Autoresearch for GPU kernels. Give it any PyTorch model, go to sleep, wake up to optimized Triton kernels.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).