Spaces:
Sleeping
Sleeping
rexsimiloluwah
commited on
Commit
•
4ddd43e
1
Parent(s):
bcb1eec
added more applications
Browse files- app.py +21 -3
- {tasks → apps}/__init__.py +0 -0
- apps/__pycache__/__init__.cpython-311.pyc +0 -0
- apps/__pycache__/asr.cpython-311.pyc +0 -0
- {tasks → apps}/asr.py +0 -0
- apps/image_captioning.py +21 -0
- apps/multimodal_visual_qa.py +24 -0
- apps/ner.py +19 -0
- apps/object_detection.py +109 -0
app.py
CHANGED
@@ -1,16 +1,34 @@
|
|
1 |
import gradio as gr
|
2 |
|
3 |
-
from
|
4 |
mic_transcribe_interface,
|
5 |
file_transcribe_interface
|
6 |
)
|
|
|
|
|
|
|
|
|
7 |
|
8 |
app = gr.Blocks()
|
9 |
|
10 |
with app:
|
11 |
gr.TabbedInterface(
|
12 |
-
[
|
13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
)
|
15 |
|
16 |
app.launch(share=True)
|
|
|
1 |
import gradio as gr
|
2 |
|
3 |
+
from apps.asr import (
|
4 |
mic_transcribe_interface,
|
5 |
file_transcribe_interface
|
6 |
)
|
7 |
+
from apps.object_detection import obj_detection_interface
|
8 |
+
from apps.image_captioning import img_captioning_interface
|
9 |
+
from apps.multimodal_visual_qa import multimodal_visual_qa_interface
|
10 |
+
from apps.ner import ner_interface
|
11 |
|
12 |
app = gr.Blocks()
|
13 |
|
14 |
with app:
|
15 |
gr.TabbedInterface(
|
16 |
+
[
|
17 |
+
mic_transcribe_interface,
|
18 |
+
file_transcribe_interface,
|
19 |
+
obj_detection_interface,
|
20 |
+
img_captioning_interface,
|
21 |
+
multimodal_visual_qa_interface,
|
22 |
+
ner_interface
|
23 |
+
],
|
24 |
+
[
|
25 |
+
"Transcribe from Microphone",
|
26 |
+
"Transcribe from Audio File",
|
27 |
+
"Detect Objects from an Image",
|
28 |
+
"Generate a Caption for an Image",
|
29 |
+
"Perform QA on an Image",
|
30 |
+
"Named Entity Recogntion"
|
31 |
+
]
|
32 |
)
|
33 |
|
34 |
app.launch(share=True)
|
{tasks → apps}/__init__.py
RENAMED
File without changes
|
apps/__pycache__/__init__.cpython-311.pyc
ADDED
Binary file (170 Bytes). View file
|
|
apps/__pycache__/asr.cpython-311.pyc
ADDED
Binary file (1.81 kB). View file
|
|
{tasks → apps}/asr.py
RENAMED
File without changes
|
apps/image_captioning.py
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import AutoProcessor
|
3 |
+
from transformers import BlipForConditionalGeneration
|
4 |
+
|
5 |
+
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
|
6 |
+
processor = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
|
7 |
+
|
8 |
+
def caption_image(image):
|
9 |
+
inputs = processor(image, return_tensors="pt")
|
10 |
+
output = model.generate(**inputs)
|
11 |
+
caption = processor.decode(output[0], skip_special_tokens=True)
|
12 |
+
|
13 |
+
return caption
|
14 |
+
|
15 |
+
img_captioning_interface = gr.Interface(
|
16 |
+
fn=caption_image,
|
17 |
+
inputs=gr.Image(label="Input Image", type="pil"),
|
18 |
+
outputs=gr.Textbox(label="Predicted Caption"),
|
19 |
+
title="Image Caption Generator App",
|
20 |
+
description="This app generates a caption for an image."
|
21 |
+
)
|
apps/multimodal_visual_qa.py
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import AutoProcessor
|
3 |
+
from transformers import BlipForQuestionAnswering
|
4 |
+
|
5 |
+
model = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
|
6 |
+
processor = AutoProcessor.from_pretrained("Salesforce/blip-vqa-base")
|
7 |
+
|
8 |
+
def process_image(image, question: str):
|
9 |
+
inputs = processor(image, question, return_tensors="pt")
|
10 |
+
output = model.generate(**inputs)
|
11 |
+
answer = processor.decode(output[0], skip_special_tokens=True)
|
12 |
+
|
13 |
+
return answer
|
14 |
+
|
15 |
+
multimodal_visual_qa_interface = gr.Interface(
|
16 |
+
fn=process_image,
|
17 |
+
inputs=[
|
18 |
+
gr.Image(label="Input Image", type="pil"),
|
19 |
+
gr.Textbox(label="Enter question to prompt the image")
|
20 |
+
],
|
21 |
+
outputs=gr.Textbox(label="Answer"),
|
22 |
+
title="Multimodal Visual QA Application",
|
23 |
+
description="This app can help you ask questions about an image"
|
24 |
+
)
|
apps/ner.py
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import pipeline
|
3 |
+
|
4 |
+
ner_pipeline = pipeline("ner")
|
5 |
+
|
6 |
+
examples = [
|
7 |
+
"Does Chicago have any stores and does Joe live here?",
|
8 |
+
]
|
9 |
+
|
10 |
+
def ner(text):
|
11 |
+
output = ner_pipeline(text)
|
12 |
+
return {"text": text, "entities": output}
|
13 |
+
|
14 |
+
ner_interface = gr.Interface(
|
15 |
+
ner,
|
16 |
+
gr.Textbox(placeholder="Enter sentence"),
|
17 |
+
gr.HighlightedText(),
|
18 |
+
examples=examples
|
19 |
+
)
|
apps/object_detection.py
ADDED
@@ -0,0 +1,109 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import io
|
2 |
+
|
3 |
+
import requests
|
4 |
+
import numpy as np
|
5 |
+
import gradio as gr
|
6 |
+
from PIL import Image
|
7 |
+
import matplotlib.pyplot as plt
|
8 |
+
from transformers import pipeline
|
9 |
+
|
10 |
+
# Load the pipeline
|
11 |
+
obj_detector = pipeline(
|
12 |
+
task="object-detection",
|
13 |
+
model="facebook/detr-resnet-50"
|
14 |
+
)
|
15 |
+
|
16 |
+
# Object detection utilities
|
17 |
+
def load_image_from_url(url: str):
|
18 |
+
return Image.open(requests.get(url, stream=True).raw).convert("RGB")
|
19 |
+
|
20 |
+
def render_results_in_image(img, detection_results):
|
21 |
+
plt.figure(figsize=(16, 10))
|
22 |
+
plt.imshow(img)
|
23 |
+
|
24 |
+
ax = plt.gca()
|
25 |
+
|
26 |
+
for prediction in detection_results:
|
27 |
+
x, y = prediction["box"]["xmin"], prediction["box"]["ymin"]
|
28 |
+
w = prediction["box"]["xmax"] - prediction["box"]["xmin"]
|
29 |
+
h = prediction["box"]["ymax"] - prediction["box"]["ymin"]
|
30 |
+
|
31 |
+
ax.add_patch(
|
32 |
+
plt.Rectangle(
|
33 |
+
(x, y),
|
34 |
+
w,
|
35 |
+
h,
|
36 |
+
fill=False,
|
37 |
+
color="green",
|
38 |
+
linewidth=2
|
39 |
+
)
|
40 |
+
)
|
41 |
+
|
42 |
+
ax.text(
|
43 |
+
x,
|
44 |
+
y,
|
45 |
+
f"{prediction['label']}: {round(prediction['score']*100, 1)}%"
|
46 |
+
)
|
47 |
+
plt.axis("off")
|
48 |
+
|
49 |
+
# save the modified image to a BytesIO object
|
50 |
+
img_buf = io.BytesIO()
|
51 |
+
plt.savefig(img_buf, format="png",
|
52 |
+
bbox_inches="tight",
|
53 |
+
pad_inches=0)
|
54 |
+
img_buf.seek(0)
|
55 |
+
modified_image = Image.open(img_buf)
|
56 |
+
|
57 |
+
# close the plot to prevent it from being displayed
|
58 |
+
plt.close()
|
59 |
+
|
60 |
+
return modified_image
|
61 |
+
|
62 |
+
def summarize_detection_results(detection_results):
|
63 |
+
summary = {}
|
64 |
+
|
65 |
+
for prediction in detection_results:
|
66 |
+
label = prediction["label"]
|
67 |
+
|
68 |
+
if label in summary:
|
69 |
+
summary[label] += 1
|
70 |
+
else:
|
71 |
+
summary[label] = 1
|
72 |
+
|
73 |
+
summary_string = "In this image, there are "
|
74 |
+
|
75 |
+
for i, (label, count) in enumerate(summary.items()):
|
76 |
+
summary_string += f"{str(count)} {label}"
|
77 |
+
|
78 |
+
if count > 1:
|
79 |
+
summary_string += "s"
|
80 |
+
|
81 |
+
summary_string += ", "
|
82 |
+
|
83 |
+
if i == len(summary) - 2:
|
84 |
+
summary_string += "and "
|
85 |
+
|
86 |
+
# remove the trailing comma and space
|
87 |
+
summary_string = summary_string.rstrip(", ") + "."
|
88 |
+
|
89 |
+
return summary_string
|
90 |
+
|
91 |
+
def detect_objects(image):
|
92 |
+
detection_results = obj_detector(image)
|
93 |
+
|
94 |
+
processed_image = render_results_in_image(image, detection_results)
|
95 |
+
|
96 |
+
summary_string = summarize_detection_results(detection_results)
|
97 |
+
|
98 |
+
return processed_image, summary_string
|
99 |
+
|
100 |
+
obj_detection_interface = gr.Interface(
|
101 |
+
fn=detect_objects,
|
102 |
+
inputs=gr.Image(label="Input Image", type="pil"),
|
103 |
+
outputs=[
|
104 |
+
gr.Image(label="Output image with predicted objects", type="pil"),
|
105 |
+
gr.Textbox(label="Object detection summary")
|
106 |
+
],
|
107 |
+
title="Object Detection Application",
|
108 |
+
description="This app detects objects from an image."
|
109 |
+
)
|