implementation detail · filed under training objective
Continued pretraining for MTP layers
A dedicated continued-pretraining stage for the MTP layers.
- 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.
NVIDIA added a dedicated continued-pretraining stage for these MTP layers.
usedunclearin Nemotron 3.5 LightningNVIDIA
Filed alongside
Other methods under training objective :: multi-token prediction objective.