implementation detail · filed under optimization
NVFP4 fine-grained micro-block scaling
NVFP4 uses 16-element micro-block scaling, alongside E4M3 block scales, an FP32 global scale, and E2M1 elements.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
The NVFP4 format features: fine-grained (16 element) micro-block scaling, block scaling factors with E4M3 format, a second-level FP32 global scale, and the E2M1 element format.
usedoptimizationin Nemotron 3 familyNVIDIA
Filed alongside
Other methods under optimization :: training precision.
NVFP4BF16NVFP4 pre-trainingFP8 mixed-precision trainingFP4+FP8 mixed precisionFP8-precision reinforcement learningMXFP8BF16 gradient reductionBF16 mixed-precision trainingBlock-wise FP8 activation quantization with offloadE2M1FP32 attention-output retentionFP32 gradient reductionFP8 storage for the residual stateHigh-precision final network layersMixed-FP8 quantizationMixed-precision trainingTwo-dimensional block quantization