This is a Llama 2 architecture model series quantized for use with llama2.c project. This repository is not official llama2.c.
Usage:
./runq ../llama2.c-models/llama2_7b_q80.bin -z ../llama2.c-models/llama2_7b_q80-tokenizer.bin -p 0.8 -n 0 -m chat
Enter system prompt (optional):
User: <s>[INST] Traduce a español lo siguiente "number of steps to run for, default 256."[/INST]
Assistant: El texto en español es:
Número de pasos para correr, predeterminado 256.
<s>
User: <s>[INST]Ahora traduce a chino[/INST]
Assistant: Sure! Here's the translation of "number of steps to run for, default 256" into Chinese:
数 для，默认为256。
Here's a breakdown of the translation:
* 数 (bùshuate) - number of steps
* для (dòu) - for
* (pàojiǔ) - to run
* 默认 (mòdé) - default
* 256 (èr bā) - a specific number
I hope this helps! Let me know if you have any other questions.
<s>
User:
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no library tag.
Model tree for elvisAR/llama2.c-models
Base model
karpathy/tinyllamas