Edit model card

Information

OpenAssistant-Alpaca-13B-4-bit working with GPTQ versions used in Oobabooga's Text Generation Webui and KoboldAI.
This was made using https://huggingface.co/chavinlo/alpaca-13b and Serpdotai's Open Assistant 13b LoRa trained for 4 epochs using Open Assistant's dataset.

python llama.py /Models/alpaca13b-oaast4ep-lora c4 --wbits 4 --true-sequential --groupsize 128 --save_safetensors oasst-alpaca13b-4ep-lora-4bit-128g.safetensors

Benchmarks

--true-sequential --groupsize 128

Wikitext2: 6.854333400726318

Ptb-New: 12.411578178405762

C4-New: 9.355494499206543

Note: This version uses --groupsize 128, resulting in better evaluations.

Downloads last month
3

Dataset used to train Monero/oasst-alpaca13b-4epoch-4bit-128g