Upload 2 files
Browse files- README.md +53 -0
- requirements.txt +6 -0
README.md
ADDED
@@ -0,0 +1,53 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Text Embedding Visualizer
|
2 |
+
|
3 |
+
This project generates embeddings for short sentences and visualizes them in 2D using PCA and t-SNE.
|
4 |
+
It works on **both CPU and GPU laptops** with the same dependencies.
|
5 |
+
|
6 |
+
## Install
|
7 |
+
```bash
|
8 |
+
## create a virtual environment
|
9 |
+
python -m venv venv
|
10 |
+
## activate the venv
|
11 |
+
source ./venv/bin/activate # Windows: ./venv/Scripts/activate
|
12 |
+
|
13 |
+
pip install -r requirements.txt
|
14 |
+
|
15 |
+
***How it Works***
|
16 |
+
## Loads a small dataset of sentences.
|
17 |
+
## Generates embeddings with all-MiniLM-L6-v2.
|
18 |
+
## Reduces dimensions using PCA and t-SNE.
|
19 |
+
## Visualizes them on a 2D plot.
|
20 |
+
|
21 |
+
|
22 |
+
## Example Output
|
23 |
+
|
24 |
+
When running the script, you will see:
|
25 |
+
followed by **two interactive plots**:
|
26 |
+
|
27 |
+
1. **PCA Visualization**
|
28 |
+
- Each dot represents a sentence.
|
29 |
+
- Sentences with similar meaning appear closer together.
|
30 |
+
- Example:
|
31 |
+
- "The Eiffel Tower is in France"
|
32 |
+
and "The capital of France is Paris"
|
33 |
+
are positioned near each other.
|
34 |
+
|
35 |
+
2. **t-SNE Visualization**
|
36 |
+
- Another dimensionality reduction method that shows natural clusters.
|
37 |
+
- Example:
|
38 |
+
- "Cats are amazing pets" and "Dogs are loyal companions"
|
39 |
+
appear together in one cluster, away from unrelated topics.
|
40 |
+
|
41 |
+
The plots help you **see how AI models understand meaning** in text.
|
42 |
+
|
43 |
+
---
|
44 |
+
|
45 |
+
## Sample Sentences Used
|
46 |
+
- Artificial intelligence is transforming the world.
|
47 |
+
- Cats are amazing pets.
|
48 |
+
- The capital of France is Paris.
|
49 |
+
- The Eiffel Tower is in France.
|
50 |
+
- Deep learning enables image recognition.
|
51 |
+
- Dogs are loyal companions.
|
52 |
+
- The sun rises in the east.
|
53 |
+
- The moon orbits the Earth.
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
torch==2.2.2
|
2 |
+
transformers==4.41.2
|
3 |
+
sentence-transformers==3.0.1
|
4 |
+
scikit-learn==1.5.0
|
5 |
+
matplotlib==3.9.0
|
6 |
+
numpy==1.26.4
|