trunk/25a44754189251b6259196617f14a002a4d87d15: [Special] Improve accuracy of erfinv near 1 (#201187)
- PyTorch 近 90 天出现 2544 次
- PyTorch 近 90 天第 2511 次发版
- 上一次:同一天稍早 · trunk/277652eef64acabd6d493a9bf28a398a7c22a1dd: [Test] Port XPU for test_masked.py (#189683)
发生了什么
By using erfc instead of erf for the approximation and adding 3rd step of Newton-Raphson approximation for double Apply the same fix to MPS. Metal erfinv never had Newton-Raphson refinement: #101507 ported the CPU rational approximation but dropped the correction steps (Metal had no erf at the time), leaving it accurate to only ~8e-6 relative (~67 fp32 eps), which was masked by the 1e-4 float32 OpInfo tolerance. CUD…
摘要按规则整理自下方来源原文