Spec Decoding (DSpark) finally works with Pipeline Parallelism in vLLM!! 🔥 A lot of the GPU poor/proletarian have been asking for this feature for a while now, but it took the recent Kimi K3's massive 2.8T parameters affecting the GPU middle class (B200) for it to be implemented! Before this change, GPU poors that needed to use pipeline parallelism, as their weights didn't fully fit on 1 server, would not be able to take advantage of DSpark. Shoutout to yongqinwang-cmd & Inferact for implementing it in vLLM!

Tired of decoding meeting notes that say, “As you can see on this slide…” without the actual visual? 🖼️ “Take notes for me” in Google Meet now includes screenshots of presented slides and visuals directly within the meeting notes to help you get a more complete picture. Learn more in our latest drop →

Yesterday's special: decoding. Today's special: new routes. ✨ 2027 new destinations: Philadelphia to Reykjavik (Airbus A321neo) Philadelphia to Vienna (A321XLR) Philadelphia to Porto (A321XLR) 2027 new routes: New York to Amsterdam (A321XLR) New York to Nice (A321XLR) Chicago to Tokyo (Boeing 787-9) Charlotte to Barcelona (777-200ER) New York to London Heathrow (787-9) (Fourth daily flight added) 2027 seasonal extensions: Charlotte to Paris (777-200ER) Miami to Milan (787-8) Which route are you most excited for? Send this to your travel buddy and start planning now. Flights go on sale this Monday. ✈️

vLLM integrated PyNvVideoCodec to offload video decoding from CPU to GPU's NVDEC. Result: 2x+ throughput at 8×H100, CPU bottleneck gone! This is huge for video captioning at scale (AV training, metadata), and ships with CUDA vLLM releases. 🔗