DGX Spark 64GB vs 128GB, from the memory math:- Llama 3.3 70B Q4_K_M @32K: ~55.7 GB, fits the 64GB- GPT-OSS 120B MXFP4 @8K: ~66.5 GB, needs the 128GB- Both share 273 GB/s, so dense 70B tops out around 7 tok/shttps://quantized.uk/cookbook/dgx-spark-64gb-vs-128gb/#LocalLLM #DGXSpark #LLM #SelfHosted
Related
#TodayILearned that Google has made SynthID #AI #slop image detection available on https://synthid.com. I used to go sep...
#TodayILearned that Google has made SynthID #AI #slop image detection available on https://synthid.com. I used to go separately to OpenAI and Gemini to check if images were slop-ge...
ソラリスではGolden Gateはどう見られているんでしょうかApple、Mac向けに不具合修正とセキュリティアップデートを含んだ「macOS 27.2.1 Golden Gate」をリリース。 https://applech2.com/...
ソラリスではGolden Gateはどう見られているんでしょうかApple、Mac向けに不具合修正とセキュリティアップデートを含んだ「macOS 27.2.1 Golden Gate」をリリース。 https://applech2.com/archives/20261110-macos-27-2-1-golden-gate-update.html#Apple...
ランスはFinal Cut Proをどう思うかな……Apple starts offering $50 extended usage packs for AI features in Creator Studio https://www.e...
ランスはFinal Cut Proをどう思うかな……Apple starts offering $50 extended usage packs for AI features in Creator Studio https://www.engadget.com/2311500/apple-starts-offering-50-dollar-extended...