Python API ReferenceΒΆ
- Common
- NdArray
NdArray
NdArray.bool_fill()
NdArray.cast()
NdArray.clear()
NdArray.clear_called
NdArray.copy_from()
NdArray.data
NdArray.data_ptr()
NdArray.dtype
NdArray.fill()
NdArray.from_numpy_array()
NdArray.get_data()
NdArray.masked_fill()
NdArray.modification_count
NdArray.narrow()
NdArray.ndim
NdArray.shape
NdArray.size
NdArray.size_from_axis()
NdArray.strides
NdArray.zero()
NdArray.zeroing
- Variable
Variable
Variable.apply()
Variable.backward()
Variable.bool_fill_()
Variable.clear_all_graph_links()
Variable.d
Variable.data
Variable.forward()
Variable.from_numpy_array()
Variable.function_references
Variable.g
Variable.get_unlinked_variable()
Variable.grad
Variable.info
Variable.masked_fill_()
Variable.ndim
Variable.need_grad
Variable.no_grad()
Variable.parent
Variable.persistent
Variable.recompute
Variable.reset_shape()
Variable.reshape()
Variable.rewire_on()
Variable.shape
Variable.size
Variable.size_from_axis()
Variable.unlinked()
Variable.visit()
Variable.visit_check()
- Computation Graph
- Functions
- Function
Function
Function.args
Function.backward()
Function.forward()
Function.grad_depends_output_data()
Function.info
Function.inplace_data()
Function.inplace_data_with()
Function.min_outputs()
Function.need_setup_recompute()
Function.recompute()
Function.set_active_input_mask()
Function.setup()
Function.setup_recompute()
Function.tags
PythonFunction
- List of Functions
- Neural Network Layers
affine()
convolution()
depthwise_convolution()
deconvolution()
depthwise_deconvolution()
deformable_convolution()
adaptive_separable_convolution()
max_pooling()
average_pooling()
global_average_pooling()
sum_pooling()
unpooling()
embed()
rnn()
lstm()
gru()
multi_head_attention()
patch_correlation()
roi_align()
- Neural Network Activation
- Normalization
batch_normalization()
fused_batch_normalization()
sync_batch_normalization()
mean_subtraction()
norm_normalization()
clip_by_value()
clip_grad_by_value()
clip_by_norm()
clip_grad_by_norm()
layer_normalization()
instance_normalization()
group_normalization()
weight_standardization()
weight_normalization()
spectral_norm()
- Reduction
- Arithmetic
- Logical
equal()
equal_scalar()
greater()
greater_equal()
greater_equal_scalar()
greater_scalar()
less()
less_equal()
less_equal_scalar()
less_scalar()
logical_and()
logical_and_scalar()
logical_not()
logical_or()
logical_or_scalar()
logical_xor()
logical_xor_scalar()
not_equal()
not_equal_scalar()
sign()
minimum2()
maximum2()
minimum_scalar()
maximum_scalar()
isnan()
isinf()
reset_nan()
reset_inf()
where()
- Math
- Array Manipulation
concatenate()
split()
stack()
slice()
gather()
gather_nd()
scatter_nd()
scatter_add()
pad()
transpose()
broadcast()
broadcast_to()
tile()
meshgrid()
flip()
shift()
sort()
shape()
reshape()
one_hot()
assign()
top_k_data()
top_k_grad()
pack_padded_sequence()
pad_packed_sequence()
searchsorted()
bool_gather()
bool_scatter()
bool_fill()
dot()
- Stochasticity
- Loss Functions
- Signal Processing
- Geometric Neural Network Layers
- Quantized Neural Network Layers
- Unsupported, Special Use
- Image Object Detection
- Validation
- Neural Network Layers
- Function
- Parametric Functions
- Parameter Management API
- List of Parametric Functions
parametric_function_api()
affine()
convolution()
depthwise_convolution()
deconvolution()
depthwise_deconvolution()
deformable_convolution()
batch_normalization()
fused_batch_normalization()
sync_batch_normalization()
mean_subtraction()
layer_normalization()
instance_normalization()
group_normalization()
rnn()
lstm()
gru()
embed()
prelu()
svd_affine()
svd_convolution()
cpd3_convolution()
binary_connect_affine()
binary_connect_convolution()
binary_weight_affine()
binary_weight_convolution()
inq_affine()
inq_convolution()
fixed_point_quantized_affine()
fixed_point_quantized_convolution()
min_max_quantized_affine()
min_max_quantized_convolution()
pow2_quantized_affine()
pow2_quantized_convolution()
pruned_affine()
pruned_convolution()
min_max_quantize()
lstm_cell()
LSTMCell
spectral_norm()
weight_normalization()
multi_head_attention()
transformer()
transformer_encode()
transformer_decode()
- Parameter Initializer
- Grad
- Solvers
- Solver
Solver
Solver.check_inf_grad()
Solver.check_inf_or_nan_grad()
Solver.check_nan_grad()
Solver.clear_parameters()
Solver.clip_grad_by_norm()
Solver.get_parameters()
Solver.get_states()
Solver.info
Solver.learning_rate()
Solver.load_states()
Solver.name
Solver.remove_parameters()
Solver.save_states()
Solver.scale_grad()
Solver.set_learning_rate()
Solver.set_parameters()
Solver.set_states()
Solver.set_states_from_protobuf()
Solver.set_states_to_protobuf()
Solver.setup()
Solver.update()
Solver.weight_decay()
Solver.weight_decay_is_fused()
Solver.zero_grad()
- List of solvers
- Solver
- Communicator
- Communicator interface
Communicator
Communicator.abort()
Communicator.add_context_and_parameters()
Communicator.all_gather()
Communicator.all_reduce()
Communicator.all_reduce_callback()
Communicator.allreduce()
Communicator.barrier()
Communicator.bcast()
Communicator.clear_context_parameters()
Communicator.find_group()
Communicator.init()
Communicator.list_groups()
Communicator.local_rank
Communicator.name
Communicator.new_group()
Communicator.rank
Communicator.reduce()
Communicator.reduce_scatter()
Communicator.size
- List of communicators
- Communicator interface
- Monitors
- Utils
- NNP save and load utilities
- Image Utils
- Data Iterators
- Debug Utils
- DLPack
- RNN Utils
- Misc
- Quantization Aware Training
- QATConfig
QATConfig
QATConfig.RecorderPosition
QATConfig.RoundingMethod
QATConfig.bn_folding
QATConfig.bn_self_folding
QATConfig.channel_last
QATConfig.channel_wise
QATConfig.dtype
QATConfig.ext_name
QATConfig.learning_rate_scale
QATConfig.narrow_range
QATConfig.niter_to_recording
QATConfig.niter_to_training
QATConfig.pow2
QATConfig.record_layers
QATConfig.recorder_activation
QATConfig.recorder_position
QATConfig.recorder_weight
QATConfig.round_mode
QATConfig.skip_bias
QATConfig.skip_inputs_layers
QATConfig.skip_outputs_layers
QATConfig.zero_point
- QATTensorRTConfig
- QATScheduler
- QATConfig
- Extensions
- Pretrained Models
- Out-of-core execution
- Modules
- Graph Definition
- Sequential
- Experimental
- Viewers
- Show Graph by Tensorboard
- Graph Converters
GraphConverter
FunctionModifier
- Function Modifiers
BatchNormalizationFoldingModifier
AddBiasModifier
BatchNormalizationFoldingModifierInner
BatchNormalizationFoldingOppositeModifierInner
BatchNormalizationSelfFoldingModifier
FusedBatchNormalizationModifier
UnfusedBatchNormalizationModifier
ChannelLastModifier
ChannelFirstModifier
RemoveFunctionModifier
BatchNormBatchStatModifier
TestModeModifier
IdentityModifier
NoGradModifier
PruningModifier
QuantizeNonQNNToRecordingModifier
QuantizeRecordingToTrainingModifier
- Trainers
- Mixed Precision Trainings
- DynamicLossScalingUpdater
DynamicLossScalingUpdater
DynamicLossScalingUpdater.solver
DynamicLossScalingUpdater.loss
DynamicLossScalingUpdater.data_feeder
DynamicLossScalingUpdater.scale
DynamicLossScalingUpdater.scaling_factor
DynamicLossScalingUpdater.N
DynamicLossScalingUpdater.clear_buffer
DynamicLossScalingUpdater.accum_grad
DynamicLossScalingUpdater.weight_decay
DynamicLossScalingUpdater.comm
DynamicLossScalingUpdater.grads
DynamicLossScalingUpdater.update()
- DynamicLossScalingUpdater
- Parametric Function Classes