https://www6v.github.io/www6vMLSys/ Recent content on MLSys Hugo en Tue, 23 Apr 2024 21:54:43 +0000 https://www6v.github.io/www6vMLSys/docs/Inference-Opt/%E6%A8%A1%E5%9E%8B%E5%B1%82%E4%BC%98%E5%8C%96/%E9%87%8F%E5%8C%96/PTQ/WeightActivation/QuantizationInt8/gptQuantizationInt8/ Thu, 12 Oct 2023 14:33:20 +0000 https://www6v.github.io/www6vMLSys/docs/Inference-Opt/%E6%A8%A1%E5%9E%8B%E5%B1%82%E4%BC%98%E5%8C%96/%E9%87%8F%E5%8C%96/PTQ/WeightActivation/QuantizationInt8/gptQuantizationInt8/ <p></p> <!-- more --> <h1 id="llmint8"> LLM.int8() <a class="anchor" href="#llmint8">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/LLM-int8-11dbfe21108480d0b6adc341aed9ec3d?pvs=4">(原理|实战)LLM.int8()</a></p> https://www6v.github.io/www6vMLSys/docs/Training/%E4%BD%8E%E7%B2%BE%E5%BA%A6/LowPrecision/gptLowPrecision/ Wed, 16 Aug 2023 11:50:58 +0000 https://www6v.github.io/www6vMLSys/docs/Training/%E4%BD%8E%E7%B2%BE%E5%BA%A6/LowPrecision/gptLowPrecision/ <p></p> <!-- more --> <h1 id="低精度训练"> 低精度训练 <a class="anchor" href="#%e4%bd%8e%e7%b2%be%e5%ba%a6%e8%ae%ad%e7%bb%83">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/cb067e2d0bc545d898cd43dd1091c8b3?pvs=4">低精度训练</a></p> https://www6v.github.io/www6vMLSys/docs/Training/%E4%BD%8E%E7%B2%BE%E5%BA%A6/Precision/gptPrecision/ Thu, 01 Feb 2024 22:29:42 +0000 https://www6v.github.io/www6vMLSys/docs/Training/%E4%BD%8E%E7%B2%BE%E5%BA%A6/Precision/gptPrecision/ <p></p> <!-- more --> <h1 id="混合精度"> 混合精度 <a class="anchor" href="#%e6%b7%b7%e5%90%88%e7%b2%be%e5%ba%a6">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/d27bdf000e7c42eabd288f9d036ea5e7?pvs=4">(原理|实战)混合精度</a></p> https://www6v.github.io/www6vMLSys/docs/Inference-Opt/%E6%A8%A1%E5%9E%8B%E5%B1%82%E4%BC%98%E5%8C%96/%E9%87%8F%E5%8C%96/PTQ/WeightActivation/QuantizationSmoothQuant/gptQuantizationSmoothQuant/ Sat, 14 Oct 2023 13:33:21 +0000 https://www6v.github.io/www6vMLSys/docs/Inference-Opt/%E6%A8%A1%E5%9E%8B%E5%B1%82%E4%BC%98%E5%8C%96/%E9%87%8F%E5%8C%96/PTQ/WeightActivation/QuantizationSmoothQuant/gptQuantizationSmoothQuant/ <p></p> <!-- more --> <h1 id="smoothquant"> SmoothQuant <a class="anchor" href="#smoothquant">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/SmoothQuant-11dbfe21108480fc83f8ea2a495092b7?pvs=4">(原理)SmoothQuant</a></p> https://www6v.github.io/www6vMLSys/docs/Inference-Opt/%E6%A8%A1%E5%9E%8B%E5%B1%82%E4%BC%98%E5%8C%96/%E9%87%8F%E5%8C%96/PTQ/WeightActivation/QuantizationFP8/gptQuantizationFP8/ Sat, 14 Oct 2023 13:33:46 +0000 https://www6v.github.io/www6vMLSys/docs/Inference-Opt/%E6%A8%A1%E5%9E%8B%E5%B1%82%E4%BC%98%E5%8C%96/%E9%87%8F%E5%8C%96/PTQ/WeightActivation/QuantizationFP8/gptQuantizationFP8/ <p></p> <!-- more --> <h1 id="fp8"> FP8 <a class="anchor" href="#fp8">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/FP8-12c3b58d744e43299dad0badc397f592?pvs=4">(原理)FP8</a></p> https://www6v.github.io/www6vMLSys/docs/LLMOps/MaaS/gptMaaSMonitor/ Tue, 23 Apr 2024 21:54:43 +0000 https://www6v.github.io/www6vMLSys/docs/LLMOps/MaaS/gptMaaSMonitor/ <p></p> <!-- more --> <h1 id="maas-监控"> MaaS 监控 <a class="anchor" href="#maas-%e7%9b%91%e6%8e%a7">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/MaaS-1d2bfe2110848079b06afffaea367fe9?pvs=4">MaaS 监控</a></p> https://www6v.github.io/www6vMLSys/docs/LLMOps/MaaS/gptLLMOpsPaaS/ Tue, 26 Sep 2023 23:18:45 +0000 https://www6v.github.io/www6vMLSys/docs/LLMOps/MaaS/gptLLMOpsPaaS/ <p></p> <!-- more --> <h1 id="llm-paas"> LLM PaaS <a class="anchor" href="#llm-paas">#</a> </h1> <p><a href="https://candied-skunk-1ca.notion.site/9ddf2032d70b4722ad34a48cb305d80b?pvs=4">LLM PaaS</a></p>