A deep learning package for many-body potential energy representation and molecular dynamics
LGPL-3.0 License
Bot releases are visible (Hide)
Published by amcadmus about 3 years ago
New features:
init-frz-model
option for se_r
and se_t
(#1144)Enhancement for developers:
Bug fixings:
Published by amcadmus about 3 years ago
New features:
Enhancements:
np.testing.assert_almost_equal
for array comparing (#1059)allow_growth
in default_tf_session_config
(#1067)Bug fixings:
all_virial
for float precision (#1069)lcueve.out
->lcurve.out
(#1077)rcut
and sel
in the example of se_e3
(#1082)std::numeric_limits
(#1113)Published by amcadmus about 3 years ago
Breaking changes to v1.3
dp convert-from
to convert old models to v2.deepmd.DeepPot
has been moved to deepmd.infer.DeepPot
.
NNPInter
has been renamed to deepmd::DeepPot
and NNPInter.h
has been renamed to DeepPot.h
. Use -ldeepmd_cc
to link instead.New features
se_e3
)hybrid
)Performance improvement:
format_nlist_b
(#832 #845)scan_nlist
kernel (#1028)Enhancements
nopbc
and related docs (#994)dp --version
(#995)tensorboard_freq
to control sampling ratio during training. (#996)viewcode
and intersphinx
(#997)model.get_ntypes()<data.get_ntypes()
(#1016)descrpt/se_e2_a
(#1017)fit/ener
(#1024)InputNlist
into API doc (#1009)Improvement of the code for developers
deepmd
namespace for the C++ APIsel
is not enough (#914)exclude_types
(#1005)Bug fixings:
using namespace std
. Solve compiling compatability problem.cuda
memory access error #566prod_virial_grad
that causes wrong results when training with virials #685prod_virial_grad
that causes wrong results when training with virials #685DeepmdData.type_map
to input type_map
(#924)deepmd
and deepmd.entrypoints
. (#922)decay_rate
to stop_lr
from old inputs (#949)except
and limit the try
clause (#977)CMAKE_COMPILER_IS_GNUCXX
(#1038)Published by amcadmus about 3 years ago
New features:
sel
from the training data. (#831)Performance improvement:
Enhancements:
Bug fixings:
InvalidArgumentError
caused by zero sel
and optimize zero matrix (#900)auto_sel
(#911)Published by amcadmus over 3 years ago
New feature:
Performance improvement:
Enhancement:
Bug fixings
box.npy
is not necessary for nopbc (#810)NOT_LOADABLE
should be tuple
(#825)Published by amcadmus over 3 years ago
New features:
Enhancements
Bug fixings:
Published by amcadmus over 3 years ago
New hardware support:
Enhancement:
sel
Bug fixing:
prod_virial_grad
that causes wrong results when training with virials #685Published by amcadmus over 3 years ago
Increment to v2.0.0-alpha:
New features:
Enhancement:
Bug fixings
Published by amcadmus over 3 years ago
What's new to v2.0.0-alpha.0
Enhancement:
Bug fixings:
Published by amcadmus over 3 years ago
The very first alpha release of deepmd-kit version 2.0.0. It includes the following new features
Enhancements
Improvement of the code for developers
Bug fixings:
using namespace std
. Solves compiling compatability problem.deepmd
namespace for the C++ APIPublished by amcadmus over 3 years ago
Bug fixing:
Published by amcadmus over 3 years ago
Improvements:
Bug fixings:
Published by amcadmus over 3 years ago
Bug fixings
Published by amcadmus over 3 years ago
New feature:
Bug fixing:
Published by amcadmus almost 4 years ago
Bug fixing:
loss
in dipole and polarizability trainingImprovement:
Published by amcadmus almost 4 years ago
New features:
Improvements:
Bug fixings:
coeff_ener
when coeff_atom_ener
is given #290Published by amcadmus about 4 years ago
New features:
Bug fixings:
Published by amcadmus over 4 years ago
Bug fixing:
Published by amcadmus over 4 years ago
New features of dp train:
stop_lr
, the decay_rate
will be computed automaticallyTF_INTRA_OP_PARALLELISM_THREADS
and TF_INTER_OP_PARALLELISM_THREADS
to control the multi-threading of tf, clean up command line options.systems
is provided with a string, all possible systems will be recursively searched within the dir given by systems
New features of dp test:
Bug fixings:
Published by amcadmus over 4 years ago
Big fixings: