llama-anon's picture
Update README.md
3bd795d
|
raw
history blame
288 Bytes
metadata
license: agpl-3.0

LLaMA-13B merged with Instruct-13B weights, unlike the bare weights it does not output gibberish.

Quantized to 4bit (with mmap enabled)

Feel free to donate:

XMR: 86Z8nLSVPx3SZ5z7iWugeK5JruAeGPUJyExD9e3wdTSxUvFMhGXNG9ucPqCm8M29y1AxP6ta56GBQ4GiEUMzeew9MfX1yct