Update app.py
Browse files
app.py
CHANGED
|
@@ -1,41 +1,65 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
-
|
|
|
|
| 3 |
from PIL import Image
|
| 4 |
|
| 5 |
-
# ---
|
|
|
|
|
|
|
| 6 |
MODEL_NAME = "microsoft/Florence-2-large"
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
trust_remote_code=True
|
| 12 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
-
# ---
|
| 15 |
def generate_alt_text(files, max_new_tokens):
|
| 16 |
-
"""
|
| 17 |
-
files: list of uploaded image files
|
| 18 |
-
max_new_tokens: how many new tokens to generate
|
| 19 |
-
"""
|
| 20 |
captions = []
|
| 21 |
for f in files or []:
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
)
|
| 31 |
-
|
| 32 |
-
captions.append(
|
| 33 |
return "\n\n".join(captions)
|
| 34 |
|
| 35 |
# --- Gradio UI ---
|
| 36 |
-
with gr.Blocks(title="Florence
|
| 37 |
-
gr.Markdown("# Florence
|
| 38 |
-
gr.Markdown("Using
|
| 39 |
|
| 40 |
with gr.Row():
|
| 41 |
file_input = gr.File(
|
|
@@ -47,23 +71,23 @@ with gr.Blocks(title="Florence‑2‑large Alt‑Text Generator") as demo:
|
|
| 47 |
minimum=16,
|
| 48 |
maximum=512,
|
| 49 |
step=16,
|
| 50 |
-
value=
|
| 51 |
label="Max New Tokens"
|
| 52 |
)
|
| 53 |
generate_btn = gr.Button("Generate Alt Text")
|
| 54 |
-
clear_btn
|
| 55 |
|
| 56 |
gallery = gr.Gallery(label="Image Preview", columns=4)
|
| 57 |
-
output
|
| 58 |
|
| 59 |
-
# Preview
|
| 60 |
file_input.change(
|
| 61 |
fn=lambda files: [Image.open(f.name).convert("RGB") for f in (files or [])],
|
| 62 |
inputs=file_input,
|
| 63 |
-
outputs=gallery
|
| 64 |
)
|
| 65 |
|
| 66 |
-
# Generate captions
|
| 67 |
generate_btn.click(
|
| 68 |
fn=generate_alt_text,
|
| 69 |
inputs=[file_input, max_new_tokens],
|
|
@@ -71,9 +95,9 @@ with gr.Blocks(title="Florence‑2‑large Alt‑Text Generator") as demo:
|
|
| 71 |
queue=True
|
| 72 |
)
|
| 73 |
|
| 74 |
-
# Clear
|
| 75 |
clear_btn.click(
|
| 76 |
-
fn=lambda: ([],
|
| 77 |
outputs=[file_input, max_new_tokens, gallery, output]
|
| 78 |
)
|
| 79 |
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
+
import torch
|
| 3 |
+
from transformers import AutoProcessor, AutoModelForCausalLM
|
| 4 |
from PIL import Image
|
| 5 |
|
| 6 |
+
# --- Configuration ---
|
| 7 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 8 |
+
torch_dtype = torch.float16 if device == "cuda" else torch.float32
|
| 9 |
MODEL_NAME = "microsoft/Florence-2-large"
|
| 10 |
+
|
| 11 |
+
# Load model and processor with remote code trust
|
| 12 |
+
processor = AutoProcessor.from_pretrained(
|
| 13 |
+
MODEL_NAME,
|
| 14 |
+
trust_remote_code=True
|
| 15 |
)
|
| 16 |
+
model = AutoModelForCausalLM.from_pretrained(
|
| 17 |
+
MODEL_NAME,
|
| 18 |
+
trust_remote_code=True,
|
| 19 |
+
torch_dtype=torch_dtype
|
| 20 |
+
).to(device)
|
| 21 |
|
| 22 |
+
# --- Alt Text Generation ---
|
| 23 |
def generate_alt_text(files, max_new_tokens):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
captions = []
|
| 25 |
for f in files or []:
|
| 26 |
+
# Load image
|
| 27 |
+
image = Image.open(f.name).convert("RGB")
|
| 28 |
+
# Use caption prompt
|
| 29 |
+
prompt = "<CAPTION>"
|
| 30 |
+
# Preprocess inputs
|
| 31 |
+
inputs = processor(
|
| 32 |
+
text=prompt,
|
| 33 |
+
images=image,
|
| 34 |
+
return_tensors="pt"
|
| 35 |
+
).to(device)
|
| 36 |
+
# Generate tokens
|
| 37 |
+
generated_ids = model.generate(
|
| 38 |
+
input_ids=inputs["input_ids"],
|
| 39 |
+
pixel_values=inputs["pixel_values"],
|
| 40 |
+
max_new_tokens=int(max_new_tokens),
|
| 41 |
+
num_beams=3,
|
| 42 |
+
do_sample=False
|
| 43 |
+
)
|
| 44 |
+
# Decode text
|
| 45 |
+
generated_text = processor.batch_decode(
|
| 46 |
+
generated_ids,
|
| 47 |
+
skip_special_tokens=True
|
| 48 |
+
)[0]
|
| 49 |
+
# Post-process for caption task
|
| 50 |
+
result = processor.post_process_generation(
|
| 51 |
+
generated_text,
|
| 52 |
+
task=prompt,
|
| 53 |
+
image_size=(image.width, image.height)
|
| 54 |
)
|
| 55 |
+
caption = result.get(prompt, "").strip().capitalize()
|
| 56 |
+
captions.append(caption)
|
| 57 |
return "\n\n".join(captions)
|
| 58 |
|
| 59 |
# --- Gradio UI ---
|
| 60 |
+
with gr.Blocks(title="Florence-2-large Alt Text Generator") as demo:
|
| 61 |
+
gr.Markdown("# Florence-2-large Alt Text Generator")
|
| 62 |
+
gr.Markdown("Using microsoft/Florence-2-large with manual model & processor loading.")
|
| 63 |
|
| 64 |
with gr.Row():
|
| 65 |
file_input = gr.File(
|
|
|
|
| 71 |
minimum=16,
|
| 72 |
maximum=512,
|
| 73 |
step=16,
|
| 74 |
+
value=128,
|
| 75 |
label="Max New Tokens"
|
| 76 |
)
|
| 77 |
generate_btn = gr.Button("Generate Alt Text")
|
| 78 |
+
clear_btn = gr.Button("Clear All")
|
| 79 |
|
| 80 |
gallery = gr.Gallery(label="Image Preview", columns=4)
|
| 81 |
+
output = gr.Textbox(label="Generated Alt Texts", lines=10, show_copy_button=True)
|
| 82 |
|
| 83 |
+
# Preview images
|
| 84 |
file_input.change(
|
| 85 |
fn=lambda files: [Image.open(f.name).convert("RGB") for f in (files or [])],
|
| 86 |
inputs=file_input,
|
| 87 |
+
outputs=gallery
|
| 88 |
)
|
| 89 |
|
| 90 |
+
# Generate captions on click
|
| 91 |
generate_btn.click(
|
| 92 |
fn=generate_alt_text,
|
| 93 |
inputs=[file_input, max_new_tokens],
|
|
|
|
| 95 |
queue=True
|
| 96 |
)
|
| 97 |
|
| 98 |
+
# Clear all fields
|
| 99 |
clear_btn.click(
|
| 100 |
+
fn=lambda: ([], 128, [], ""),
|
| 101 |
outputs=[file_input, max_new_tokens, gallery, output]
|
| 102 |
)
|
| 103 |
|