Update app.py
Browse files
app.py
CHANGED
@@ -2,9 +2,9 @@ import os
|
|
2 |
import streamlit as st
|
3 |
from PIL import Image
|
4 |
import torch
|
5 |
-
from transformers import
|
6 |
|
7 |
-
# Get Hugging Face API key from
|
8 |
HF_TOKEN = os.getenv("HF_KEY")
|
9 |
|
10 |
# Ensure API key is available
|
@@ -12,12 +12,12 @@ if not HF_TOKEN:
|
|
12 |
st.error("β Hugging Face API key not found! Set it as 'HF_KEY' in Spaces secrets.")
|
13 |
st.stop()
|
14 |
|
15 |
-
# Load the
|
16 |
@st.cache_resource
|
17 |
def load_model():
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
return processor, model
|
22 |
|
23 |
processor, model = load_model()
|
@@ -31,33 +31,23 @@ if uploaded_file:
|
|
31 |
image = Image.open(uploaded_file).convert("RGB")
|
32 |
st.image(image, caption="Uploaded Image", use_container_width=True)
|
33 |
|
34 |
-
# User
|
35 |
task = st.selectbox(
|
36 |
"Select a task:",
|
37 |
["Generate a caption", "Answer a question", "Detect objects", "Generate segmentation"]
|
38 |
)
|
39 |
|
40 |
-
# User
|
41 |
prompt = st.text_area("Enter a prompt (e.g., 'Describe the image' or 'What objects are present?')")
|
42 |
|
43 |
if st.button("Run"):
|
44 |
if prompt:
|
45 |
-
inputs = processor(text=prompt, images=image, return_tensors="pt")
|
|
|
46 |
|
47 |
-
with torch.
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
# Handle different outputs
|
53 |
-
if task == "Generate a caption":
|
54 |
-
answer = raw_output
|
55 |
-
elif task == "Answer a question":
|
56 |
-
answer = raw_output
|
57 |
-
elif task == "Detect objects":
|
58 |
-
answer = f"Object bounding boxes: {raw_output}"
|
59 |
-
elif task == "Generate segmentation":
|
60 |
-
answer = f"Segmentation codes: {raw_output}"
|
61 |
|
62 |
st.success(f"β
Result: {answer}")
|
63 |
-
|
|
|
2 |
import streamlit as st
|
3 |
from PIL import Image
|
4 |
import torch
|
5 |
+
from transformers import PaliGemmaProcessor, PaliGemmaForConditionalGeneration
|
6 |
|
7 |
+
# Get Hugging Face API key from environment variables
|
8 |
HF_TOKEN = os.getenv("HF_KEY")
|
9 |
|
10 |
# Ensure API key is available
|
|
|
12 |
st.error("β Hugging Face API key not found! Set it as 'HF_KEY' in Spaces secrets.")
|
13 |
st.stop()
|
14 |
|
15 |
+
# Load the model and processor
|
16 |
@st.cache_resource
|
17 |
def load_model():
|
18 |
+
model_id = "google/paligemma2-3b-mix-224"
|
19 |
+
model = PaliGemmaForConditionalGeneration.from_pretrained(model_id, torch_dtype=torch.bfloat16, device_map="auto").eval()
|
20 |
+
processor = PaliGemmaProcessor.from_pretrained(model_id)
|
21 |
return processor, model
|
22 |
|
23 |
processor, model = load_model()
|
|
|
31 |
image = Image.open(uploaded_file).convert("RGB")
|
32 |
st.image(image, caption="Uploaded Image", use_container_width=True)
|
33 |
|
34 |
+
# User input for task selection
|
35 |
task = st.selectbox(
|
36 |
"Select a task:",
|
37 |
["Generate a caption", "Answer a question", "Detect objects", "Generate segmentation"]
|
38 |
)
|
39 |
|
40 |
+
# User prompt
|
41 |
prompt = st.text_area("Enter a prompt (e.g., 'Describe the image' or 'What objects are present?')")
|
42 |
|
43 |
if st.button("Run"):
|
44 |
if prompt:
|
45 |
+
inputs = processor(text=prompt, images=image, return_tensors="pt").to(torch.bfloat16).to(model.device)
|
46 |
+
input_len = inputs["input_ids"].shape[-1] # Get input length
|
47 |
|
48 |
+
with torch.inference_mode():
|
49 |
+
generation = model.generate(**inputs, max_new_tokens=100, do_sample=False)
|
50 |
+
generation = generation[0][input_len:] # Remove input tokens from output
|
51 |
+
answer = processor.decode(generation, skip_special_tokens=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
|
53 |
st.success(f"β
Result: {answer}")
|
|