Update preprocessing.py

preprocessing.py

@@ -8,6 +8,15 @@ import torch
 from torch.autograd import Variable
 import pdb
 import torch.nn.functional as F
+from sentence_transformers import SentenceTransformer
+
+model = SentenceTransformer('sentence-transformers/all-MiniLM-L12-v2')
+
+def process_caption(text):
+    # Encode text and get model output
+    embeddings = model.encode(text)
+    ouput = np.expand_dims(embeddings, axis=0)
+    return ouput
 
 class Text2ImageDataset(Dataset):