🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Founder mode

Should we pre-install Astral‘s uv along with pip and Miniforge‘s conda, in our dstack default Docker image?

New release alert! 🚨 Most importantly, it includes a fix for the Lambda provider now correctly showing all supported regions! Also a number of bugfixes. Thanks everyone who contributed! github.com/dstackai/dstac…

A month ago, we announced our performance roadmap. Today, we are happy to share that the latest release achieves 🚀2.7x higher throughput and is 5x faster for output latency on Llama 8B, and 1.8x higher throughput and 2x faster on Llama 70B for H100s. blog.vllm.ai/2024/09/05/per…

That's huge! 😍

👀

Interested in using Google Cloud TPUs for fine-tuning and deploying LLMs? Learn how dstack helps by leveraging open-source tools like @HuggingFace’s Optimum TPU and vLLM! The blog post includes examples and covers limitations and workarounds. dstack.ai/blog/tpu-on-gc…

👇 Nice blog post detailing how to use dstack for fine-tuning and deploying large language models on TPUs:

Who else thinks Harari’s new book is scam?

A new update is out! The CLI now works on Windows without needing WSL 2. Check out this and other changes in the update here: github.com/dstackai/dstac…

Let’s make AI more open and secure by hiding the model’s chain of thought from the user.

We aimed for AGI and ended up with LLMs that need a moment to gather their thoughts 😂

RAG, compound systems, agents, CoT, what's going on?

Our founder and CEO Andrey Cheptsov, explains why we built a new container orchestrator, how it streamlines AI development, and removes friction between AI and Ops teams youtube.com/watch?v=yzVMp5…

Our blog post on using TPUs for fine-tuning and deploying LLMs with dstack, Hugging Face's Optimum TPU and vLLM is now featured on Google Cloud's community blog! 😍 Thanks Romin Irani for featuring it! medium.com/google-cloud/u…

empathy is a premium model of our brain; when we run out of credits, the brain switches to the default model

Attending the HPC Roundtable event in Munich. Amazing networking experience with some of the HPC/AI infra vendors here incl. AMD Dell Huawei and others. Looking forward to discussing the use of containers in HPC/AI datacenters

eu/decel 🙁🇪🇺

What’s the best open-source LLM for generating code these days? 🧑🏽‍💻