CoolFace
Modelpublic

rizju/JuriMind-Indian-Law-RAFT

sourceHugging Faceapache-2.0updated 10mo agoView on Hugging Face
0likes12downloads
Model Card

JuriMind-Indian-Law-RAFT

First open-source Indian Supreme Court AI trained with RAFT (2024)

  • —Base: Meta-Llama-3.1-8B-Instruct (merged 16-bit)
  • —Training: 800 high-quality RAFT examples from real Supreme Court judgments
  • —Method: LoRA (r=64) → 5 epochs → merged
  • —Specialty: Handles noisy retrieval, cites sources with ##begin_quote## ... ##end_quote##
  • —Perfect for legal research, case law analysis, judgment summarization

base_model: unsloth/meta-llama-3.1-8b-instruct-bnb-4bit tags:

  • —text-generation-inference
  • —transformers
  • —unsloth
  • —llama license: apache-2.0 language:
  • —en ---

Uploaded finetuned model

  • —Developed by: Rizwan
  • —License: apache-2.0
  • —Finetuned from model : unsloth/meta-llama-3.1-8b-instruct-bnb-4bit

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

Usage

python
from transformers import pipeline
pipe = pipeline("text-generation", model="rizju/JuriMind-Indian-Law-RAFT", device_map="auto")