Abdullah-Rizwan commited on
Commit
d342e49
Β·
verified Β·
1 Parent(s): 9e2236d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +102 -14
README.md CHANGED
@@ -1,14 +1,102 @@
1
- ---
2
- title: Ai Lecture Generator
3
- emoji: 🏒
4
- colorFrom: yellow
5
- colorTo: pink
6
- sdk: gradio
7
- sdk_version: 5.23.3
8
- app_file: app.py
9
- pinned: false
10
- license: mit
11
- short_description: 'An AI agent that converts lecture audio into clean Notes '
12
- ---
13
-
14
- Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # πŸŽ“ AI Lecture Notes Generator
2
+
3
+ Transform lecture audio into structured, comprehensive notes using AI.
4
+
5
+ ## Features
6
+
7
+ - πŸŽ™οΈ Transcribes lecture audio using OpenAI's Whisper Large V3 model
8
+ - πŸ“ Generates well-structured markdown notes using Meta-Llama-3.1-8B
9
+ - 🧠 Identifies key concepts, points, and takeaways automatically
10
+ - πŸš€ Simple web interface built with Gradio
11
+ - ⚑ Optimized with 4-bit quantization for efficient inference
12
+
13
+ ## Demo
14
+
15
+ ![AI Lecture Notes Generator Demo](https://path-to-your-demo-image-or-gif.gif)
16
+
17
+ ## How It Works
18
+
19
+ 1. **Audio Transcription**: Converts lecture audio to text using Whisper Large V3
20
+ 2. **Note Generation**: Processes transcription through Llama 3.1 to create structured notes
21
+ 3. **Format Cleaning**: Ensures consistent markdown formatting with proper headers
22
+
23
+ ## Installation
24
+
25
+ ```bash
26
+ # Clone the repository
27
+ git clone https://github.com/yourusername/ai-lecture-notes-generator.git
28
+ cd ai-lecture-notes-generator
29
+
30
+ # Install dependencies
31
+ pip install -r requirements.txt
32
+ ```
33
+
34
+ ## Usage
35
+
36
+ ### Option 1: Run the web interface
37
+
38
+ ```python
39
+ python app.py
40
+ ```
41
+
42
+ Then open your browser and navigate to the displayed URL (typically http://127.0.0.1:7860/).
43
+
44
+ ### Option 2: Use the API programmatically
45
+
46
+ ```python
47
+ from lecture_notes_generator import process_audio
48
+
49
+ # Generate notes from an audio file
50
+ notes = process_audio("path_to_lecture_audio.mp3")
51
+ print(notes)
52
+ ```
53
+
54
+ ## Requirements
55
+
56
+ - Python 3.8+
57
+ - PyTorch
58
+ - Transformers
59
+ - Gradio
60
+ - Hugging Face account with appropriate model access
61
+ - CUDA-compatible GPU (recommended)
62
+
63
+ ## Output Format
64
+
65
+ The generated notes follow this structure:
66
+
67
+ ```markdown
68
+ # Topic: [Lecture Topic]
69
+ ## Instructor
70
+ [Instructor name if detected]
71
+ ## Summary
72
+ [Concise summary of the lecture content]
73
+ ## Key Concepts
74
+ - [Important concept 1]
75
+ - [Important concept 2]
76
+ ...
77
+ ## Key Points
78
+ - [Major point 1]
79
+ - [Major point 2]
80
+ ...
81
+ ## Takeaways
82
+ - [Takeaway 1]
83
+ - [Takeaway 2]
84
+ ...
85
+ ```
86
+
87
+ ## Limitations
88
+
89
+ - Best suited for academic lectures with clear structure
90
+ - Audio quality significantly impacts transcription accuracy
91
+ - May require fine-tuning for specialized terminology
92
+
93
+ ## License
94
+
95
+ MIT
96
+
97
+ ## Acknowledgements
98
+
99
+ - OpenAI for the Whisper model
100
+ - Meta for the Llama 3.1 model
101
+ - Hugging Face for model hosting and libraries
102
+ - Gradio for the web interface framework