GLM-130B by 
zai-org
Bilingual model for research and evaluation
Top 6.9% on sourcepulse
GLM-130B is an open-source, 130-billion parameter, bilingual (English/Chinese) language model designed for researchers and developers working with large-scale NLP models. It offers strong performance on various benchmarks and supports efficient inference through quantization and optimized libraries.
How It Works
GLM-130B utilizes the General Language Model (GLM) pre-training approach, which combines autoregressive blank infilling with bidirectional context. This allows it to excel at both left-to-right generation and filling in masked segments of text. The model is designed for efficient deployment, supporting INT4 quantization to enable inference on consumer-grade hardware.
Quick Start & Requirements
2 years ago
Inactive
Explore Similar Projects
OpenGVLab
Cornell-RelaxML
microsoft
rmihaylov
johnsmith0031
Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems), 
Jared Palmer(Ex-VP of AI at Vercel; Founder of Turborepo; Author of Formik, TSDX), and mpt-30B-inference by 
abacaj
Lysandre Debut(Chief Open-Source Officer at Hugging Face), Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems), and AQLM by 
Vahe1994
SandAI-org
yandex Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems), 
Georgios Konstantopoulos(CTO, General Partner at Paradigm), and 
qwopqwop200
lyogavin