refactor: Update GenerateEmbeddingsForm to support batch processing

refactor: Update embedding batch size handling in RAG configuration

refactor: add query_doc query caching

refactor: update logging statements in generate_chat_completion function

change embedding_batch_size to Optional
This commit is contained in:
Peter De-Ath
2024-09-26 23:28:47 +01:00
parent 09f34a7561
commit 885b9f1ece
6 changed files with 116 additions and 132 deletions

View File

@@ -986,10 +986,10 @@ RAG_EMBEDDING_MODEL_TRUST_REMOTE_CODE = (
os.environ.get("RAG_EMBEDDING_MODEL_TRUST_REMOTE_CODE", "").lower() == "true"
)
RAG_EMBEDDING_OPENAI_BATCH_SIZE = PersistentConfig(
"RAG_EMBEDDING_OPENAI_BATCH_SIZE",
"rag.embedding_openai_batch_size",
int(os.environ.get("RAG_EMBEDDING_OPENAI_BATCH_SIZE", "1")),
RAG_EMBEDDING_BATCH_SIZE = PersistentConfig(
"RAG_EMBEDDING_BATCH_SIZE",
"rag.embedding_batch_size",
int(os.environ.get("RAG_EMBEDDING_BATCH_SIZE", "1")),
)
RAG_RERANKING_MODEL = PersistentConfig(