Scale up AV foundation model pre-training and fine-tuning with data to its maximum capacity – into billions+ of examples - across different sources, delivering the maximum value to the model through every additional example.
Must-have
Nice-to-have
Not specified