Find answers from the community

s
sodeep
Offline, last seen last week
Joined September 25, 2024
Is it possible to use a local TEI server for reranking?
26 comments
L
s
Hmm. When using Qdrant (gRPC) with LlamaIndex, I see this warning with each script invocation.
Plain Text
WARNING:root:Unable to close grpc_channel. Connection was interrupted on the server side
2 comments
s
L
This code takes over 90 seconds to ingest a small 32KB markdown file into FalkorDB.
Plain Text
reader = SimpleDirectoryReader(input_dir='data', recursive=False)
documents = reader.load_data()

Settings.llm = OpenAI(model='gpt-4o-mini', api_key=os.getenv('OPENAI_API_KEY'), temperature=0, max_tokens=100)
Settings.chunk_size = 512

graph_store = FalkorDBGraphStore('redis://localhost:6379', decode_responses=True)
storage_context = StorageContext.from_defaults(graph_store=graph_store)

knowledge_graph_index = KnowledgeGraphIndex.from_documents(
    documents,
    max_triplets_per_chunk=5,
    storage_context=storage_context,
)
Does LlamaIndex offer anything similar to the "Contextual Embeddings" found in this Anthropic post? https://www.anthropic.com/news/contextual-retrieval
17 comments
s
L
R
@Logan M I see an entry in the docs about using the Jina AI embedder via SageMaker, but I don't see anything related to the Jina reranker via SageMaker. Is the latter not possible just yet?
2 comments
s
L