specific method · filed under inference & serving
FP8 E4M3 quantization
Uses the E4M3 FP8 format for quantization; the cited evidence does not specify the tensors being quantized.
- sources
- 2
- model
- 1
- labs adopt it
- 0
- strongest
- evaluated
How sources treat it
One count per evidence span, weakest treatment to strongest.
evaluated 2
Documented in
Evidence
2 spans quoted from the sources, strongest treatment first.
FP8 E4M3 quantization degrades accuracy
evaluatedpost trainingin Nemotron 3 SuperNVIDIA
FP8 E4M3 quantization degrades accuracy
evaluatedinference servingin Nemotron 3 SuperNVIDIA
Filed alongside
Other methods under inference & serving :: inference quantization.
FP8 KV-cache quantizationFP8FP4 KV-cache quantizationNVFP4 quantizationFour-Over-SixMXFP4 weight quantizationPost-training quantizationQuantizationBlock-scaled INT8 quantization with stochastic roundingGGUFMax-based scalingMSE-based scalingMXFP8 activation quantizationNVFP4 KV-cache quantizationNVFP4 ModelOpt re-quantizationSSM cache quantizationAWQ INT4 weight quantization (W4A16)BF16 inferenceBlock-wise E4M3 FP8 weight quantizationChannel-wise quantizationDynamic activation scalingEmbedding and KV-cache quantizationEmpirical bits-per-element budget selectionFine-grained FP8 quantization