3 ms·
DeepSeek has shown that significantly less costly training is possible when incentivized. Even for SOTA models.
by timschmidt 1y ago
DeepSeek has shown that significantly less costly training is possible when incentivized. Even for SOTA models.