|
--- |
|
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"> |
|
|
|
|
