Downloads · 30 days
36
20% of all-time downloads
tarekys5/qwen2.5-3B-egyptian-legalv2-wip
qwen2.5-3B-egyptian-legalv2-wip is a question answering model from tarekys5. Use it when the input is a question plus a passage. It is set up for transformers. The card lists the license as apache-2.0.
Research Preview: This is the second iteration (v2) of the Egyptian Legal fine-tuned model. It is currently in a Beta stage for evaluation and community feedback.
Downloads · 30 days
36
20% of all-time downloads
All-time downloads
178
Public
Parameters
3.1B
6.2 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors6.2 GB · 100%
From the Hugging Face model README
Research Preview: This is the second iteration (v2) of the Egyptian Legal fine-tuned model. It is currently in a Beta stage for evaluation and community feedback.
Building on the foundations of v1, this version (v2) aims to bridge the gap between complex Egyptian legislative language and AI understanding. By leveraging the advanced reasoning of Qwen2.5-3B, this model is tailored to handle the specific linguistic complexities and formal structures found in Egyptian Civil, Criminal, and Administrative codes.
Key Objective: To assist legal researchers and practitioners in navigating the Egyptian legal landscape through a highly specialized, instruction-following AI.
This version introduces a more refined training pipeline compared to its predecessor:
Qwen/Qwen2.5-3B-InstructCompared to general-purpose models, v2 excels in:
This model is a work-in-progress. * Accuracy: It may generate hallucinations or cite outdated articles.
from transformers import pipeline
pipe = pipeline("question-answering", model="tarekys5/qwen2.5-3B-egyptian-legalv2-wip")