Suggestion to simplify (#2)

- Suggestion to simplify (ea98132386519a165e842c4c4ffedefdb3aa37dc)
- Upload 5 files (a2f01e5b572df78727ff7d3f4198cad5f327ec67)
- Update app.py (d413d63c2fa94c25e73076432da36ed66280e375)


Co-authored-by: Apolinário from multimodal AI art <multimodalart@users.noreply.huggingface.co>

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+0_6as5rHi0sgG4W2Tq.png filter=lfs diff=lfs merge=lfs -text
+1_sTXgMwDUW0pk-1yK4iHYFw.png filter=lfs diff=lfs merge=lfs -text

app.py

@@ -14,39 +14,35 @@ processor = AutoProcessor.from_pretrained(config.base_model_name_or_path)
 model = model.to(device)
 model.eval()
 
-def predict(prompt, image_url, image_pil=None, max_length=64):
-    if image_pil is not None:
-        image = image_pil
-    else:  
-        image = processor.image_processor.fetch_images(image_url)
+#Pre-determined best prompt for this fine-tune
+prompt="Describe the following image:"
+
+#Max generated tokens for your prompt
+max_length=64
+
+def predict(image):
     prompts = [[image, prompt]]
     inputs = processor(prompts[0], return_tensors="pt").to(device)
     generated_ids = model.generate(**inputs, max_length=max_length)
     generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
-    return image, generated_text
-
-
+    generated_text = generated_text.replace(prompt,"")
+    return generated_text
 
 title = "Midjourney-like Image Captioning with IDEFICS"
 description = "Gradio Demo for generating *Midjourney* like captions (describe functionality) with **IDEFICS**"
 
 examples = [
-    ["Describe the following image:", "https://miro.medium.com/v2/resize:fit:0/1*sTXgMwDUW0pk-1yK4iHYFw.png", None, 64],
-    ["Describe the following image:", "https://miro.medium.com/v2/resize:fit:1400/0*6as5rHi0sgG4W2Tq.png", None, 64],
-    ["Describe the following image:", "https://cdn.arstechnica.net/wp-content/uploads/2023/06/zoomout_2-1440x807.jpg", None, 64],
-    ["Describe the following image:", "https://framerusercontent.com/images/inZdRVn7eafZNvaVre2iW1a538.png", None, 64],
-    ["Describe the following image:", "https://hips.hearstapps.com/hmg-prod/images/cute-photos-of-cats-in-grass-1593184777.jpg", None, 64]
-    
+    ["1_sTXgMwDUW0pk-1yK4iHYFw.png"],
+    ["0_6as5rHi0sgG4W2Tq.png"],
+    ["zoomout_2-1440x807.jpg"],
+    ["inZdRVn7eafZNvaVre2iW1a538.webp"],
+    ["cute-photos-of-cats-in-grass-1593184777.jpg"]
 ]
 io = gr.Interface(fn=predict, 
                   inputs=[
-                      gr.Textbox(label="Prompt", value="Describe the following image:", interactive=False),
-                      gr.Textbox(label="image URL", placeholder="Insert the URL of the image to be described"),
-                      gr.Image(label="or upload an image", type="pil"),
-                      gr.Slider(label="Max tokens", value=64, max=128, min=16, step=8)
+                      gr.Image(label="Upload an image", type="pil"),
                   ],
                   outputs=[
-                      gr.Image(type='pil', label="Image"),
                       gr.Textbox(label="IDEFICS Description")
                   ],
                   title=title, description=description, examples=examples,