back to deepmind model

app.py

@@ -23,8 +23,8 @@ feature_extractor = ViTFeatureExtractor()
 #feature_extractor = ViTFeatureExtractor.from_pretrained("google/vit-base-patch16-224")
 
 #the following gets called by classify_image() 
-feature_extractor = PerceiverFeatureExtractor.from_pretrained("google/vit-base-patch16-224")
-model = PerceiverForImageClassificationConvProcessing.from_pretrained("google/vit-base-patch16-224")
+feature_extractor = PerceiverFeatureExtractor.from_pretrained("deepmind/vision-perceiver-con")
+model = PerceiverForImageClassificationConvProcessing.from_pretrained("deepmind/vision-perceiver-con")
 #google/vit-base-patch16-224, deepmind/vision-perceiver-con
 image_pipe = ImageClassificationPipeline(model=model, feature_extractor=feature_extractor)
 
@@ -59,6 +59,7 @@ def self_caption(image):
   print("Pred dictionary")
   print(pred_dictionary)
   #return(pred_dictionary)
+  preds = ' '.join(preds)
   return preds