# Llamma index Saving and Loading **URL:** \ **Category:** Intermediate\ **Created:** [December 30, 2023, 6:07pm UTC](https://discuss.huggingface.co/t/llamma-index-saving-and-loading/67491 "2023-12-30T18:07:24Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![AgelZin](https://avatars.discourse-cdn.com/v4/letter/a/6bbea6/32.png) [@AgelZin](https://discuss.huggingface.co/u/AgelZin)\ **Post date:** [December 30, 2023, 6:07pm UTC](https://discuss.huggingface.co/t/llamma-index-saving-and-loading/67491/1 "2023-12-30T18:07:24Z")
I create a llamma2 model and could query with it with no problems and then i saved it like this. > index.storage\_context.persist(persist\_dir=“\”) After loading i encountered a weird error # rebuild storage context storage\_context = StorageContext.from\_defaults(persist\_dir=“\”) # load index index = load\_index\_from\_storage(storage\_context) query\_engine = index.as\_query\_engine() response = query\_engine.query(“random message”) * * * ValueError Traceback (most recent call last) Cell In[12], line 2 1 query\_engine = index.as\_query\_engine() ----\> 2 response = query\_engine.query(“a”) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/core/base\_query\_engine.py:30, in BaseQueryEngine.query(self, str\_or\_query\_bundle) 28 if isinstance(str\_or\_query\_bundle, str): 29 str\_or\_query\_bundle = QueryBundle(str\_or\_query\_bundle) —\> 30 return self.\_query(str\_or\_query\_bundle) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/query\_engine/retriever\_query\_engine.py:170, in RetrieverQueryEngine.\_query(self, query\_bundle) 166 “”“Answer a query.”“” 167 with self.callback\_manager.event( 168 CBEventType.QUERY, payload={EventPayload.QUERY\_STR: query\_bundle.query\_str} 169 ) as query\_event: → 170 nodes = self.retrieve(query\_bundle) 171 response = self.\_response\_synthesizer.synthesize( 172 query=query\_bundle, 173 nodes=nodes, 174 ) 176 query\_event.on\_end(payload={EventPayload.RESPONSE: response}) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/query\_engine/retriever\_query\_engine.py:126, in RetrieverQueryEngine.retrieve(self, query\_bundle) 125 def retrieve(self, query\_bundle: QueryBundle) → List[NodeWithScore]: → 126 nodes = self.\_retriever.retrieve(query\_bundle) 127 return self.\_apply\_node\_postprocessors(nodes, query\_bundle=query\_bundle) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/core/base\_retriever.py:54, in BaseRetriever.retrieve(self, str\_or\_query\_bundle) 49 with self.callback\_manager.as\_trace(“query”): 50 with self.callback\_manager.event( 51 CBEventType.RETRIEVE, 52 payload={EventPayload.QUERY\_STR: query\_bundle.query\_str}, 53 ) as retrieve\_event: —\> 54 nodes = self.\_retrieve(query\_bundle) 55 retrieve\_event.on\_end( 56 payload={EventPayload.NODES: nodes}, 57 ) 58 return nodes File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/indices/vector\_store/retrievers/retriever.py:88, in VectorIndexRetriever.\_retrieve(self, query\_bundle) 82 if query\_bundle.embedding is None and len(query\_bundle.embedding\_strs) \> 0: 83 query\_bundle.embedding = ( 84 self.\_service\_context.embed\_model.get\_agg\_embedding\_from\_queries( 85 query\_bundle.embedding\_strs 86 ) 87 ) —\> 88 return self.\_get\_nodes\_with\_embeddings(query\_bundle) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/indices/vector\_store/retrievers/retriever.py:164, in VectorIndexRetriever.\_get\_nodes\_with\_embeddings(self, query\_bundle\_with\_embeddings) 160 def \_get\_nodes\_with\_embeddings( 161 self, query\_bundle\_with\_embeddings: QueryBundle 162 ) → List[NodeWithScore]: 163 query = self.\_build\_vector\_store\_query(query\_bundle\_with\_embeddings) → 164 query\_result = self.\_vector\_store.query(query, \*\*self.\_kwargs) 165 return self.\_build\_node\_list\_from\_query\_result(query\_result) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/vector\_stores/simple.py:274, in SimpleVectorStore.query(self, query, \*\*kwargs) 266 top\_similarities, top\_ids = get\_top\_k\_mmr\_embeddings( 267 query\_embedding, 268 embeddings, (…) 271 mmr\_threshold=mmr\_threshold, 272 ) 273 elif query.mode == VectorStoreQueryMode.DEFAULT: → 274 top\_similarities, top\_ids = get\_top\_k\_embeddings( 275 query\_embedding, 276 embeddings, 277 similarity\_top\_k=query.similarity\_top\_k, 278 embedding\_ids=node\_ids, 279 ) 280 else: 281 raise ValueError(f"Invalid query mode: {query.mode}") File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/indices/query/embedding\_utils.py:31, in get\_top\_k\_embeddings(query\_embedding, embeddings, similarity\_fn, similarity\_top\_k, embedding\_ids, similarity\_cutoff) 29 similarity\_heap: List[Tuple[float, Any]] = 30 for i, emb in enumerate(embeddings\_np): —\> 31 similarity = similarity\_fn(query\_embedding\_np, emb) 32 if similarity\_cutoff is None or similarity \> similarity\_cutoff: 33 heapq.heappush(similarity\_heap, (similarity, embedding\_ids[i])) File /srv/scratch3/zinonos/zin/lib/python3.10/site-packages/llama\_index/embeddings/base.py:47, in similarity(embedding1, embedding2, mode) 45 return np.dot(embedding1, embedding2) 46 else: —\> 47 product = np.dot(embedding1, embedding2) 48 norm = np.linalg.norm(embedding1) \* np.linalg.norm(embedding2) 49 return product / norm ValueError: shapes (1536,) and (768,) not aligned: 1536 (dim 0) != 768 (dim 0) ---
**Author:** ![AgelZin](https://avatars.discourse-cdn.com/v4/letter/a/6bbea6/32.png) [@AgelZin](https://discuss.huggingface.co/u/AgelZin)\ **Post date:** [January 2, 2024, 12:39pm UTC](https://discuss.huggingface.co/t/llamma-index-saving-and-loading/67491/2 "2024-01-02T12:39:27Z")
Solved. Have to adjust code like this else it loads default embedings > index = load\_index\_from\_storage(storage\_context, service\_context=service\_context)