specific method · filed under optimization
Fused loss computation
Computes policy-gradient and OPD losses together in one kernel, optionally including metrics.
- source
- 1
- models
- 2
- 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.
All loss computation is fused into one kernel: the policy-gradient loss (with or without top-prenormalization) and the OPD loss, optionally with metrics such as entropy, label logit, and top-p mass.
usedoptimizationin MiMo-V2.6Xiaomi
Filed alongside
Other methods under optimization.