Actions: google-research/neuralgcm
Actions
318 workflow runs
318 workflow runs
BatchGaussianRandomFieldModule
. These parameters will often be (effectively) constant in space/time, which means there are initialized with huge (correlation) values. As such, keeping things stable also required using jnp.expm1
to implement (1 - self.phi**2)
in the GaussianRandomField
.
CI
#311:
Pull request #89
synchronize
by
copybara-service
bot
BatchGaussianRandomFieldModule
. These parameters will often be (effectively) constant in space/time, which means there are initialized with huge (correlation) values. As such, keeping things stable also required using jnp.expm1
to implement (1 - self.phi**2)
in the GaussianRandomField
.
CI
#309:
Pull request #89
synchronize
by
copybara-service
bot
numpy<2
to avoid datetime incompatibilities.
CI
#299:
Commit a37f804
pushed
by
copybara-service
bot
numpy<2
to avoid datetime incompatibilities.
CI
#298:
Pull request #90
synchronize
by
copybara-service
bot
numpy<2
to avoid datetime incompatibilities.
CI
#297:
Pull request #90
synchronize
by
copybara-service
bot
numpy<2
to avoid datetime incompatibilities.
CI
#296:
Pull request #90
synchronize
by
copybara-service
bot
BatchGaussianRandomFieldModule
. These parameters will often be (effectively) constant in space/time, which means there are initialized with huge (correlation) values. As such, keeping things stable also required using jnp.expm1
to implement (1 - self.phi**2)
in the GaussianRandomField
.
CI
#295:
Pull request #89
synchronize
by
copybara-service
bot
numpy<2
to avoid datetime incompatibilities.
CI
#294:
Pull request #90
synchronize
by
copybara-service
bot