OpenBMB, known for open-sourcing the UltraFeedback dataset, has released a new series of language and vision models, including MiniCPM and MiniCPM-V, aimed at edge computing. MiniCPM, a 2 billion parameter model, is reported to compete with and even outperform the 7 billion parameter Mistral model and the 13 billion parameter Llama2 model in comprehensive rankings. MiniCPM's overall score also surpasses that of the 40 billion parameter Falcon model. MiniCPM-V, a 3 billion parameter vision language model, is designed for edge deployment, outperforming similar-sized models and is ideal for mobile deployment. These models are available on Huggingface and are intended for research and limited commercial use. OpenBMB's release also includes a 2.4 billion parameter base model comparable to Mistral-7B and a 2.4 DPO model that surpasses the performance of Llama-70B on MT Bench. Additionally, a 3 billion bilingual VLLM and a 12 billion RLHF VLLM have been released, further expanding the capabilities of edge computing models.
OpenBMB, the creators of UltraFeedback, silently released a series of very strong edge models! - 2.4B base model close to Mistral 7B - 2.4 DPO outperforming Llama 70B on MT Bench - A 3B bilingual VLLM (+12B version RLHF VLLM) Check the models at https://t.co/sXHjVpvgDh 🚀
MiniCPM from @OpenBMB is on the Hub 🔥 ✨ Competitive performance to Mistral-7B on general benchmarks ✨ MiniCPM-V outperforms models with similar size ✨ Ideal for mobile deployment ✨ For research and limited commercial use https://t.co/UwVt95CjCm
MiniCPM from Openbmb is on the Hub 🔥 ✨ Competitive performance to Mistral-7B on general benchmarks ✨ MiniCPM-V outperforms models with similar size ✨ Ideal for mobile deployment ✨ For research and limited commercial use https://t.co/UwVt95CjCm
MiniCPM from Openbmb is on the Hub 🔥 ✨ Competitive performance to Mistral-7B on general benchmarks ✨ MiniCPM-V outperforms models with similar size ✨ Ideal for mobile deployment ✨ For research and limited commercial use https://t.co/UwVt95CjCm More details:…
MiniCPM from Openbmb is on the Hub 🔥 ✨ Competitive performance to Mistral-7B on general benchmarks ✨ MiniCPM-V outperforms models with similar size ✨ Ideal for mobile deployment ✨ For research and limited commercial use More details: https://t.co/HahZZ6BvYn…
Vision LLM for #edgecomputing ? Check out this new release on @huggingface @OpenBMB, who OS'ed the UltraFeedback dataset before, released a series of strong eco-friendly yet powerful LLMs - MiniCPM: 2B model that competes with Mistral-7B - MiniCPM-V: 3B vision LLM on edge! https://t.co/Cz8HP4EUUL
🌟Meet #MiniCPM: An end-side LLM outperforms Llama2-13B. 🏆It is similar to Mistral-7B in the comprehensive rankings and the overall score surpasses Llama2-13B,Falcon -40B and other models. GitHub:https://t.co/qRfCCD7C9O Huggingface:https://t.co/vpZrk7NOAM #MiniCPM https://t.co/T4uOyk4jwC