zylon-ai
Any way to change size and amount of embeddings?
Right now each embedding is a maximum of 500 tokens and there are 4 embeds max. What if I want embeds to include less tokens but for the AI to use more of them?
The values for embedding size and overlap are in both ingest.py and privategpt.py
chunk_size = 500
chunk_overlap = 50
That is in bytes not tokens.
in privategpt.py you could change the 4 in:
target_source_chunks = int(os.environ.get('TARGET_SOURCE_CHUNKS',4))
to the value of your choice or define a different value for TARGET_SOURCE_CHUNKS in your .env file.
It seems the code has changed a lot. Is there any way now to change the amount of embeddings?
It seems the code has changed a lot. Is there any way now to change the amount of embeddings?
Also interested but this answer ! I need to analyse some logs into a JSON file. The file upload is okay, then I have the "Maximum number of input tokens"
