Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -12,7 +12,8 @@ title = """ # 🙋🏻♂️Welcome to Tonic's🦅Falcon Vision👁️Langua
|
|
| 12 |
description = """
|
| 13 |
Falcon2-11B-vlm is an 11B parameters causal decoder-only model built by TII and trained on over 5,000B tokens of RefinedWeb enhanced with curated corpora. To bring vision capabilities, , we integrate the pretrained CLIP ViT-L/14 vision encoder with our Falcon2-11B chat-finetuned model and train with image-text data. For enhancing the VLM's perception of fine-grained details w.r.t small objects in images, we employ a dynamic encoding mechanism at high-resolution for image inputs.
|
| 14 |
|
| 15 |
-
|
|
|
|
| 16 |
"""
|
| 17 |
|
| 18 |
processor = LlavaNextProcessor.from_pretrained("tiiuae/falcon-11B-vlm", tokenizer_class='PreTrainedTokenizerFast')
|
|
@@ -31,42 +32,51 @@ def generate_paragraph(image_url):
|
|
| 31 |
|
| 32 |
return generated_captions
|
| 33 |
|
| 34 |
-
# Function to set the URL and generate the paragraph
|
| 35 |
def set_and_generate(url):
|
| 36 |
generated_paragraph = generate_paragraph(url)
|
| 37 |
return url, generated_paragraph
|
| 38 |
|
| 39 |
-
|
| 40 |
with gr.Blocks() as demo:
|
|
|
|
| 41 |
gr.Markdown(title)
|
| 42 |
gr.Markdown(description)
|
| 43 |
|
|
|
|
| 44 |
with gr.Row():
|
|
|
|
| 45 |
with gr.Column():
|
|
|
|
| 46 |
image_url_input = gr.Textbox(label="Image URL")
|
| 47 |
generate_button = gr.Button("Generate Paragraph")
|
| 48 |
|
| 49 |
-
example_1 = gr.
|
| 50 |
-
example_2 = gr.
|
| 51 |
-
example_3 = gr.
|
| 52 |
-
|
|
|
|
| 53 |
with gr.Column():
|
| 54 |
generated_paragraph_output = gr.Textbox(label="Generated Paragraph")
|
| 55 |
|
| 56 |
generate_button.click(generate_paragraph, inputs=image_url_input, outputs=generated_paragraph_output)
|
| 57 |
|
|
|
|
|
|
|
| 58 |
example_1.click(
|
| 59 |
lambda: set_and_generate("https://www.animalspot.net/wp-content/uploads/2020/01/Types-of-Falcons.jpg"),
|
| 60 |
outputs=[image_url_input, generated_paragraph_output]
|
| 61 |
)
|
|
|
|
| 62 |
example_2.click(
|
| 63 |
lambda: set_and_generate("https://www.leaders-mena.com/leaders/uploads/2023/01/The-Traditional-Camel-Racing-In-Saudi-Arabia-Unique-Sport-Activity-1024x576.jpg"),
|
| 64 |
outputs=[image_url_input, generated_paragraph_output]
|
| 65 |
)
|
|
|
|
| 66 |
example_3.click(
|
| 67 |
lambda: set_and_generate("http://embed.robertharding.com/embed/1161-4342.jpg"),
|
| 68 |
outputs=[image_url_input, generated_paragraph_output]
|
| 69 |
)
|
| 70 |
|
|
|
|
| 71 |
# Launch the Gradio interface
|
| 72 |
demo.launch()
|
|
|
|
| 12 |
description = """
|
| 13 |
Falcon2-11B-vlm is an 11B parameters causal decoder-only model built by TII and trained on over 5,000B tokens of RefinedWeb enhanced with curated corpora. To bring vision capabilities, , we integrate the pretrained CLIP ViT-L/14 vision encoder with our Falcon2-11B chat-finetuned model and train with image-text data. For enhancing the VLM's perception of fine-grained details w.r.t small objects in images, we employ a dynamic encoding mechanism at high-resolution for image inputs.
|
| 14 |
|
| 15 |
+
### Join us :
|
| 16 |
+
🌟TeamTonic🌟 is always making cool demos! Join our active builder's 🛠️community 👻 [](https://discord.gg/GWpVpekp) On 🤗Huggingface:[MultiTransformer](https://huggingface.co/MultiTransformer) Math 🔍 [introspector](https://huggingface.co/introspector) On 🌐Github: [Tonic-AI](https://github.com/tonic-ai) & contribute to🌟 [MultiTonic](https://github.com/multitonic/)🤗Big thanks to Yuvi Sharma and all the folks at huggingface for the community grant 🤗
|
| 17 |
"""
|
| 18 |
|
| 19 |
processor = LlavaNextProcessor.from_pretrained("tiiuae/falcon-11B-vlm", tokenizer_class='PreTrainedTokenizerFast')
|
|
|
|
| 32 |
|
| 33 |
return generated_captions
|
| 34 |
|
|
|
|
| 35 |
def set_and_generate(url):
|
| 36 |
generated_paragraph = generate_paragraph(url)
|
| 37 |
return url, generated_paragraph
|
| 38 |
|
| 39 |
+
|
| 40 |
with gr.Blocks() as demo:
|
| 41 |
+
|
| 42 |
gr.Markdown(title)
|
| 43 |
gr.Markdown(description)
|
| 44 |
|
| 45 |
+
|
| 46 |
with gr.Row():
|
| 47 |
+
|
| 48 |
with gr.Column():
|
| 49 |
+
|
| 50 |
image_url_input = gr.Textbox(label="Image URL")
|
| 51 |
generate_button = gr.Button("Generate Paragraph")
|
| 52 |
|
| 53 |
+
example_1 = gr.Image(value="https://www.animalspot.net/wp-content/uploads/2020/01/Types-of-Falcons.jpg", label="Types of Falcons", interactive=True, shape=(150, 150))
|
| 54 |
+
example_2 = gr.Image(value="https://www.leaders-mena.com/leaders/uploads/2023/01/The-Traditional-Camel-Racing-In-Saudi-Arabia-Unique-Sport-Activity-1024x576.jpg", label="Camel Racing - Saudi Arabia", interactive=True, shape=(150, 150))
|
| 55 |
+
example_3 = gr.Image(value="http://embed.robertharding.com/embed/1161-4342.jpg", label="Urban Street Scene - India", interactive=True, shape=(150, 150))
|
| 56 |
+
|
| 57 |
+
|
| 58 |
with gr.Column():
|
| 59 |
generated_paragraph_output = gr.Textbox(label="Generated Paragraph")
|
| 60 |
|
| 61 |
generate_button.click(generate_paragraph, inputs=image_url_input, outputs=generated_paragraph_output)
|
| 62 |
|
| 63 |
+
|
| 64 |
+
|
| 65 |
example_1.click(
|
| 66 |
lambda: set_and_generate("https://www.animalspot.net/wp-content/uploads/2020/01/Types-of-Falcons.jpg"),
|
| 67 |
outputs=[image_url_input, generated_paragraph_output]
|
| 68 |
)
|
| 69 |
+
|
| 70 |
example_2.click(
|
| 71 |
lambda: set_and_generate("https://www.leaders-mena.com/leaders/uploads/2023/01/The-Traditional-Camel-Racing-In-Saudi-Arabia-Unique-Sport-Activity-1024x576.jpg"),
|
| 72 |
outputs=[image_url_input, generated_paragraph_output]
|
| 73 |
)
|
| 74 |
+
|
| 75 |
example_3.click(
|
| 76 |
lambda: set_and_generate("http://embed.robertharding.com/embed/1161-4342.jpg"),
|
| 77 |
outputs=[image_url_input, generated_paragraph_output]
|
| 78 |
)
|
| 79 |
|
| 80 |
+
|
| 81 |
# Launch the Gradio interface
|
| 82 |
demo.launch()
|