Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
license_link: https://huggingface.co/microsoft/phi-2/resolve/main/LICENSE
|
4 |
+
language:
|
5 |
+
- en
|
6 |
+
widget:
|
7 |
+
- text: Hello who are you?
|
8 |
+
example_title: Identity
|
9 |
+
- text: What can you do?
|
10 |
+
example_title: Capabilities
|
11 |
+
- text: Create a fastapi endpoint to retrieve the weather given a zip code.
|
12 |
+
example_title: Coding
|
13 |
+
tags:
|
14 |
+
- convAI
|
15 |
+
- conversational
|
16 |
+
pipeline_tag: text-generation
|
17 |
+
inference: false
|
18 |
+
---
|
19 |
+
|
20 |
+
<!-- description start -->
|
21 |
+
# Phi-2 Super (SFT + cDPO)
|
22 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/62ceeb27e7f6014c0e9d9268/5-LQCMrXi8FN_ewcWL47v.png)
|
23 |
+
- **Model creator:** [Anton Bacaj](https://huggingface.co/abacaj)
|
24 |
+
- **Original model:** [Phi-2 Super](https://huggingface.co/abacaj/phi-2-super)
|
25 |
+
|
26 |
+
## Description
|
27 |
+
This repo contains 4-bit Marlin format model files for [abacaj's Phi-2 Super](https://huggingface.co/abacaj/phi-2-super)
|
28 |
+
|
29 |
+
### Phi-2-super (SFT + cDPO)
|
30 |
+
|
31 |
+
Base Model: [microsoft/phi-2](https://huggingface.co/microsoft/phi-2)
|
32 |
+
|
33 |
+
### Chat template
|
34 |
+
|
35 |
+
The model uses the same chat template as found in Mistral instruct models:
|
36 |
+
|
37 |
+
```python
|
38 |
+
text = "<|endoftext|>[INST] What is your favourite condiment? [/INST]"
|
39 |
+
"Well, I'm quite partial to a good squeeze of fresh lemon juice. It adds just the right amount of zesty flavour to whatever I'm cooking up in the kitchen!<|endoftext|> "
|
40 |
+
"[INST] Do you have mayonnaise recipes? [/INST]"
|
41 |
+
```
|
42 |
+
|
43 |
+
### MT-bench / heval
|
44 |
+
|
45 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/62ceeb27e7f6014c0e9d9268/lnFu3x1ufdpQVysIrX4-G.png)
|
46 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/62ceeb27e7f6014c0e9d9268/mJfBpH8dIW7Ii2KAGI_A7.png)
|
47 |
+
<!-- description end -->
|