AIHPC x Large Language Model - BLOOM 176B Large Language Model Results Showcase!
What are the hottest AI innovation trends in 2023?
According to VentureBeat, an American technology blog, AI experts are particularly optimistic about three trends in AI innovation in 2023: 1. Workflow automation and "human-centric" AI solutions; 2. Data-driven AI decision-making; and 3. Generative AI applications. Natural Language Processing (NLP) and Large Language Models (LMM), two technologies within the first trend, are particularly promising for their ability to scale business operations and their progress.Obsessed with chatting on ChatGPT?
The recently popular ChatGPT is a chatbot developed based on a large-scale language model (LLM). An LLM is a natural language processing system with over 100 billion parameters, applicable to content generation, text summarization, chatbots, coding, and customized AI applications such as predicting protein structure and biomolecular properties. BLOOM (BigScience Large Open-science Open-access Multilingual Language Model) is the world's largest multilingual model, with 176 billion parameters, encompassing 46 languages and 13 programming languages. However, due to its massive data volume and model size, the development and maintenance of LLMs are extremely difficult, making them unsuitable for most companies to use and train.TWS Ready-to-Go for LLM!
Tai-Alibaba Cloud's world-class AIHPC supercomputer, with its GPU computing power, is the best practice platform for "high-performance computing." It can provide massive 176B model training and inference for BLOOM to run rapidly on TWCC. The training time gradually decreases as the number of nodes increases, achieving a near-linear and impressive acceleration result, solving the challenge of multi-node inference. At the same time, it has launched Taiwan's only "AI High-Computing Power Consultant" one-stop integrated service, providing AI experts, technical environment resources, and LLM development services to accelerate the transformation of enterprise needs into usable models and applications, realizing AI 2.0.
| Date: February 10, 2023 (Friday)
| Time: 15:00-16:00
| Target audience: NLP labs and related practitioners, data scientists, AI engineers, data analysts, cloud platform (next-generation IT) managers, machine learning engineers, technology professionals, computer science professors and students, etc.
You will learn:
- Trends and Applications of LLM
- TWCC's achievements in training large BLOOM models.
- TWCC helps optimize large model inference systems, successfully overcoming the challenges of multi-node inference.
- TWCC offers the only AI consulting service in Taiwan, realizing AI 2.0.
Organizer: Taiwan Smart Cloud Services Co., Ltd.
Co-organizers: Taiwan Cloud & IoT Industry Association, Artificial Intelligence Technology Foundation