Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -2,7 +2,6 @@ import torch
|
|
2 |
import fastapi
|
3 |
import numpy as np
|
4 |
from PIL import Image
|
5 |
-
from typing import Any, Type
|
6 |
|
7 |
class TorchTensor(torch.Tensor):
|
8 |
pass
|
@@ -10,23 +9,21 @@ class TorchTensor(torch.Tensor):
|
|
10 |
class Prediction:
|
11 |
prediction: TorchTensor
|
12 |
|
13 |
-
app = fastapi.FastAPI()
|
14 |
-
|
15 |
-
model = torch.load("model67.bin", map_location='cpu')
|
16 |
|
|
|
17 |
# Define a function to preprocess the input image
|
18 |
-
def preprocess_input(input:
|
19 |
-
image = Image.open(
|
20 |
image = image.resize((224, 224))
|
21 |
input = np.array(image)
|
22 |
input = torch.from_numpy(input).float()
|
23 |
-
input = input.permute(2, 0, 1)
|
24 |
input = input.unsqueeze(0)
|
25 |
return input
|
26 |
|
27 |
# Define an endpoint to make predictions
|
28 |
@app.post("/predict")
|
29 |
-
async def predict_endpoint(input:
|
30 |
"""Make a prediction on an image uploaded by the user."""
|
31 |
|
32 |
# Preprocess the input image
|
@@ -35,12 +32,8 @@ async def predict_endpoint(input: Any):
|
|
35 |
# Make a prediction
|
36 |
prediction = model(input)
|
37 |
|
38 |
-
|
39 |
predicted_class = prediction.argmax(1).item()
|
40 |
|
41 |
# Return the predicted class in JSON format
|
42 |
-
return {"prediction": predicted_class}
|
43 |
-
|
44 |
-
if __name__ == "__main__":
|
45 |
-
import uvicorn
|
46 |
-
uvicorn.run(app, host="0.0.0.0", port=8000)
|
|
|
2 |
import fastapi
|
3 |
import numpy as np
|
4 |
from PIL import Image
|
|
|
5 |
|
6 |
class TorchTensor(torch.Tensor):
|
7 |
pass
|
|
|
9 |
class Prediction:
|
10 |
prediction: TorchTensor
|
11 |
|
12 |
+
app = fastapi.FastAPI(docs_url="/")
|
|
|
|
|
13 |
|
14 |
+
model = torch.load("best_model.pth", map_location='cpu')
|
15 |
# Define a function to preprocess the input image
|
16 |
+
def preprocess_input(input: fastapi.UploadFile):
|
17 |
+
image = Image.open(input.file)
|
18 |
image = image.resize((224, 224))
|
19 |
input = np.array(image)
|
20 |
input = torch.from_numpy(input).float()
|
|
|
21 |
input = input.unsqueeze(0)
|
22 |
return input
|
23 |
|
24 |
# Define an endpoint to make predictions
|
25 |
@app.post("/predict")
|
26 |
+
async def predict_endpoint(input:fastapi.UploadFile):
|
27 |
"""Make a prediction on an image uploaded by the user."""
|
28 |
|
29 |
# Preprocess the input image
|
|
|
32 |
# Make a prediction
|
33 |
prediction = model(input)
|
34 |
|
35 |
+
|
36 |
predicted_class = prediction.argmax(1).item()
|
37 |
|
38 |
# Return the predicted class in JSON format
|
39 |
+
return {"prediction": predicted_class}
|
|
|
|
|
|
|
|