Lab notes from v37: reproducibility recipes, infrastructure write-ups, research.https://v37.ai/en-ushttps://v37.ai/research/deterministic-stack-mi100/https://v37.ai/research/deterministic-stack-mi100/What it actually takes to make a Qwen3.6-MoE LoRA train reproducibly on AMD MI100s. Three patches, one well-known invariant, and why the same inputs really do produce the same weights.Mon, 27 Apr 2026 01:00:00 GMTreproducibilitytrainingrocmqwenmoev37