_buffer = io.BytesIO()
+        output_image.save(output_buffer, format="png")
+        base64_output = base64.b64encode(output_buffer.getvalue()).decode('utf-8')
+        return {"output": base64_output}
+
+
+    def _decode_base64_image(self, base64_image: str) -> torch.Tensor:
+        image_decoded = base64.b64decode(base64_image)
+        image = Image.open(io.BytesIO(image_decoded)).convert('RGB')
+        image_tensor: torch.Tensor = self.transform(image)
+        image_tensor = image_tensor.unsqueeze(0)
+        return image_tensor