NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO · Hugging Face

eneral-purpose models

• 1.1B: TinyDolphin 2.8 1.1B. Takes about ~700MB RAM and tested on my Pi 4 with 2 gigs of RAM. Hallucinates a lot, but works for basic conversation.


• 2.7B: Dolphin 2.6 Phi-2. Takes over ~2GB RAM and tested on my 3GB 32-bit phone via llama.cpp on Termux.


• 7B: Nous Hermes Mistral 7B DPO. Takes about ~4-5GB RAM depending on contex

... See more

pair-preference-model-LLaMA3-8B by RLHFlow: Really strong reward model, trained to take in two inputs at once, which is the top open reward model on RewardBench (beating one of Cohere’s).

DeepSeek-V2 by deepseek-ai (21B active, 236B total param.): Another strong MoE base model from the DeepSeek team. Some people are questioning the very high MMLU sc... See more

Microsoft introduced Phi 1.5 – a compact AI model with multimodal capabilities, meaning it can process images as well as text. Despite being significantly smaller than OpenAI's GPT-4, with only 1.3 billion parameters, it demonstrates advanced features like those found in larger models. Phi 1.5 is open-source, emphasizing the trend towards efficient

... See more