SCBconsulting commited on
Commit
04a1838
Β·
verified Β·
1 Parent(s): 2f5fee7

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +39 -12
README.md CHANGED
@@ -1,12 +1,12 @@
1
  ---
2
  license: mit
3
- title: AI Contract Intelligence Demo
4
  sdk: gradio
5
  emoji: πŸ“š
6
  colorFrom: purple
7
  colorTo: indigo
8
  thumbnail: >-
9
- https://cdn-uploads.huggingface.co/production/uploads/6893ae91c2759b97d106132e/ApY-ZT4UGyVbI2lPDt3_J.png
10
  short_description: 'AI-powered contract analysis: summary, metadata, clause risk'
11
  ---
12
 
@@ -24,16 +24,43 @@ short_description: 'AI-powered contract analysis: summary, metadata, clause risk
24
  ---
25
  # SYNCLM – AI Contract Intelligence Demo πŸ€–πŸ“„
26
 
27
- A live demo to showcase contract analysis using Hugging Face Transformers & Gradio.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
 
29
- **Features:**
30
- - Contract upload (.pdf, .docx, .txt)
31
- - Executive summarization
32
- - Metadata extraction (NER)
33
- - Clause risk detection
34
- - Fallback clause suggestions
35
- - Translation (EN β†’ PT via DeepL or Hugging Face)
36
 
37
- Built with: `Gradio`, `Transformers`, `Sentence Transformers`, `DeepL API`
 
 
 
 
 
 
 
38
 
39
- > Created by Sabrina Colirri – SCBconsulting
 
1
  ---
2
  license: mit
3
+ title: 'SYNCLM: AI Contract Intelligence Demo'
4
  sdk: gradio
5
  emoji: πŸ“š
6
  colorFrom: purple
7
  colorTo: indigo
8
  thumbnail: >-
9
+ https://cdn-uploads.huggingface.co/production/uploads/6893ae91c2759b97d106132e/zIJmobHAiLPyf-9FItm-W.png
10
  short_description: 'AI-powered contract analysis: summary, metadata, clause risk'
11
  ---
12
 
 
24
  ---
25
  # SYNCLM – AI Contract Intelligence Demo πŸ€–πŸ“„
26
 
27
+ # πŸ€– SYNCLM – AI Contract Intelligence Demo
28
+
29
+ An interactive legal tech prototype that showcases intelligent contract review using open-source models from Hugging Face.
30
+
31
+ ### πŸ” Features
32
+ - Upload legal documents (.pdf, .docx, .txt)
33
+ - Executive summarization (EN)
34
+ - Metadata extraction using NER (Named Entity Recognition)
35
+ - Clause risk detection (rule-based)
36
+ - Fallback clause suggestions (semantic similarity)
37
+ - Real-time translation (EN β†’ PT via Unicamp-dl)
38
+
39
+ ## 🧠 Models Used
40
+
41
+ | Task | Model Used |
42
+ |-------------------|---------------------------------------------------------|
43
+ | Summarization | `facebook/bart-large-cnn` |
44
+ | Metadata (NER) | `pierreguillou/ner-legal-ptbr-electra-small` |
45
+ | Risk Detection | Custom classifier (fine-tuned LegalBERT) |
46
+ | Translation | `unicamp-dl/translation-en-pt-t5` |
47
+
48
+ > All models sourced from [Hugging Face](https://huggingface.co/models)
49
+
50
+ ## πŸš€ Run Locally
51
+
52
+ Install dependencies:
53
 
54
+ ```bash
55
+ pip install -r requirements.txt
 
 
 
 
 
56
 
57
+ ### 🧠 Built With
58
+ - `Gradio`
59
+ - `Transformers`
60
+ - `Sentence Transformers`
61
+ - `Unicamp-dl`
62
+ - `Torch`
63
+ - `PyMuPDF`
64
+ - `python-docx`
65
 
66
+ > Created by **Sabrina Colirri** β€” SCB Consulting