ai-agents-embeddings-integrati-llamacppembeddings-embed-documents-raises-typeerro-5b3ee2d9

THEN Ensure the LlamaCppEmbeddings instance is initialized with the `embedding=True` argument, or upgrade llama-cpp-python to a version that correctly returns single-sequence embeddings. For llama-cpp-python v0.2+, set `embedding=True` in the model constructor. Alternatively, flatten the nested list output by modifying line 114 to `return [list(map(float, sublist)) for e in embeddings for sublist in e]` if you cannot upgrade.

from langchain_community.embeddings import LlamaCppEmbeddings
llama_embed = LlamaCppEmbeddings(model_path="./model.gguf", n_gpu_layers=10, embedding=True)
embeddings = llama_embed.embed_documents(["text"])