File size: 1,501 Bytes
d576ac4
 
 
 
 
 
 
 
 
 
a1c453a
d576ac4
0b2bc4c
 
 
d576ac4
 
0b2bc4c
 
d576ac4
 
0b2bc4c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d576ac4
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
---
language:
- en
license: apache-2.0
tags:
- text-generation-inference
- transformers
- unsloth
- mistral
- trl
- sft
base_model: unsloth/mistral-7b-bnb-4bit
datasets:
- m-a-p/Code-Feedback
- knowrohit07/saraswati-stem
---

# **StemCoder-7B Test**
_This is not a full, comprehensive finetune. Stem-Coder-7B was a test to finetune Mistral-7B on 9k instructions (albeit, many are multi-turn) via a free Google Colab instance with [unsloth](https://github.com/unslothai/unsloth)._


<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/FRI6tBRlPqJb_qncxtFxO.png" width="500">

This model is finetuned for:

- Math
- Multi-Turn Coding
- STEM / Science / Medicine

StemCoder gives really descriptive answers, particularly for science, math based questions.

# **Examples**

## **Science**
<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/UZQy1Otn7xmjMKx0UG1_4.png" width="800">
<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/Z1hiK-YrWt_A5q21DWL3v.png" width="800">


## **Coding**
<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/XttCe5JvpXgdwDn8xERWR.png" width="800">
<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/UL0I2q8YjrSooQsxUb27S.png" width="800">

## **Medicine**
<img src="https://cdn-uploads.huggingface.co/production/uploads/648a374f00f7a3374ee64b99/SfyvKBAUqba03ggHyxS52.png" width="800">