IBM (International Business Machines), US software giant IBM is looking into ways to develop Malayalam-focused large language models (LLMs). In this context, an announcement is anticipated on Friday, the last day of the nation’s inaugural GenAI Conclave in Kochi.
International GenAI Conclave – Kochi | IBM
InstructLab, a new platform created by IBM and Red Hat for augmenting LLMs with a client’s own data, is the foundation of IBM’s Gen AI Innovation Centre in Kochi. Additionally, IBM Watsonx.AI, the data platform, and AI assistant technologies will be utilized.
What is InstructLab? (redhat.com)
Among the luminaries in attendance at the International GenAI Conclave were Pinarayi Vijayan, the Chief Minister of Kerala, and Dinesh Nirmal, Senior Vice President, Products, IBM Software.
The occasion provided a forum for talking about the most recent developments, difficulties, and possibilities in GenAI. Discussions about its function in promoting innovation in business, government, startups, and academia were held by the participants.
Chief Minister Pinarayi Vijayan spoke to the crowd and said Kerala was proud to host the first-ever GenAI Conclave in India, emphasizing the event’s importance in promoting fresh perspectives, teamwork, and homegrown inventions.
In recent years, there has been a notable advancement and use of large language models, or LLMs. The following are some noteworthy LLMs created by different businesses:
GPT-4o (OpenAI): The newest and most sophisticated language model was released by OpenAI. It delivers faster token production and is 50% less expensive than GPT-4. Interestingly, it has multimodal text, image, video, and voice capabilities.
Claude 3 (Human): The Claude 3 from Anthropic is a formidable rival to GPT-4. In contrast to GPT-4’s 32,000 tokens, it can analyze up to 200,000 tokens, or 150,000 words. Anthropic has received substantial funding from Amazon, which values the company at $15 billion.
With 314 billion parameters, Grok-1 (xAI), created by Elon Musk’s xAI, is presently the biggest open-source LLM.
Gemini (Google): Another notable LLM2 is Google’s Gemini, which is utilized in their chatbot of the same name.
LLaMA (Meta): The LLaMA model family from Meta is especially noteworthy.
Mistral AI Models: Mistral AI has made a number of models available, such as Mixtral 8x22B1 and Mistral 7B.
These LLMs are essential to software development, content generation, and chatbot applications.
