3 ms·
I've always found the Gemma models to vastly under-perform on vision tasks compared to Qwen so that's nothing new.
by thot_experiment 4mo ago
I've always found the Gemma models to vastly under-perform on vision tasks compared to Qwen so that's nothing new.
- mountainriver 4mo agoThe Qwen series adopted vision wayyy earlier than anyone else. No idea why the other labs were sleeping on it but they had about 2 years of experimentation without any competition.