ullama2.c: run llama2 on MCUs (minimum 480KB SRAM)
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).