Upload README.md
Browse files
README.md
CHANGED
|
@@ -9,4 +9,24 @@ app_file: app.py
|
|
| 9 |
pinned: false
|
| 10 |
---
|
| 11 |
|
| 12 |
-
Check out the
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
pinned: false
|
| 10 |
---
|
| 11 |
|
| 12 |
+
Check out the deployed demo version at https://huggingface.co/spaces/Entz/llm_5
|
| 13 |
+
|
| 14 |
+
This repository is the prototype in A prototype written in Python to illustrate/demonstrate querying a tiny dataset extracted from the Learning Path Index Dataset (see [Kaggle Dataset](https://www.kaggle.com/datasets/neomatrix369/learning-path-index-dataset)
|
| 15 |
+
|
| 16 |
+
The dataset can be found in the same folder.
|
| 17 |
+
|
| 18 |
+
## Pre-requisites
|
| 19 |
+
|
| 20 |
+
- Python 3.9.x or above
|
| 21 |
+
- OpenAI API Key
|
| 22 |
+
- Install dependencies from `requirements.txt`
|
| 23 |
+
- Basic Command-line experience
|
| 24 |
+
- Basic git and GitHub experience
|
| 25 |
+
|
| 26 |
+
## Known issue
|
| 27 |
+
|
| 28 |
+
- The AI can't do simple counting well. e.g. there are 3 courses in the dataset, but AI always returns 4 when its being asked how many courses in the database.
|
| 29 |
+
|
| 30 |
+
- The AI can't interpret course fee correctly when the fee is zero. Perhaps this can be fixed in prompt engineering.
|
| 31 |
+
|
| 32 |
+
- To get better accuracy in AI's interpretation on the dataset, further cleaning has to be done on the dataset.
|