feat: add llama2_7b

README.md

@@ -3,4 +3,33 @@ base_model:
 - karpathy/tinyllamas
 ---
 This is a Llama 2 architecture model series quantized for use with [llama2.c](https://github.com/karpathy/llama2.c) project.
-This repository is not official llama2.c.
+This repository is not official llama2.c.
+
+Usage:
+```bash
+./runq ~/Data/GitHub/llama2.c-models/llama2_7b_q80.bin -z ~/Data/GitHub/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:
+```

llama2_7b_q80-tokenizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50a52ef822ee9e83de5ce9d0be0a025a773d019437f58b5ff9dcafb063ece361
+size 433869

llama2_7b_q80.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bf687ddfdc4eb749e9a1d51b97110bb988fe3b4180901b290b22818bb505b97
+size 7160348928