Spaces:
Sleeping
Sleeping
Update inference.py
Browse files- inference.py +31 -15
inference.py
CHANGED
@@ -1,9 +1,7 @@
|
|
1 |
import os
|
2 |
import torch
|
3 |
-
import config
|
4 |
import spacy
|
5 |
-
|
6 |
-
|
7 |
from utils import (
|
8 |
load_dataset,
|
9 |
get_model_instance,
|
@@ -15,6 +13,9 @@ from PIL import Image
|
|
15 |
import torchvision.transforms as transforms
|
16 |
import streamlit as st
|
17 |
|
|
|
|
|
|
|
18 |
# Define device
|
19 |
DEVICE = 'cpu'
|
20 |
|
@@ -26,6 +27,7 @@ TRANSFORMS = transforms.Compose([
|
|
26 |
])
|
27 |
|
28 |
|
|
|
29 |
def load_model():
|
30 |
"""
|
31 |
Loads the model with the vocabulary and checkpoint.
|
@@ -39,9 +41,13 @@ def load_model():
|
|
39 |
|
40 |
if can_load_checkpoint():
|
41 |
st.write("Loading checkpoint...")
|
42 |
-
|
|
|
|
|
|
|
|
|
43 |
else:
|
44 |
-
st.
|
45 |
|
46 |
model.eval() # Set the model to evaluation mode
|
47 |
st.write("Model is ready for inference.")
|
@@ -53,9 +59,13 @@ def preprocess_image(image_path):
|
|
53 |
Preprocess the input image for the model.
|
54 |
"""
|
55 |
st.write(f"Preprocessing image: {image_path}")
|
56 |
-
|
57 |
-
|
58 |
-
|
|
|
|
|
|
|
|
|
59 |
|
60 |
|
61 |
def generate_report(model, image):
|
@@ -63,12 +73,15 @@ def generate_report(model, image):
|
|
63 |
Generates a report for a given image using the model.
|
64 |
"""
|
65 |
st.write("Generating report...")
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
|
|
|
|
|
|
72 |
|
73 |
|
74 |
# Streamlit App
|
@@ -79,6 +92,9 @@ st.write("Upload an X-ray image to generate a report.")
|
|
79 |
uploaded_file = st.file_uploader("Choose an image file", type=["png", "jpg", "jpeg"])
|
80 |
|
81 |
if uploaded_file is not None:
|
|
|
|
|
|
|
82 |
# Save uploaded file to disk
|
83 |
image_path = os.path.join("temp", uploaded_file.name)
|
84 |
with open(image_path, "wb") as f:
|
@@ -94,4 +110,4 @@ if uploaded_file is not None:
|
|
94 |
# Display the image and the report
|
95 |
st.image(image_path, caption="Uploaded Image", use_column_width=True)
|
96 |
st.write("Generated Report:")
|
97 |
-
st.write(report)
|
|
|
1 |
import os
|
2 |
import torch
|
|
|
3 |
import spacy
|
4 |
+
import config
|
|
|
5 |
from utils import (
|
6 |
load_dataset,
|
7 |
get_model_instance,
|
|
|
13 |
import torchvision.transforms as transforms
|
14 |
import streamlit as st
|
15 |
|
16 |
+
# Download Spacy model (only once during runtime)
|
17 |
+
spacy.cli.download("en_core_web_sm")
|
18 |
+
|
19 |
# Define device
|
20 |
DEVICE = 'cpu'
|
21 |
|
|
|
27 |
])
|
28 |
|
29 |
|
30 |
+
@st.cache_resource
|
31 |
def load_model():
|
32 |
"""
|
33 |
Loads the model with the vocabulary and checkpoint.
|
|
|
41 |
|
42 |
if can_load_checkpoint():
|
43 |
st.write("Loading checkpoint...")
|
44 |
+
try:
|
45 |
+
load_checkpoint(model)
|
46 |
+
except RuntimeError as e:
|
47 |
+
st.error(f"Error loading checkpoint: {e}")
|
48 |
+
st.stop()
|
49 |
else:
|
50 |
+
st.warning("No checkpoint found, starting with untrained model.")
|
51 |
|
52 |
model.eval() # Set the model to evaluation mode
|
53 |
st.write("Model is ready for inference.")
|
|
|
59 |
Preprocess the input image for the model.
|
60 |
"""
|
61 |
st.write(f"Preprocessing image: {image_path}")
|
62 |
+
try:
|
63 |
+
image = Image.open(image_path).convert("RGB")
|
64 |
+
image = TRANSFORMS(image).unsqueeze(0)
|
65 |
+
return image.to(DEVICE)
|
66 |
+
except Exception as e:
|
67 |
+
st.error(f"Error preprocessing image: {e}")
|
68 |
+
st.stop()
|
69 |
|
70 |
|
71 |
def generate_report(model, image):
|
|
|
73 |
Generates a report for a given image using the model.
|
74 |
"""
|
75 |
st.write("Generating report...")
|
76 |
+
try:
|
77 |
+
with torch.no_grad():
|
78 |
+
output = model.generate_caption(image, max_length=25)
|
79 |
+
report = " ".join(output)
|
80 |
+
st.write(f"Generated report: {report}")
|
81 |
+
return report
|
82 |
+
except Exception as e:
|
83 |
+
st.error(f"Error generating report: {e}")
|
84 |
+
st.stop()
|
85 |
|
86 |
|
87 |
# Streamlit App
|
|
|
92 |
uploaded_file = st.file_uploader("Choose an image file", type=["png", "jpg", "jpeg"])
|
93 |
|
94 |
if uploaded_file is not None:
|
95 |
+
# Ensure 'temp' directory exists
|
96 |
+
os.makedirs("temp", exist_ok=True)
|
97 |
+
|
98 |
# Save uploaded file to disk
|
99 |
image_path = os.path.join("temp", uploaded_file.name)
|
100 |
with open(image_path, "wb") as f:
|
|
|
110 |
# Display the image and the report
|
111 |
st.image(image_path, caption="Uploaded Image", use_column_width=True)
|
112 |
st.write("Generated Report:")
|
113 |
+
st.write(report)
|