AIbase
EN
Home
Llama 3 8b Patent Small Dataset
L

Llama 3 8b Patent Small Dataset

Developed by kimhyeongjun
A model fine-tuned on a small dataset of 16,000 English translations of Korean patents based on Meta-Llama-3-8B-Instruct, for testing purposes only.
Large Language Model
Transformers
 EnglishOpen Source License:Other#Patent text processing#Korean patent translation#Few-shot fine-tuning
Downloads 17
Release Time : 4/30/2024

Model Overview

This language model is optimized for patent text processing, fine-tuned from Llama-3-8B-Instruct, and suitable for patent-related text generation and comprehension tasks.

Model Features

Patent text optimization
Specially fine-tuned for patent texts, improving accuracy in handling patent-related content.
Small dataset training
Trained on 16,000 English translations of Korean patents, suitable for niche applications.
Efficient fine-tuning
Uses optimized training parameters and mixed-precision training to enhance training efficiency.

Model Capabilities

Patent text generation
Patent content comprehension
Technical document processing

Use Cases

Intellectual property
Patent abstract generation
Automatically generates abstracts based on patent technical descriptions
Patent technical analysis
Analyzes key technical points in patent documents
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase