Update README.md
Browse files
README.md
CHANGED
@@ 16,3 +16,27 @@ tags:


16 
 trl

17 
 sft

18 



















































16 
 trl

17 
 sft

18 


19 
+

20 
+
# xsanskarx/qwen20.5b_numina_mathinstruct

21 
+

22 
+
This repository contains a finetuned version of the Qwen2 0.5B model specifically optimized for mathematical instruction understanding and reasoning. It builds upon the Numina dataset, which provides a rich source of mathematical problems and solutions designed to enhance reasoning capabilities even in smaller language models.

23 
+

24 
+
## Motivation

25 
+

26 
+
My primary motivation is the hypothesis that highquality datasets focused on mathematical reasoning can significantly improve the performance of smaller models on tasks that require logical deduction and problemsolving. Uploading benchmarks is the next step in evaluating this claim.

27 
+

28 
+
## Model Details

29 
+

30 
+
* **Base Model:** Qwen2 0.5B

31 
+
* **Finetuning Dataset:** Numina COT

32 
+
* **Key Improvements:** Enhanced ability to parse mathematical instructions, solve problems, and provide stepbystep explanations.

33 
+

34 
+
## Usage

35 
+

36 
+
You can easily load and use this model with the Hugging Face Transformers library:

37 
+

38 
+
```python

39 
+
from transformers import AutoTokenizer, AutoModelForCausalLM

40 
+

41 
+
tokenizer = AutoTokenizer.from_pretrained("xsanskarx/qwen20.5b_numina_mathinstruct")

42 
+
model = AutoModelForCausalLM.from_pretrained("xsanskarx/qwen20.5b_numina_mathinstruct")
