3 ms·
Really impressive for the size. The 4.2B vision model apparently outperforms GPT-4V-Turbo on 5 out of 7 vision benchmarks.
by causal 2y ago
Really impressive for the size. The 4.2B vision model apparently outperforms GPT-4V-Turbo on 5 out of 7 vision benchmarks.