Nous Research introduces the new flagship 7B model in the Hermes series, named Hermes 2-Mistral-7B DPO. The model, derived from OpenHermes 2.5, excelled in various benchmarks like AGIEval, BigBench Reasoning, GPT4All, and TruthfulQA. The company acknowledges @fluidstackio as the compute sponsor for this achievement. Teknium1 also presents their version, Nous-Hermes 2 Mistral 7B DPO, which outperformed the original Hermes model in key benchmarks, becoming their new flagship 7B model. Google releases competitive 2B and 7B open-weight models challenging Mistral and LLaMA2. The community reacts positively to these advancements, with plans to test and utilize the new models for further development.
Hermes 2 7B DPO is now on OpenRouter :) https://t.co/cKI4L3vj2m
Now live on OpenRouter: @NousResearch Hermes 2 Mistral DPO, by @Teknium1 Compare it side-by-side with Mistral: https://t.co/99Pb6uqo1K https://t.co/oeziboDufX
This is huge. @Google on a mega release of a family of Open LLM, Gemma (2B, 7B). Performance >= Mistral-7B Why huge? - support Keras/JAX/PyTorch/@huggingface safetensor - extensive resources for all kinds of ML ecosystem - JAX serving framework (saxml) https://t.co/WCagbsGi8m
Someone finally beats Mistral at the 7b size? We'll test these models ASAP as base models for fine tuning and share the results here! https://t.co/hR3QDRqLJ3
Google is on a roll! New 2B & 7B open-weight models with competitive performance against LLaMA2 and Mistral, a huge win for everyone building with local LLMs. Let's see how others will react. https://t.co/7Dv7HG0xIK https://t.co/i0CkMaTmqD
Damn finally a base 7B better than mistral https://t.co/2B9ZgZGBgq https://t.co/E7UWFY8hxD
Download Nous Hermes 2 Mistral 7B DPO (Q4_K_M) on Shoggoth: https://t.co/thmCHZDCdr model by Nous Research (@NousResearch) https://t.co/VrVDeIWCPN
Introducing our DPO'd version of the original OpenHermes 2.5 7B model - Nous-Hermes 2 Mistral 7B DPO! This model improved significantly on AGIEval, BigBench, GPT4All, and TruthfulQA compared to the original Hermes model, and is our new flagship 7B model! We at Nous are finding… https://t.co/OcdVwbuDbG https://t.co/oYa7K5YkuX
Announcing the new flagship 7B model in the Hermes series: Hermes 2-Mistral-7B DPO. A very special thanks to our compute sponsor for this run, @fluidstackio. https://t.co/QXsnxU8SvF This model was DPO'd from OpenHermes 2.5 and improved on all benchmarks tested - AGIEval,…
Announcing the new flagship 7B model in the Hermes series: Hermes 2-Mistral-7B DPO. This model was DPO'd from OpenHermes 2.5 and improved on all benchmarks tested - AGIEval, BigBench Reasoning, GPT4All, and TruthfulQA. A very special thanks to our compute sponsor for this run,…
Announcing the new flagship 7B model in the Hermes series: Hermes 2-Mistral-7B DPO. This model was DPO'd from OpenHermes 2.5 and improved on all benchmarks tested - AGIEval, BigBench Reasoning, GPT4All, and TruthfulQA. https://t.co/CTsU0NQkj2 A very special thanks to our…