Apply for community grant: Academic project
We are building an open-sourced LLM for Japanese. It is very important to utilize LLM for languages other than English, but it is very challenging. We believe our project will gain popularity, and our project should be widely tested with easy and rich interfaces like Hugging face spaces. Our team has already successfully published BERT model in Japanese, especially for the financial market, and we have state-of-the-art Japanese BERT and ELECTRA for financial markets. For example, we have published bert-small-japanese, electra-small-japanese-fin-discriminator, and so on. Therefore, also this time, we are planning to build and publish our LLMs. (Now, we published a dataset and a model.) Currently, our budget is very limited, and we are using tiny instances. However, to publish bigger and more beneficial models, we need richer backends, such as A10 instances. We highly require those instances with lower budgets. We appreciate your understanding and support!