File size: 1,485 Bytes
67d6075
e296f1a
 
67d6075
17a52b2
 
 
 
 
 
 
 
e296f1a
ced5145
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
db525a1
 
 
 
 
 
 
 
 
 
 
 
58b4744
 
 
 
 
 
db525a1
58b4744
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
---
language:
- en
library_name: transformers
tags:
- physics
- lora
- Qlora
- phi-2
datasets:
- ayoubkirouane/arxiv-physics
pipeline_tag: text-generation
---


## Phi-2-arxiv-physics  : 

Phi-2-arxiv-physics is a small language model (SLM) derived from Microsoft Research's Phi-2, a Transformer-based model with 2.7 billion parameters. 
Specifically fine-tuned using the innovative QLoRA (Quantized Low-rank Adapters) technique, this model exhibits remarkable capabilities in the domain of physics-related text.




## Fine-Tuning:

The model has been fine-tuned using **QLoRA**, a technique that reduces computational and memory requirements during the fine-tuning process. 

This two-stage approach involves quantizing parameters and introducing learnable low-rank adapters, resulting in faster fine-tuning with reduced memory demands.

## Data Source:

**Phi-2-arxiv-physics** is specifically trained on the **arXiv-physics** dataset, consisting of questions and answers related to physics. 

This dataset is sourced from the arXiv repository, encompassing a wide range of topics within the field of physics.


Usage (High-level) : 
```
!pip install -q -U git+https://github.com/huggingface/transformers.git

```

``` python

from transformers import pipeline

pipe = pipeline("text-generation", model="ayoubkirouane/phi-2-arxiv-physics")

## or :

# load it directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ayoubkirouane/phi-2-arxiv-physics")

```