Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,43 @@
|
|
1 |
---
|
2 |
license: cc-by-sa-4.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: cc-by-sa-4.0
|
3 |
---
|
4 |
+
|
5 |
+
<div style="width: 800px; margin: auto;">
|
6 |
+
<h3>This is a sharded fp16 version of <a rel="noopener nofollow" href="https://huggingface.co/Tap-M/Luna-AI-Llama2-Uncensored-FP16/tree/main">Tap-M/Luna-AI-Llama2-Uncensored-FP16</a></h3>
|
7 |
+
|
8 |
+
<h2>Model Description</h2>
|
9 |
+
<p>“Luna AI Llama2 Uncensored” is a Llama2 based Chat model <br />fine-tuned on over 40,000 long form chat discussions <br />
|
10 |
+
This model was fine-tuned by Tap, the creator of Luna AI. <br />
|
11 |
+
|
12 |
+
<h2>Model Training</h2>
|
13 |
+
<p>The fine-tuning process was performed on an 8x a100 80GB machine.
|
14 |
+
<br />The model was trained on synthetic outputs which include multiple rounds of chats between Human & AI.
|
15 |
+
</p>
|
16 |
+
|
17 |
+
<a rel="noopener nofollow" href="https://huggingface.co/TheBloke/Luna-AI-Llama2-Uncensored-GPTQ">4bit GPTQ Version provided by @TheBloke - for GPU inference</a><br />
|
18 |
+
<a rel="noopener nofollow" href="https://huggingface.co/TheBloke/Luna-AI-Llama2-Uncensored-GGML">GGML Version provided by @TheBloke - For CPU inference</a>
|
19 |
+
|
20 |
+
|
21 |
+
<h2>Prompt Format</h2>
|
22 |
+
<p>The model follows the Vicuna 1.1/ OpenChat format:</p>
|
23 |
+
|
24 |
+
```
|
25 |
+
USER: I have difficulties in making friends, and I really need someone to talk to. Would you be my friend?
|
26 |
+
|
27 |
+
ASSISTANT: Of course! Friends are always here for each other. What do you like to do?
|
28 |
+
|
29 |
+
```
|
30 |
+
|
31 |
+
<h2>Benchmark Results</h2>
|
32 |
+
|
33 |
+
||||||
|
34 |
+
|---:|---:|---:|---:|---:|
|
35 |
+
|Task|Version| Metric |Value |Stderr|
|
36 |
+
|arc_challenge|0|acc_norm|0.5512|0.0146|
|
37 |
+
|hellaswag|0||||
|
38 |
+
|mmlu|1|acc_norm|0.46521|0.036|
|
39 |
+
|truthfulqa_mc|1|mc2|0.4716|0.0155|
|
40 |
+
|Average|-|-|0.5114|0.0150|
|
41 |
+
|
42 |
+
</div>
|
43 |
+
|