Update app.py
Browse files
app.py
CHANGED
@@ -288,8 +288,18 @@ def chunk_documents(texts: List[str], metadatas: List[dict], dense_model: Optimu
|
|
288 |
documents_and_metadatas = [(chunk.content, chunk.metadata) for sub_chunk in chunks for chunk in sub_chunk]
|
289 |
documents, metadatas_docs = [list(t) for t in zip(*documents_and_metadatas)]
|
290 |
print(f'CHUNKS : {documents}')
|
|
|
|
|
291 |
dense_embeddings = dense_model(documents, 32, convert_to_numpy=True)
|
|
|
|
|
|
|
|
|
|
|
292 |
sparse_embeddings = list(sparse_model.embed(documents, 32, 0))
|
|
|
|
|
|
|
293 |
|
294 |
return documents, metadatas_docs, dense_embeddings, sparse_embeddings
|
295 |
|
|
|
288 |
documents_and_metadatas = [(chunk.content, chunk.metadata) for sub_chunk in chunks for chunk in sub_chunk]
|
289 |
documents, metadatas_docs = [list(t) for t in zip(*documents_and_metadatas)]
|
290 |
print(f'CHUNKS : {documents}')
|
291 |
+
|
292 |
+
start_dense = time.time()
|
293 |
dense_embeddings = dense_model(documents, 32, convert_to_numpy=True)
|
294 |
+
end_dense = time.time()
|
295 |
+
final_dense = end_dense - start_dense
|
296 |
+
print(f'DENSE TIME: {final_dense}')
|
297 |
+
|
298 |
+
start_sparse = time.time()
|
299 |
sparse_embeddings = list(sparse_model.embed(documents, 32, 0))
|
300 |
+
end_sparse = time.time()
|
301 |
+
final_sparse = end_sparse - start_sparse
|
302 |
+
print(f'SPARSE TIME: {final_sparse}')
|
303 |
|
304 |
return documents, metadatas_docs, dense_embeddings, sparse_embeddings
|
305 |
|