Upload app.py

app.py

@@ -0,0 +1,48 @@
+import torch
+from transformers import pipeline, AutoTokenizer
+import gradio as gr
+
+# Load tokenizer
+tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
+tokenizer.clean_up_tokenization_spaces = False  # Explicitly set the parameter if needed
+
+# Load CLIP model for zero-shot classification
+clip_checkpoint = "openai/clip-vit-base-patch16"
+clip_detector = pipeline(model=clip_checkpoint, task="zero-shot-image-classification")
+
+# Postprocess the output from CLIP
+def postprocess(output):
+    return {out["label"]: float(out["score"]) for out in output}
+
+# Inference function for CLIP
+def infer(image, candidate_labels):
+    candidate_labels = [label.lstrip(" ") for label in candidate_labels.split(",")]
+    clip_out = clip_detector(image, candidate_labels=candidate_labels)
+    return postprocess(clip_out)
+
+# Gradio interface
+with gr.Blocks() as app:
+    gr.Markdown("# Custom Classification")
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(type="pil")
+            text_input = gr.Textbox(label="Input a list of labels")
+            run_button = gr.Button("Run")
+
+        with gr.Column():
+            clip_output = gr.Label(label="Output", num_top_classes=3)
+    
+    examples = [["image_8.webp", "girl, boy, lgbtq"]]
+    gr.Examples(
+        examples=examples, 
+        inputs=[image_input, text_input],
+        outputs=[clip_output],
+        fn=infer,
+        cache_examples=True
+    )
+    
+    run_button.click(fn=infer,
+                     inputs=[image_input, text_input],
+                     outputs=[clip_output])
+
+app.launch()