FP8 E4M3 quantization degrades accuracy
evaluatedpost trainingin Nemotron 3 SuperNVIDIA
1 more from this model
FP8 E4M3 quantization degrades accuracy
evaluatedinference servingin Nemotron 3 SuperNVIDIA
specific method · filed under inference & serving
Quantization using the FP8 E4M3 format, without a more specific target or scaling method established.
The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See inference quantization for the methods filed beside it.
2 quotes from 2 sources, one card per model, each opening with its strongest.
FP8 E4M3 quantization degrades accuracy
FP8 E4M3 quantization degrades accuracy