{"slug": "ktransformers-0-7-expands-avx-512-support-to-benefit-amd-epyc-servers", "title": "KTransformers 0.7 Expands AVX-512 Support to Benefit AMD EPYC Servers", "summary": "KTransformers 0.7, the heterogeneous LLM inference and fine-tuning framework, is now available with full AVX-512 support for LoRA fine-tuning that no longer requires Advanced Matrix Extensions (AMX), benefiting AMD EPYC Zen 4, Zen 5, and Zen 6 servers and older Intel Xeon processors. The release also adds VLM fine-tuning support, native FP8 LoRA support, improved DeepSeek V4 deployment, and CPU activation reuse.", "body_md": "# KTransformers 0.7 Expands AVX-512 Support To Benefit AMD EPYC Servers\n\nKTransformers as the framework for heterogeneous LLM inference and fine-tune optimizations is out today with its v0.7 feature release.\n\nWith KTransformers 0.7 there is now full\n\nThe\n\nKTransformers 0.7 also adds VLM fine-tuning support, native FP8 LoRA support, improved DeepSeek V4 deployment, and CPU activation reuse.\n\nMore details on KTransformers 0.7 for those using it for LLM inference optimizations and fine-tuning can find all the details via the release announcement on\n\nWith KTransformers 0.7 there is now full\n\n[AVX-512](https://www.phoronix.com/search/AVX-512)support for LoRA fine-tuning without depending upon Advanced Matrix Extensions ([AMX](https://www.phoronix.com/search/AMX)) also being present. This AVX-512-only without AMX benefits AMD EPYC Zen 4 / Zen 5 / Zen 6 servers with excellent AVX-512 support while lacking AMX and also older Intel Xeon processors with AVX-512 prior to the introduction of AMX with Sapphire Rapids.The\n\n[merge request](https://github.com/kvcache-ai/ktransformers/pull/2141)noted the testing on AMD hardware and the foxus on AVX-512 without AMX platforms. The KTransformers runtime will automatically select the proper CPU implementation and in turn allowing MoE expert training to happen on a wider range of large-memory servers.KTransformers 0.7 also adds VLM fine-tuning support, native FP8 LoRA support, improved DeepSeek V4 deployment, and CPU activation reuse.\n\nMore details on KTransformers 0.7 for those using it for LLM inference optimizations and fine-tuning can find all the details via the release announcement on", "url": "https://wpnews.pro/news/ktransformers-0-7-expands-avx-512-support-to-benefit-amd-epyc-servers", "canonical_source": "https://www.phoronix.com/news/KTransformers-0.7", "published_at": "2026-08-17 17:20:02+00:00", "updated_at": "2026-08-17 17:41:20.233452+00:00", "lang": "en", "topics": ["machine-learning", "large-language-models", "developer-tools"], "entities": ["KTransformers", "AMD", "Intel", "DeepSeek"], "alternates": {"html": "https://wpnews.pro/news/ktransformers-0-7-expands-avx-512-support-to-benefit-amd-epyc-servers", "markdown": "https://wpnews.pro/news/ktransformers-0-7-expands-avx-512-support-to-benefit-amd-epyc-servers.md", "text": "https://wpnews.pro/news/ktransformers-0-7-expands-avx-512-support-to-benefit-amd-epyc-servers.txt", "jsonld": "https://wpnews.pro/news/ktransformers-0-7-expands-avx-512-support-to-benefit-amd-epyc-servers.jsonld"}}