The effect is quite good
It only requires 90 billion parameters to surpass widely used commercial models such as GPT-4o and Gemini 2.0 Pro. This is truly exciting🤩. In use, the effect is quite good.
It only requires 90 billion parameters to surpass widely used commercial models such as GPT-4o and Gemini 2.0 Pro. This is truly exciting🤩. In use, the effect is quite good.
Thanks for the kind words! Just a small correction — MiniCPM-o 4.5 has 9 billion parameters (9B), not 90 billion.
It only requires 90 billion parameters to surpass widely used commercial models such as GPT-4o and Gemini 2.0 Pro. This is truly exciting🤩. In use, the effect is quite good.
Thanks for the kind words! Just a small correction — MiniCPM-o 4.5 has 9 billion parameters (9B), not 90 billion.
Real decent speech chat model!
It only requires 90 billion parameters to surpass widely used commercial models such as GPT-4o and Gemini 2.0 Pro. This is truly exciting🤩. In use, the effect is quite good.
Thanks for the kind words! Just a small correction — MiniCPM-o 4.5 has 9 billion parameters (9B), not 90 billion.
You guys clearly deserve to pat yourselves in the back. I legit don't understand HOW you guys were able to make such a pathetically small model (being barely a fraction of your average generalist nowadays) be so smart and efficient, ON TOP of being multimodal. Its just crazy. I'm also a huge fan of VoxCPM, so this isn't the first time I was blown away by a product of yours.
You guys rock man, keep up the good work! The AI Revolution needs more people like you.
appreciation corner, +1 on this