A couple/multi of GGML conversions of the FrankesteinsMonster-13B merge by Blackroot. Due "Vocab size mismatch" error added_tokens.json was padded with dummy tokens to make up the difference and allow conversion. This may or may not affect model performance.
Please see the original repo for full description and full credits, short summary follows below:
This is a frankenstein model merge of Nous-Hermes-Llama2-13b + Llama-2-13B-Storywriter-LORA + limarp-llama2
. Both lora merges were done at full weight (1:1:1 ratio)
Prompt Format
The model follows the Alpaca prompt format:
Instruction:
<prompt>
Response:
<leave a newline blank for model to respond>
This model does not have enough activity to be deployed to Inference API (serverless) yet.
Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.