Alibaba's Qwen1.5 series has introduced a new model, the Qwen1.5-32B, which is generating significant interest in the tech community. This 32 billion parameter model boasts impressive language understanding, multilingual support, coding, and mathematical abilities, making it competitive with larger models like the 72B version. The Qwen1.5-32B model, with a 32k context size and using DPO for preference training, is designed to be a mid-size option, providing a balance between capability and ease of deployment, particularly on devices with limited resources. It has been praised for its high-accuracy, low-latency performance across various benchmarks, including code, reasoning, and math, and for its decent multilingual capabilities. The release of Qwen1.5-32B and its chat variant has been met with enthusiasm, with applications already being deployed in arenas like Chatbot Arena, where it's expected to perform well. The model is available for commercial use under a custom license and has been integrated into platforms like Together API, indicating a broad potential for application.
Alibaba-Qwen Releases Qwen1.5 32B: A New Multilingual dense LLM with a context of 32k and Outperforming Mixtral on the Open LLM Leaderboard Quick read: https://t.co/GUSQPacRoW Demo: https://t.co/aH8eS8N4hh @Alibaba_Qwen #ArtificialInteligence #LLMs
Another week in Open ML 🥳 - Cohere Command R+ - Google Gemma Instruct 1.1 - Qwen 1.5 32B model family is out - JetMoE - Sailor: LMs for South-East Asia - Mixture of Depths replication - Two different bitnet 1.5 open-source replications Open ML going brrr 🚀
More options for Qwen, more choices for MetaGPT, Qwen1.5 - 32B has just been released. 🙌Head over to MetaGPT to configure and try it out: https://t.co/1P8w3gIb8l 👇 https://t.co/52OBvDSJOL
More choices for Qwen, more choices for MetaGPT Qwen1.5 - 32B just released. 🙌Come configure and experience it on MetaGPT: https://t.co/1P8w3gIb8l https://t.co/52OBvDSJOL
New Gemma instruct models (2B & 7B) now on MLX 🚀 You can run inference, and (Q)LoRA fine-tuning locally on your Mac. > pip install -U mlx_lm I’m getting 19-22 tokens/s on my M1 Air 🚀 Model cards 👇🏾 https://t.co/74CoyjsB4J https://t.co/wx6RoTBgHK
Qwen 1.5 is out! @ollama #ai https://t.co/iFcrYUCIzt
What an incredible way to end the week -- C4AI Command R+ claims #1 spot on @huggingface leaderboard!🎉 We released 104B open weights with the goal of making generative AI breakthroughs accessible to the research community - so exciting to see such a positive response. 🔥 https://t.co/PqrYAfuRBR
Qwen1.5-32B is here! - Very strong metrics across benchmarks (code, reasoning, and math benchmarks) - Mid-size (32 billion params, easier to run on-device) - Decent multilingual capabilities Website: https://t.co/jrb1N2753k Demo: https://t.co/Mn3RyxZ9aU https://t.co/MAh1bwK6LW
Now in a 30B model Qwen1.5 series, is a useful medium weight local model. Link: https://t.co/ue6J7Tyvd6 https://t.co/ptVwVearIc
New models from @Alibaba_Qwen available on the Together API! https://t.co/W71pajXkXA https://t.co/djYtPkP9JB Read more on the Qwen blog post: https://t.co/RnrCMV3PJu https://t.co/1OCRWpAyIO
Congrats @Alibaba_Qwen on the new Qwen1.5-32B release🔥 We have shipped it to Arena, thanks to the fast & reliable endpoint support by @togethercompute! Qwen1.5 72B has been the best open model on Chatbot Arena leaderboard. Very excited to see how the 32B performs! https://t.co/raPCjdNibJ https://t.co/kyrWjij4VL
Qwen-32B is out, with very competitive performance against Mixtral. https://t.co/TG2tkRq626 https://t.co/2skaFGhRy5
Today, we release a new model of the Qwen1.5 series: Qwen1.5-32B and Qwen1.5-32B-Chat! Blog: https://t.co/HG9xXU3Bn1 HF: https://t.co/oE1DBcrRNq , search repos with “Qwen1.5-32B” in model names. GitHub: https://t.co/5vKV1KFwfy For a long time, our users have been requesting us… https://t.co/EtpmtB36rT
New open LLM from @Alibaba_Qwen! Qwen1.5 32B is a new multilingual dense LLM with a context of 32k, outperforming Mixtral on the open LLM Leaderboard! 🌍🚀 TL;DR 🧮 32B with 32k context size 💬 Chat model used DPO for preference training 📜 Custom License, commercially useable… https://t.co/8FkH021SPz
🤔 Have you met this problem before? 14B is not capable enough, but 72B is too large? 🎾 A 30B model might be the sweet spot, and now we finally have it, the new member of Qwen1.5 series, Qwen1.5-32B! Blog: https://t.co/b5VeO7H6Ep HF: https://t.co/VmLL1lGOa0 , search repos… https://t.co/ZawClzFzS3
🏠 Welcome to the Qwen1.5 family, the new dense model member, Qwen1.5-32B! This model has shown competitive performance comparable to the 72B model, especially impressing in language understanding, multilingual support, coding and mathematical abilities. But beyond that,… https://t.co/O4gcL1WeDM
Super happy to announce the release of h2o-danube2-1.8b, a 1.8 billion parameter foundation LLM. We adapted our original model and continued training with an additional 2T tokens which makes it the best model to date on Open LLM Leaderboard benchmark. https://t.co/wZGT2uFogM https://t.co/s2630mVam3
Thrilled to unveil how to deploy a quantized Llama2-7B model using MLC on NVIDIA Jetson Orin NX 16GB! 🚀 Edge devices can now leverage high-accuracy, low-latency language models. Perfect for smart assistants & more. Dive into the future of edge AI:https://t.co/Hqd3T6WSWw 🌐…
Qwen1.5-72B outperforms GPT-4 on our own cultural benchmark. Y’all should all try Qwen for other languages. https://t.co/S6BA3ArSZQ