# \#furiosa-llm **URL:** https://forums.furiosa.ai/tag/furiosa-llm/7.md [Latest](https://forums.furiosa.ai/latest.md) · [Categories](https://forums.furiosa.ai/categories.md) · [Tags](https://forums.furiosa.ai/tags.md) --- ## [Furiosa-LLM (FLM) 2026.4.0 release and its Q4 roadmap](https://forums.furiosa.ai/t/furiosa-llm-flm-2026-4-0-release-and-its-q4-roadmap/490)
**Author:** [@hyunsik](https://forums.furiosa.ai/u/hyunsik)\ **Replies:** 0\ **Last updated:** [October 13, 2026, 3:46pm UTC](https://forums.furiosa.ai/t/furiosa-llm-flm-2026-4-0-release-and-its-q4-roadmap/490 "2026-10-13T15:46:59Z")
Hi folks, We released Furiosa SDK 2026.4.0 today! See the release announcement for details. Looking ahead, here are our priorities for Q4: Speech recognition models and Transcribe API support Video modality support K… --- ## [Qwen3-VL-32B FXB가 furiosa-llm 2026.4.0b7에서 로드 실패 (unknown variant ShardingContext)](https://forums.furiosa.ai/t/qwen3-vl-32b-fxb-furiosa-llm-2026-4-0b7-unknown-variant-shardingcontext/455)
**Author:** [@jaeyeon.kim](https://forums.furiosa.ai/u/jaeyeon.kim)\ **Replies:** 1\ **Last updated:** [August 21, 2026, 4:28pm UTC](https://forums.furiosa.ai/t/qwen3-vl-32b-fxb-furiosa-llm-2026-4-0b7-unknown-variant-shardingcontext/455 "2026-08-21T16:28:28Z")
본문 요약: 32B는 furiosa-llm 2026.3.0 + v2026.3 아티팩트로 정상 구동됨을 확인했습니다(우회 완료). 다만 (a) 공개 인덱스의 최신 런타임 2026.4.0b7에서는 공개된 32B 아티팩트가 전부 로드 실패하고, (b) furiosa-llm serve furiosa-ai/Qwen3-VL-32B-Instruct가 기본으로 받는 main(b13) 아티팩트에 맞는 b8… --- ## [fxb build Qwen/Qwen3-8B-FP8 수행 시 tcc subprocess failed / os error 2 발생 문의](https://forums.furiosa.ai/t/fxb-build-qwen-qwen3-8b-fp8-tcc-subprocess-failed-os-error-2/444)
**Author:** [@k92492](https://forums.furiosa.ai/u/k92492)\ **Replies:** 1\ **Last updated:** [July 28, 2026, 4:55pm UTC](https://forums.furiosa.ai/t/fxb-build-qwen-qwen3-8b-fp8-tcc-subprocess-failed-os-error-2/444 "2026-07-28T16:55:07Z")
안녕하세요. RNGD 환경에서 furiosa-llm 기반 custom model smoke test를 진행 중입니다. 표준 LLM(…).generate() 경로는 정상 동작하지만, fxb build를 이용한 custom build 경로에서 host native와 Docker container 양쪽 모두 동일한 에러가 발생하여 문의드립니다. \[환경\] OS: Ubuntu 22.04.5 … --- ## [Furiosa SDK 2026.3.0 release](https://forums.furiosa.ai/t/furiosa-sdk-2026-3-0-release/417)
**Author:** [@hyunsik](https://forums.furiosa.ai/u/hyunsik)\ **Replies:** 0\ **Last updated:** [June 30, 2026, 1:31pm UTC](https://forums.furiosa.ai/t/furiosa-sdk-2026-3-0-release/417 "2026-06-30T13:31:47Z")
Hello! We are excited to announce the Furiosa SDK 2026.3.0 release. SDK 2026.3.0 has officially been published today. Here are the release note and documents: Release Note (2026.3.0) Furiosa Docs (2026.3.0) Upgrade Gui… --- ## [Throughput 성능 최적화를 위해 TP=1일 때, furiosa artifact를 만드는 과정에서 에러가 나왔습니다](https://forums.furiosa.ai/t/throughput-tp-1-furiosa-artifact/378)
**Author:** [@cake](https://forums.furiosa.ai/u/cake)\ **Replies:** 2\ **Last updated:** [December 5, 2025, 4:44am UTC](https://forums.furiosa.ai/t/throughput-tp-1-furiosa-artifact/378 "2025-12-05T04:44:48Z")
안녕하세요. 퓨리오사 레니게이드에서 빌드 중에 궁금한 것이 있어 문의드립니다. Throughput 성능 최적화 방법 문의 글을 참고하여 모델 artifact를 추가해 빌드를 시도했습니다. from furiosa\_llm.artifact.builder import ArtifactBuilder RELEASE\_PREFILL\_BUCKETS = \[ (1, 256), (1, 320), (1,… --- ## [Models with sliding window attention](https://forums.furiosa.ai/t/models-with-sliding-window-attention/372)
**Author:** [@huijong.jeong](https://forums.furiosa.ai/u/huijong.jeong)\ **Replies:** 3\ **Last updated:** [November 4, 2025, 2:50am UTC](https://forums.furiosa.ai/t/models-with-sliding-window-attention/372 "2025-11-04T02:50:08Z")
\[EN\] I want to benchmark LLM models involving sliding window attention via furiosa-llm. Unfortunately, my model’s architecture is not on your supported models list and and it seems like the only way to meet my goal is … --- ## [Throughput 성능 최적화 방법 문의](https://forums.furiosa.ai/t/throughput/362)
**Author:** [@gspark](https://forums.furiosa.ai/u/gspark)\ **Replies:** 1\ **Last updated:** [September 10, 2025, 5:26pm UTC](https://forums.furiosa.ai/t/throughput/362 "2025-09-10T17:26:39Z")
저희 모델(Llama3.1 모델을 post training한것)을 furiosa\_llm으로 빌드한것보다 Furiosa에서 이미 최적화하여 업로드 하신 동일 아키텍처 artifact가 30배 이상 token generation throughput 성능이 좋은 결과를 실험적으로 확인하였습니다. 이와 관련하여 동일 아키텍처 모델의 throughput 성능 최적화 방법에 대하여 문의드립니다. F… --- ## [Elice furiosa-smi info command error](https://forums.furiosa.ai/t/elice-furiosa-smi-info-command-error/326)
**Author:** [@jinpark](https://forums.furiosa.ai/u/jinpark)\ **Replies:** 1\ **Last updated:** [August 6, 2025, 3:04am UTC](https://forums.furiosa.ai/t/elice-furiosa-smi-info-command-error/326 "2025-08-06T03:04:59Z")
3 lspci -nn | grep FuriosaAI 4 sudo apt update && sudo apt install -y pciutils 5 sudo update-pciids 6 sudo apt update && sudo apt install -y curl gnupg 7 curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | s… --- ## [Furiosa-llm server 명령어 실패](https://forums.furiosa.ai/t/furiosa-llm-server/316)
**Author:** [@jinpark](https://forums.furiosa.ai/u/jinpark)\ **Replies:** 1\ **Last updated:** [July 31, 2025, 2:06am UTC](https://forums.furiosa.ai/t/furiosa-llm-server/316 "2025-07-31T02:06:41Z")
uriosa-smi info furiosa-smi: /workspace/source/furiosa-smi/target/release/build/furiosa-smi-6d8e10dd8c9557b1/out/hwloc-2.10.0/hwloc/topology-linux.c:4694: look\_sysfscpukinds: Assertion nr\_pus \> 0’ failed. Aborted (core… --- ## [Vision Language Model이 지원될 수 있는 시기가 궁금합니다](https://forums.furiosa.ai/t/vision-language-model/307)
**Author:** [@saibie1677](https://forums.furiosa.ai/u/saibie1677)\ **Replies:** 1\ **Last updated:** [July 23, 2025, 3:18pm UTC](https://forums.furiosa.ai/t/vision-language-model/307 "2025-07-23T15:18:23Z")
25년 국산 AI 반도체 활용사업에 참여하여, 엘리스 클라우드를 통해 rngd를 활용하고 있습니다. 사업에서 제공한 상품 선택 안내서에서는 Qwen이나 Gemma를 지원한다고 되어 있었는데 막상 사업에 참여하게 되니 해당 모델들은 Future Release에 포함되어 있었습니다. 이 사업은 연말에 상품개발이 완료되는 것을 목표로 진행이 되는데 해당 모델들은 언제 지원이 가능하게 될 지 … --- ## [A brief introduction to Furiosa SDK 2025.2](https://forums.furiosa.ai/t/a-brief-introduction-to-furiosa-sdk-2025-2/283)
**Author:** [@hyunsik](https://forums.furiosa.ai/u/hyunsik)\ **Replies:** 0\ **Last updated:** [June 1, 2025, 10:54pm UTC](https://forums.furiosa.ai/t/a-brief-introduction-to-furiosa-sdk-2025-2/283 "2025-06-01T22:54:16Z")
Here is a brief introduction to Furiosa SDK 2025.2. Especially, this slide introduces a lot of new features of Furiosa-LLM.