sync with upstream codebase

app.py

@@ -44,7 +44,7 @@ def strings2ints(a):
     idx = {v: i for i, v in enumerate(set([*a]))}
     return torch.Tensor([idx[e] for e in a]).to(dtype=torch.int64), {v: k for k, v in idx.items()}
 
-def predict(image):
+def predict(image, n_individuals_to_return=20):
     with torch.inference_mode():
 
         output = ear_detector(image)
@@ -68,14 +68,15 @@ def predict(image):
         similarity = torch.matmul(F.normalize(embedding), F.normalize(features).T)
 
         similarity_sorted_idx = torch.argsort(similarity[0], descending=True).cpu().numpy().reshape(-1)
-        candidates = identities.reshape(-1)[similarity_sorted_idx].tolist()
-        candidates_similarity = similarity[0, similarity_sorted_idx].tolist()
+        candidates, candidates_unique_idx = np.unique(identities.reshape(-1)[similarity_sorted_idx], return_index=True)
 
-    return f"We are about {max(0, candidates_similarity[0]):.0%} confident that this elephant is {'_'.join(candidates[0].split('_')[:-1])}"
+        candidates = candidates[np.argsort(candidates_unique_idx)]
+        candidates_similarity = similarity[0, similarity_sorted_idx].numpy()[np.argsort(candidates_unique_idx)]
 
+    return "Individuals ranked by similarity:\n\n" + "\n\n".join([f"{'_'.join(candidate.split('_')[:-1])}" for candidate, candidate_similarity in zip(candidates[:n_individuals_to_return], candidates_similarity[:n_individuals_to_return])])
 
 gr.Interface(
     fn=predict,
     inputs=gr.Image(type="pil"),
-    outputs=gr.Label(),
-).launch(share=True)
+    outputs=gr.Markdown(),
+).launch(share=True, auth=("user", "mkb6makCBRYGUAd3"))