Precision (original) (raw)
Toggle site navigation sidebar
Toggle table of contents sidebar
Toggle table of contents sidebar
class composer.core.Precision(value)[source]#
Enum class for the numerical precision to be used by the model.
FP32#
Use 32-bit floating-point precision. Compatible with CPUs and GPUs.
AMP_FP16#
Use torch.cuda.amp with 16-bit floating-point precision. Only compatible with GPUs.
AMP_BF16#
Use torch.cuda.amp with 16-bit BFloat precision.
AMP_FP8#
Use transformer_engine.pytorch.fp8_autocast
with 8-bit FP8 precison.