SOURCE-LINKED INTELLIGENCE

UE5M3 FP4 Block Scaling for Stable Language Model Pretraining

arXiv · AI, language, vision and robotics · article · Sep 2, 2026 · UTC

Stable 4-bit floating-point (FP4) pretraining is difficult because the E2M1 payload represents only a narrow range of magnitudes. NVIDIA's Transformer Engine \nv{} recipe addresses this with current-tensor scaling, a randomized Hadamard transform (RHT), and bfloat16 (BF16) final layers, adding work outside the FP4 matrix multiplications. We instead pair E2M1 payloads with unsigned E5M3 (\ue{}) block scales. Their wider range permits periodic tensor scaling, while our recipe applies selective stochastic rounding to backward gradients, omits RHT, and uses FP4 in all eligible internal linears. We

Read original source ↗ Open in workspace

recordType

paper

region

Global