dragon
A Computation Graph Virtual Machine based ML Framework
File Explorer
Download Latest Version (.zip)- bug-report.md
- documentation-issue.md
- Codegen.cmake
- Dependencies.cmake
- FindBLAS.cmake
- FindCUDA.cmake
- FindCUDNN.cmake
- FindMLU.cmake
- FindMPI.cmake
- FindMPS.cmake
- FindNumPy.cmake
- FindOpenMP.cmake
- FindProtobuf.cmake
- FindPythonLibs.cmake
- LinkLibrary.cmake
- MiscCheck.cmake
- SelectCudaArch.cmake
- StripDebugInfo.cmake
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- context.py
- iterator.py
- pipeline.py
- types.py
- __init__.py
- bbox_ops.py
- builtin_ops.py
- decoder_ops.py
- generic_ops.py
- image_ops.py
- math_ops.py
- random_ops.py
- reader_ops.py
- __init__.py
- __init__.py
- pytorch.py
- __init__.py
- Dockerfile
- Dockerfile
- README.md
- dragon.png
- logo.png
- tick.png
- favicon.ico
- CPUContext.rst
- CUDAContext.rst
- Graph.rst
- MLUContext.rst
- MPSContext.rst
- Operator.rst
- OpSchema.rst
- Tensor.rst
- TypeMeta.rst
- UnifiedMemory.rst
- Workspace.rst
- core.rst
- footer.html
- header.html
- layout.xml
- conf.py
- Doxyfile
- index.rst
- make.bat
- Makefile
- BbFlip.rst
- BBoxPaste.rst
- Brightness.rst
- BrightnessContrast.rst
- Cast.rst
- CoinFlip.rst
- ColorSpaceConversion.rst
- ColorTwist.rst
- Contrast.rst
- CropMirrorNormalize.rst
- Erase.rst
- ExternalSource.rst
- GaussianBlur.rst
- Hsv.rst
- ImageDecoder.rst
- ImageDecoderRandomCrop.rst
- Normalize.rst
- Pad.rst
- Paste.rst
- RandomBBoxCrop.rst
- RandomResizedCrop.rst
- Reshape.rst
- Resize.rst
- Rotate.rst
- Slice.rst
- TFRecordReader.rst
- Uniform.rst
- WarpAffine.rst
- device.rst
- get_device_type.rst
- get_distributed_info.rst
- Iterator.rst
- ops.rst
- Pipeline.rst
- set_optimization.rst
- set_scheduler.rst
- set_verbosity.rst
- bitwise_and.rst
- bitwise_or.rst
- bitwise_xor.rst
- invert.rst
- current_device.rst
- get_device_capability.rst
- get_device_count.rst
- get_device_name.rst
- is_available.rst
- memory_allocated.rst
- set_cublas_flags.rst
- set_cudnn_flags.rst
- set_default_device.rst
- set_device.rst
- set_random_seed.rst
- Stream.rst
- synchronize.rst
- all_gather.rst
- all_reduce.rst
- broadcast.rst
- get_backend.rst
- get_group.rst
- get_rank.rst
- get_world_size.rst
- is_cncl_available.rst
- is_initialized.rst
- is_mpi_available.rst
- is_nccl_available.rst
- new_group.rst
- reduce_scatter.rst
- from_dlpack.rst
- to_dlpack.rst
- debug.rst
- error.rst
- fatal.rst
- get_verbosity.rst
- info.rst
- log.rst
- set_directory.rst
- set_verbosity.rst
- warning.rst
- ctc_loss.rst
- l1_loss.rst
- l2_loss.rst
- nll_loss.rst
- sigmoid_cross_entropy_loss.rst
- sigmoid_focal_loss.rst
- smooth_l1_loss.rst
- softmax_cross_entropy_loss.rst
- abs.rst
- add.rst
- affine.rst
- argmax.rst
- argmin.rst
- atan2.rst
- ceil.rst
- clip.rst
- cos.rst
- cummax.rst
- cummin.rst
- cumsum.rst
- div.rst
- equal.rst
- exp.rst
- floor.rst
- gemm.rst
- greater.rst
- greater_equal.rst
- is_finite.rst
- is_inf.rst
- is_nan.rst
- less.rst
- less_equal.rst
- log.rst
- logical_and.rst
- logical_not.rst
- logical_or.rst
- logical_xor.rst
- matmul.rst
- max.rst
- maximum.rst
- mean.rst
- min.rst
- minimum.rst
- mul.rst
- nan_to_num.rst
- negative.rst
- norm.rst
- not_equal.rst
- pow.rst
- reciprocal.rst
- round.rst
- rsqrt.rst
- sigmoid.rst
- sign.rst
- sin.rst
- sqrt.rst
- square.rst
- sub.rst
- sum.rst
- tanh.rst
- top_k.rst
- var.rst
- current_device.rst
- get_device_capability.rst
- get_device_count.rst
- get_device_name.rst
- is_available.rst
- memory_allocated.rst
- set_cnnl_flags.rst
- set_default_device.rst
- set_device.rst
- set_random_seed.rst
- synchronize.rst
- current_device.rst
- get_device_count.rst
- get_device_family.rst
- is_available.rst
- memory_allocated.rst
- set_default_device.rst
- set_device.rst
- set_random_seed.rst
- synchronize.rst
- batch_norm.rst
- bias_add.rst
- channel_norm.rst
- channel_shuffle.rst
- conv.rst
- conv1d.rst
- conv1d_transpose.rst
- conv2d.rst
- conv2d_transpose.rst
- conv3d.rst
- conv3d_transpose.rst
- conv_transpose.rst
- depth_to_space.rst
- drop_block.rst
- drop_path.rst
- dropout.rst
- elu.rst
- gelu.rst
- group_norm.rst
- GRU.rst
- hardsigmoid.rst
- hardswish.rst
- instance_norm.rst
- layer_norm.rst
- leaky_relu.rst
- local_response_norm.rst
- log_softmax.rst
- lp_norm.rst
- LSTM.rst
- moments.rst
- pool.rst
- pool1d.rst
- pool2d.rst
- pool3d.rst
- prelu.rst
- relu.rst
- relu6.rst
- RNN.rst
- selu.rst
- silu.rst
- softmax.rst
- space_to_depth.rst
- sync_batch_norm.rst
- BackendRep.rst
- export.rst
- prepare_backend.rst
- record.rst
- run_model.rst
- supports_device.rst
- Adam.rst
- AdamW.rst
- Optimizer.rst
- RMSprop.rst
- SGD.rst
- glorot_normal.rst
- glorot_uniform.rst
- multinomial.rst
- normal.rst
- normal_like.rst
- permutation.rst
- set_seed.rst
- truncated_normal.rst
- uniform.rst
- uniform_like.rst
- get_build_info.rst
- get_include.rst
- get_lib.rst
- extract_patches.rst
- resize.rst
- roi_align.rst
- roi_pool.rst
- argsort.rst
- assign.rst
- autograph.rst
- bitwise.rst
- boolean_mask.rst
- broadcast_to.rst
- cast.rst
- concat.rst
- constant.rst
- cuda.rst
- device.rst
- DeviceSpec.rst
- distributed.rst
- dlpack.rst
- eager_mode.rst
- expand_dims.rst
- eye.rst
- eye_like.rst
- fill.rst
- flatten.rst
- function.rst
- gather.rst
- gather_elements.rst
- get_num_threads.rst
- get_workspace.rst
- GradientTape.rst
- graph_mode.rst
- identity.rst
- linspace.rst
- load_library.rst
- logging.rst
- losses.rst
- math.rst
- mlu.rst
- mps.rst
- name_scope.rst
- nn.rst
- nonzero.rst
- one_hot.rst
- ones.rst
- ones_like.rst
- onnx.rst
- optimizers.rst
- pad.rst
- python_plugin.rst
- random.rst
- range.rst
- repeat.rst
- reset_workspace.rst
- reshape.rst
- reverse.rst
- roll.rst
- scatter_add.rst
- scatter_elements.rst
- set_num_threads.rst
- shape.rst
- slice.rst
- sort.rst
- split.rst
- squeeze.rst
- stack.rst
- stop_gradient.rst
- sysconfig.rst
- Tensor.rst
- tile.rst
- transpose.rst
- tril.rst
- triu.rst
- unique.rst
- unstack.rst
- variable_scope.rst
- vision.rst
- where.rst
- Workspace.rst
- zeros.rst
- zeros_like.rst
- bitwise_and.rst
- bitwise_or.rst
- bitwise_xor.rst
- invert.rst
- as_dtype.rst
- DType.rst
- elu.rst
- exponential.rst
- get.rst
- hard_sigmoid.rst
- linear.rst
- relu.rst
- selu.rst
- sigmoid.rst
- softmax.rst
- swish.rst
- tanh.rst
- Constant.rst
- get.rst
- GlorotNormal.rst
- GlorotUniform.rst
- Initializer.rst
- Ones.rst
- RandomNormal.rst
- RandomUniform.rst
- TruncatedNormal.rst
- VarianceScaling.rst
- Zeros.rst
- Activation.rst
- Add.rst
- AveragePooling1D.rst
- AveragePooling2D.rst
- AveragePooling3D.rst
- BatchNormalization.rst
- Concatenate.rst
- Conv1D.rst
- Conv1DTranspose.rst
- Conv2D.rst
- Conv2DTranspose.rst
- Conv3D.rst
- Conv3DTranspose.rst
- Dense.rst
- Dropout.rst
- ELU.rst
- Flatten.rst
- GlobalAveragePooling1D.rst
- GlobalAveragePooling2D.rst
- GlobalAveragePooling3D.rst
- GlobalMaxPool1D.rst
- GlobalMaxPool2D.rst
- GlobalMaxPool3D.rst
- Layer.rst
- LayerNormalization.rst
- LeakyReLU.rst
- Maximum.rst
- MaxPool1D.rst
- MaxPool2D.rst
- MaxPool3D.rst
- Minimum.rst
- Multiply.rst
- Permute.rst
- ReLU.rst
- Reshape.rst
- SELU.rst
- Softmax.rst
- Subtract.rst
- UpSampling1D.rst
- UpSampling2D.rst
- UpSampling3D.rst
- ZeroPadding1D.rst
- ZeroPadding2D.rst
- ZeroPadding3D.rst
- binary_crossentropy.rst
- BinaryCrossentropy.rst
- categorical_crossentropy.rst
- CategoricalCrossentropy.rst
- get.rst
- Loss.rst
- mean_absolute_error.rst
- mean_squared_error.rst
- MeanAbsoluteError.rst
- MeanSquaredError.rst
- sparse_categorical_crossentropy.rst
- SparseCategoricalCrossentropy.rst
- Adam.rst
- Optimizer.rst
- RMSprop.rst
- SGD.rst
- get.rst
- L1.rst
- l1_l2.rst
- L1L2.rst
- L2.rst
- Regularizer.rst
- activations.rst
- initializers.rst
- Input.rst
- layers.rst
- losses.rst
- optimizers.rst
- regularizers.rst
- Sequential.rst
- matmul.rst
- abs.rst
- add.rst
- add_n.rst
- argmax.rst
- argmin.rst
- atan2.rst
- ceil.rst
- cos.rst
- cumsum.rst
- divide.rst
- equal.rst
- exp.rst
- floor.rst
- greater.rst
- greater_equal.rst
- is_finite.rst
- is_inf.rst
- is_nan.rst
- l2_normalize.rst
- less.rst
- less_equal.rst
- log.rst
- multiply.rst
- negative.rst
- not_equal.rst
- pow.rst
- reciprocal.rst
- reduce_max.rst
- reduce_mean.rst
- reduce_min.rst
- reduce_sum.rst
- reduce_variance.rst
- round.rst
- rsqrt.rst
- sigmoid.rst
- sign.rst
- sin.rst
- sqrt.rst
- square.rst
- subtract.rst
- tanh.rst
- top_k.rst
- avg_pool.rst
- avg_pool1d.rst
- avg_pool2d.rst
- avg_pool3d.rst
- bias_add.rst
- conv1d.rst
- conv1d_transpose.rst
- conv2d.rst
- conv2d_transpose.rst
- conv3d.rst
- conv3d_transpose.rst
- conv_transpose.rst
- convolution.rst
- depth_to_space.rst
- dropout.rst
- elu.rst
- fused_batch_norm.rst
- gelu.rst
- l2_loss.rst
- leaky_relu.rst
- local_response_normalization.rst
- log_softmax.rst
- max_pool.rst
- max_pool1d.rst
- max_pool2d.rst
- max_pool3d.rst
- moments.rst
- relu.rst
- relu6.rst
- selu.rst
- sigmoid_cross_entropy_with_logits.rst
- silu.rst
- softmax.rst
- softmax_cross_entropy_with_logits.rst
- space_to_depth.rst
- sparse_softmax_cross_entropy_with_logits.rst
- normal.rst
- truncated_normal.rst
- uniform.rst
- argsort.rst
- bitwise.rst
- broadcast_to.rst
- cast.rst
- clip_by_value.rst
- concat.rst
- constant.rst
- convert_to_tensor.rst
- device.rst
- dtypes.rst
- expand_dims.rst
- eye.rst
- fill.rst
- function.rst
- gather.rst
- GradientTape.rst
- identity.rst
- keras.rst
- linalg.rst
- linspace.rst
- math.rst
- Module.rst
- name_scope.rst
- nn.rst
- one_hot.rst
- ones.rst
- ones_like.rst
- pad.rst
- random.rst
- range.rst
- reshape.rst
- reverse.rst
- roll.rst
- shape.rst
- slice.rst
- sort.rst
- split.rst
- squeeze.rst
- TensorShape.rst
- TensorSpec.rst
- tile.rst
- transpose.rst
- unique.rst
- unique_with_counts.rst
- unstack.rst
- zeros.rst
- zeros_like.rst
- BackendRep.rst
- prepare.rst
- run_model.rst
- run_node.rst
- supports_device.rst
- Binding.rst
- Engine.rst
- onnx.rst
- GradScaler.rst
- backward.rst
- cnnl.rst
- cuda.rst
- cudnn.rst
- mps.rst
- openmp.rst
- current_device.rst
- device_count.rst
- get_device_capability.rst
- get_device_name.rst
- is_available.rst
- manual_seed.rst
- manual_seed_all.rst
- memory_allocated.rst
- set_device.rst
- synchronize.rst
- all_gather.rst
- all_gather_into_tensor.rst
- all_reduce.rst
- broadcast.rst
- reduce_scatter.rst
- reduce_scatter_tensor.rst
- trace.rst
- current_device.rst
- device_count.rst
- get_device_capability.rst
- get_device_name.rst
- is_available.rst
- manual_seed.rst
- manual_seed_all.rst
- memory_allocated.rst
- set_device.rst
- synchronize.rst
- current_device.rst
- device_count.rst
- get_device_family.rst
- get_device_name.rst
- is_available.rst
- manual_seed.rst
- manual_seed_all.rst
- memory_allocated.rst
- set_device.rst
- synchronize.rst
- adaptive_avg_pool1d.rst
- adaptive_avg_pool2d.rst
- adaptive_avg_pool3d.rst
- adaptive_max_pool1d.rst
- adaptive_max_pool2d.rst
- adaptive_max_pool3d.rst
- affine.rst
- avg_pool1d.rst
- avg_pool2d.rst
- avg_pool3d.rst
- batch_norm.rst
- binary_cross_entropy_with_logits.rst
- channel_norm.rst
- channel_shuffle.rst
- conv1d.rst
- conv2d.rst
- conv3d.rst
- conv_transpose1d.rst
- conv_transpose2d.rst
- conv_transpose3d.rst
- cosine_similarity.rst
- cross_entropy.rst
- ctc_loss.rst
- drop_block2d.rst
- drop_path.rst
- dropout.rst
- elu.rst
- gelu.rst
- group_norm.rst
- hardsigmoid.rst
- hardswish.rst
- interpolate.rst
- kl_div.rst
- l1_loss.rst
- layer_norm.rst
- leaky_relu.rst
- linear.rst
- local_response_norm.rst
- log_softmax.rst
- max_pool1d.rst
- max_pool2d.rst
- max_pool3d.rst
- mse_loss.rst
- multi_head_attention_forward.rst
- nll_loss.rst
- normalize.rst
- one_hot.rst
- pad.rst
- pixel_shuffle.rst
- pixel_unshuffle.rst
- prelu.rst
- relu.rst
- relu6.rst
- selu.rst
- sigmoid.rst
- sigmoid_focal_loss.rst
- silu.rst
- smooth_l1_loss.rst
- softmax.rst
- sync_batch_norm.rst
- tanh.rst
- unfold.rst
- upsample.rst
- upsample_bilinear.rst
- upsample_nearest.rst
- calculate_gain.rst
- constant_.rst
- dirac_.rst
- eye_.rst
- kaiming_normal_.rst
- kaiming_uniform_.rst
- normal_.rst
- ones_.rst
- trunc_normal_.rst
- uniform_.rst
- xavier_normal_.rst
- xavier_uniform_.rst
- zeros_.rst
- AdaptiveAvgPool1d.rst
- AdaptiveAvgPool2d.rst
- AdaptiveAvgPool3d.rst
- AdaptiveMaxPool1d.rst
- AdaptiveMaxPool2d.rst
- AdaptiveMaxPool3d.rst
- Affine.rst
- AvgPool1d.rst
- AvgPool2d.rst
- AvgPool3d.rst
- BatchNorm1d.rst
- BatchNorm2d.rst
- BatchNorm3d.rst
- BCEWithLogitsLoss.rst
- ChannelShuffle.rst
- ConstantPad1d.rst
- ConstantPad2d.rst
- ConstantPad3d.rst
- Conv1d.rst
- Conv2d.rst
- Conv3d.rst
- ConvTranspose1d.rst
- ConvTranspose2d.rst
- ConvTranspose3d.rst
- CosineSimilarity.rst
- CrossEntropyLoss.rst
- CTCLoss.rst
- DropBlock2d.rst
- Dropout.rst
- DropPath.rst
- ELU.rst
- Flatten.rst
- functional.rst
- GELU.rst
- GroupNorm.rst
- GRU.rst
- GumbelSoftmax.rst
- Hardsigmoid.rst
- Hardswish.rst
- Identity.rst
- init.rst
- KLDivLoss.rst
- L1Loss.rst
- LayerNorm.rst
- LeakyReLU.rst
- Linear.rst
- LocalResponseNorm.rst
- LogSoftmax.rst
- LSTM.rst
- LSTMCell.rst
- MaxPool1d.rst
- MaxPool2d.rst
- MaxPool3d.rst
- Module.rst
- ModuleList.rst
- MSELoss.rst
- MultiheadAttention.rst
- NLLLoss.rst
- Parameter.rst
- PixelShuffle.rst
- PixelUnshuffle.rst
- PReLU.rst
- ReflectionPad1d.rst
- ReflectionPad2d.rst
- ReflectionPad3d.rst
- ReLU.rst
- ReLU6.rst
- ReplicationPad1d.rst
- ReplicationPad2d.rst
- ReplicationPad3d.rst
- RNN.rst
- SELU.rst
- Sequential.rst
- Sigmoid.rst
- SigmoidFocalLoss.rst
- SiLU.rst
- SmoothL1Loss.rst
- Softmax.rst
- SyncBatchNorm.rst
- Tanh.rst
- TransformerDecoder.rst
- TransformerDecoderLayer.rst
- TransformerEncoder.rst
- TransformerEncoderLayer.rst
- Unfold.rst
- Upsample.rst
- UpsamplingBilinear2d.rst
- UpsamplingNearest2d.rst
- ZeroPad2d.rst
- export.rst
- Adam.rst
- AdamW.rst
- LARS.rst
- Optimizer.rst
- RMSprop.rst
- SGD.rst
- ZeroAdamW.rst
- checkpoint.rst
- checkpoint_sequential.rst
- no_checkpoint.rst
- from_dlpack.rst
- to_dlpack.rst
- checkpoint.rst
- dlpack.rst
- abs.rst
- add.rst
- addmm.rst
- amp.rst
- arange.rst
- argmax.rst
- argmin.rst
- argsort.rst
- as_tensor.rst
- atan2.rst
- autograd.rst
- backends.rst
- baddbmm.rst
- bitwise_and.rst
- bitwise_not.rst
- bitwise_or.rst
- bitwise_xor.rst
- bmm.rst
- cat.rst
- ceil.rst
- chunk.rst
- clamp.rst
- cos.rst
- cuda.rst
- cummax.rst
- cummin.rst
- cumsum.rst
- device.rst
- distributed.rst
- div.rst
- dtype.rst
- empty.rst
- enable_grad.rst
- eq.rst
- exp.rst
- eye.rst
- flatten.rst
- flip.rst
- fliplr.rst
- flipud.rst
- floor.rst
- from_numpy.rst
- full.rst
- full_like.rst
- gather.rst
- ge.rst
- gt.rst
- index_select.rst
- inference_mode.rst
- isfinite.rst
- isinf.rst
- isnan.rst
- jit.rst
- le.rst
- linspace.rst
- log.rst
- logical_and.rst
- logical_not.rst
- logical_or.rst
- logical_xor.rst
- logsumexp.rst
- lt.rst
- manual_seed.rst
- masked_select.rst
- matmul.rst
- max.rst
- maximum.rst
- mean.rst
- min.rst
- minimum.rst
- mlu.rst
- mm.rst
- mps.rst
- mul.rst
- multinomial.rst
- nan_to_num.rst
- narrow.rst
- ne.rst
- neg.rst
- nn.rst
- no_grad.rst
- nonzero.rst
- norm.rst
- ones.rst
- ones_like.rst
- onnx.rst
- optim.rst
- permute.rst
- pow.rst
- rand.rst
- randn.rst
- randperm.rst
- reciprocal.rst
- reshape.rst
- roll.rst
- round.rst
- rsqrt.rst
- scatter.rst
- scatter_add.rst
- set_grad_enabled.rst
- sigmoid.rst
- sign.rst
- sin.rst
- Size.rst
- sort.rst
- split.rst
- sqrt.rst
- square.rst
- squeeze.rst
- stack.rst
- sub.rst
- sum.rst
- tanh.rst
- tensor.rst
- Tensor_.rst
- tile.rst
- topk.rst
- transpose.rst
- tril.rst
- triu.rst
- unbind.rst
- unique.rst
- unsqueeze.rst
- var.rst
- var_mean.rst
- where.rst
- zeros.rst
- zeros_like.rst
- roi_align.rst
- roi_pool.rst
- ops.rst
- conf.py
- dali.rst
- dragon.rst
- index.rst
- make.bat
- Makefile
- tensorflow.rst
- tensorrt.rst
- torch.rst
- banner_repository.png
- README.md
- CMakeLists.txt
- common.h
- context.cc
- context.h
- context_cuda.cc
- context_cuda.h
- context_mlu.cc
- context_mlu.h
- context_mps.h
- context_mps.mm
- gradient.cc
- gradient.h
- graph.cc
- graph.h
- graph_optimizer.cc
- graph_optimizer.h
- memory.cc
- memory.h
- operator.cc
- operator.h
- operator_schema.cc
- operator_schema.h
- registry.h
- tensor.cc
- tensor.h
- typeid.h
- types.h
- workspace.cc
- workspace.h
- dropblock_kernel.cc
- dropout_kernel.cc
- elu_kernel.cc
- gelu_kernel.cc
- prelu_kernel.cc
- relu_kernel.cc
- selu_kernel.cc
- sigmoid_kernel.cc
- softmax_kernel.cc
- swish_kernel.cc
- tanh_kernel.cc
- dropblock_kernel.cu
- dropout_kernel.cu
- elu_kernel.cu
- gelu_kernel.cu
- prelu_kernel.cu
- relu_kernel.cu
- selu_kernel.cu
- sigmoid_kernel.cu
- softmax_kernel.cu
- swish_kernel.cu
- tanh_kernel.cu
- dropout_kernel.mlu
- relu_kernel.mlu
- sigmoid_kernel.mlu
- swish_kernel.mlu
- dropout_kernel.mm
- elu_kernel.mm
- gelu_kernel.mm
- relu_kernel.mm
- sigmoid_kernel.mm
- softmax_kernel.mm
- swish_kernel.mm
- tanh_kernel.mm
- CMakeLists.txt
- op_kernels.h
- assign_kernel.cc
- boolean_kernel.cc
- gather_kernel.cc
- initialize_kernel.cc
- onehot_kernel.cc
- pad_kernel.cc
- repeat_kernel.cc
- reverse_kernel.cc
- roll_kernel.cc
- scatter_kernel.cc
- slice_kernel.cc
- tile_kernel.cc
- unique_kernel.cc
- assign_kernel.cu
- boolean_kernel.cu
- gather_kernel.cu
- initialize_kernel.cu
- onehot_kernel.cu
- pad_kernel.cu
- repeat_kernel.cu
- reverse_kernel.cu
- roll_kernel.cu
- scatter_kernel.cu
- slice_kernel.cu
- tile_kernel.cu
- unique_kernel.cu
- assign_kernel.mm
- initialize_kernel.mm
- slice_kernel.mm
- CMakeLists.txt
- op_kernels.h
- cross_entropy_kernel.cc
- focal_loss_kernel.cc
- generic_loss_kernel.cc
- l1_loss_kernel.cc
- nll_loss_kernel.cc
- cross_entropy_kernel.cu
- focal_loss_kernel.cu
- generic_loss_kernel.cu
- l1_loss_kernel.cu
- nll_loss_kernel.cu
- cross_entropy_kernel.mlu
- generic_loss_kernel.mlu
- l1_loss_kernel.mlu
- cross_entroy_kernel.mm
- focal_loss_kernel.mm
- generic_loss_kernel.mm
- l1_loss_kernel.mm
- nll_loss_kernel.mm
- CMakeLists.txt
- op_kernels.h
- arg_kernel.cc
- clip_kernel.cc
- cum_kernel.cc
- elementwise_kernel.cc
- moments_kernel.cc
- nonzero_kernel.cc
- reduce_kernel.cc
- topk_kernel.cc
- arg_kernel.cu
- clip_kernel.cu
- cum_kernel.cu
- eltmentwise_kernel.cu
- moments_kernel.cu
- nonzero_kernel.cu
- reduce_kernel.cu
- topk_kernel.cu
- clip_kernel.mlu
- elementwise_kernel.mlu
- clip_kernel.mm
- elementwise_kernel.mm
- moments_kernel.mm
- reduce_kernel.mm
- CMakeLists.txt
- op_kernels.h
- batch_norm_kernel.cc
- channel_norm_kernel.cc
- group_norm_kernel.cc
- layer_norm_kernel.cc
- lp_norm_kernel.cc
- batch_norm_kernel.cu
- channel_norm_kernel.cu
- group_norm_kernel.cu
- layer_norm_kernel.cu
- lp_norm_kernel.cu
- batch_norm_kernel.mm
- channel_norm_kernel.mm
- group_norm_kernel.mm
- layer_norm_kernel.mm
- lp_norm_kerenl.mm
- CMakeLists.txt
- op_kernels.h
- lstm_kernel.cc
- lstm_kernel.cu
- CMakeLists.txt
- op_kernels.h
- adam_kernel.cc
- gradient_kernel.cc
- rmsprop_kernel.cc
- sgd_kernel.cc
- adam_kernel.cu
- gradient_kernel.cu
- rmsprop_kernel.cu
- sgd_kernel.cu
- adam_kernel.mlu
- gradient_kernel.mlu
- rmsprop_kernel.mlu
- sgd_kernel.mlu
- adam_kernel.mm
- rmsprop_kernel.mm
- sgd_kernel.mm
- CMakeLists.txt
- op_kernels.h
- avgpool_kernel.cc
- bias_add_kernel.cc
- conv_kernel.cc
- maxpool_kernel.cc
- resize_linear_kernel.cc
- resize_nearest_kernel.cc
- roi_align_kernel.cc
- roi_pool_kernel.cc
- avgpool_kernel.cu
- bias_add_kernel.cu
- conv_kernel.cu
- maxpool_kernel.cu
- resize_linear_kernel.cu
- resize_nearest_kernel.cu
- roi_align_kernel.cu
- roi_pool_kernel.cu
- bias_add_kernel.mm
- roi_align_kernel.mm
- CMakeLists.txt
- op_kernels.h
- CMakeLists.txt
- op_kernels.h
- CMakeLists.txt
- common.h
- cuda.h
- dlpack.h
- gradient.h
- mlu.h
- module.cc
- mpi.h
- mps.h
- numpy.h
- plugin_op.cc
- plugin_op.h
- proto.h
- sysconfig.h
- tensor.h
- types.h
- CMakeLists.txt
- device.cc
- dragon_runtime.h
- module.cc
- CMakeLists.txt
- onnx_attibute.cc
- onnx_backend.cc
- onnx_backend.h
- onnx_importer.cc
- onnx_initializer.cc
- CMakeLists.txt
- dropblock_op.cc
- dropblock_op.h
- dropout_op.cc
- dropout_op.h
- dropout_op.mm
- dropout_op_cudnn.cc
- droppath_op.cc
- droppath_op.h
- droppath_op.mm
- elu_op.cc
- elu_op.h
- elu_op_cnnl.cc
- elu_op_cudnn.cc
- gelu_op.cc
- gelu_op.h
- gelu_op_cnnl.cc
- hardsigmoid_op.cc
- hardsigmoid_op.h
- hardsigmoid_op_cnnl.cc
- hardswish_op.cc
- hardswish_op.h
- hardswish_op_cnnl.cc
- log_softmax_op.cc
- log_softmax_op.h
- log_softmax_op_cnnl.cc
- prelu_op.cc
- prelu_op.h
- relu_op.cc
- relu_op.h
- relu_op_cnnl.cc
- relu_op_cudnn.cc
- selu_op.cc
- selu_op.h
- sigmoid_op.cc
- sigmoid_op.h
- sigmoid_op_cnnl.cc
- sigmoid_op_cudnn.cc
- silu_op.cc
- silu_op.h
- silu_op_cnnl.cc
- softmax_op.cc
- softmax_op.h
- softmax_op_cnnl.cc
- softmax_op_cudnn.cc
- tanh_op.cc
- tanh_op.h
- tanh_op_cnnl.cc
- tanh_op_cudnn.cc
- assign_op.cc
- assign_op.h
- assign_op_cnnl.cc
- boolean_mask_op.cc
- boolean_mask_op.h
- CMakeLists.txt
- concat_op.cc
- concat_op.h
- expand_op.cc
- expand_op.h
- expand_op_cnnl.cc
- gather_elements_op.cc
- gather_elements_op.mm
- gather_elements_op_cnnl.cc
- gather_op.cc
- gather_op.h
- gather_op.mm
- gather_op_cnnl.cc
- initialize_op.cc
- initialize_op.h
- initialize_op.mm
- initialize_op_cnnl.cc
- multinomial_op.cc
- multinomial_op.h
- one_hot_op.cc
- one_hot_op.h
- one_hot_op.mm
- one_hot_op_cnnl.cc
- pad_op.cc
- pad_op.h
- pad_op.mm
- pad_op_cnnl.cc
- repeat_op.cc
- repeat_op.h
- reshape_op.cc
- reshape_op.h
- reverse_op.cc
- reverse_op.h
- reverse_op.mm
- reverse_op_cnnl.cc
- roll_op.cc
- roll_op.h
- roll_op_cnnl.cc
- scatter_add_op.cc
- scatter_add_op.mm
- scatter_elements_op.cc
- scatter_elements_op.mm
- scatter_elements_op_cnnl.cc
- scatter_op.h
- shape_op.cc
- shape_op.h
- shuffle_op.cc
- shuffle_op.h
- shuffle_op_cnnl.cc
- slice_op.cc
- slice_op.h
- slice_op_cnnl.cc
- sort_op.cc
- sort_op.h
- split_op.cc
- split_op.h
- stack_op.cc
- stack_op.h
- tile_op.cc
- tile_op.h
- tile_op.mm
- tile_op_cnnl.cc
- transpose_op.cc
- transpose_op.h
- transpose_op_cnnl.cc
- transpose_op_impl_cnnl.h
- trilu_op.cc
- trilu_op.h
- unique_op.cc
- unique_op.h
- CMakeLists.txt
- collective_op.cc
- collective_op.h
- collective_op_impl.h
- collective_op_impl_cncl.cc
- collective_op_impl_cncl.h
- collective_op_impl_mpi.cc
- collective_op_impl_mpi.h
- collective_op_impl_nccl.cc
- collective_op_impl_nccl.h
- checkpoint_op.cc
- CMakeLists.txt
- gradient_op.cc
- gradient_op.h
- CMakeLists.txt
- cross_entropy_loss_op.h
- cross_entropy_sigmoid_loss_op.cc
- cross_entropy_sigmoid_loss_op_cnnl.cc
- cross_entropy_softmax_loss_op.cc
- cross_entropy_softmax_loss_op_cnnl.cc
- ctc_loss_op.cc
- ctc_loss_op.h
- ctc_loss_op_cudnn.cc
- focal_loss_op.h
- focal_sigmoid_loss_op.cc
- l1_loss_op.cc
- l1_loss_op.h
- l1_loss_op_cnnl.cc
- l1_smooth_loss_op.cc
- l1_smooth_loss_op_cnnl.cc
- l2_loss_op.cc
- l2_loss_op.h
- l2_loss_op_cnnl.cc
- nll_loss_op.cc
- nll_loss_op.h
- nll_loss_op_cnnl.cc
- affine_op.cc
- affine_op.h
- affine_op_cnnl.cc
- arg_op.cc
- arg_op.h
- arg_op.mm
- arg_op_cnnl.cc
- cast_op.cc
- cast_op.h
- clip_op.cc
- clip_op.h
- CMakeLists.txt
- cum_op.cc
- cum_op.h
- cum_op_cnnl.cc
- elementwise_gradient_cnnl.cc
- elementwise_gradient_impl.h
- elementwise_gradient_op.cc
- elementwise_op.cc
- elementwise_op.h
- elementwise_op_cnnl.cc
- gemm_op.cc
- gemm_op.h
- gemm_op.mm
- gemm_op_cnnl.cc
- gemm_op_cuda.cc
- gemm_op_impl_cnnl.h
- gemm_op_impl_cuda.h
- matmul_op.cc
- matmul_op.h
- matmul_op.mm
- matmul_op_cnnl.cc
- moments_op.cc
- moments_op.h
- nonzero_op.cc
- nonzero_op.h
- pow_op.cc
- reduce_op.cc
- reduce_op.h
- reduce_op_cnnl.cc
- reduce_op_impl_cnnl.h
- reduce_op_impl_cudnn.h
- topk_op.cc
- topk_op.h
- topk_op.mm
- topk_op_cnnl.cc
- where_op.cc
- batch_norm_op.cc
- batch_norm_op.h
- batch_norm_op_cnnl.cc
- batch_norm_op_cudnn.cc
- batch_norm_op_sync.cc
- channel_norm_op.cc
- channel_norm_op.h
- CMakeLists.txt
- group_norm_op.cc
- group_norm_op.h
- group_norm_op_cnnl.cc
- layer_norm_op.cc
- layer_norm_op.h
- layer_norm_op_cnnl.cc
- lp_norm_op.cc
- lp_norm_op.h
- lp_norm_op_cnnl.cc
- lrn_op.cc
- lrn_op.h
- lrn_op_cudnn.cc
- CMakeLists.txt
- embedding_op.cc
- embedding_op.h
- embedding_op_cnnl.cc
- lstm_cell_op.cc
- lstm_cell_op.h
- mha_op.cc
- mha_op.h
- mha_op_base.h
- mha_op_cnnl.cc
- mha_op_cudnn.cc
- rnn_op.cc
- rnn_op.h
- rnn_op_base.h
- rnn_op_cudnn.cc
- rnn_param_op.cc
- rnn_param_op.h
- adam_op.cc
- CMakeLists.txt
- rmsprop_op.cc
- sgd_op.cc
- update_op.h
- update_op_base.cc
- zero_update_op.h
- zero_update_op_base.cc
- bias_add_op.cc
- bias_add_op.h
- bias_add_op_cudnn.cc
- CMakeLists.txt
- conv_op.cc
- conv_op.h
- conv_op.mm
- conv_op_algo.h
- conv_op_base.cc
- conv_op_base.h
- conv_op_base.mm
- conv_op_cnnl.cc
- conv_op_cudnn.cc
- conv_op_impl.h
- conv_op_impl_cnnl.h
- conv_op_impl_cudnn.h
- conv_transpose_op.cc
- conv_transpose_op.h
- conv_transpose_op.mm
- conv_transpose_op_cnnl.cc
- conv_transpose_op_cudnn.cc
- im2col_op.cc
- im2col_op.h
- pool_op.cc
- pool_op.h
- pool_op.mm
- pool_op_base.cc
- pool_op_base.h
- pool_op_base.mm
- pool_op_cnnl.cc
- pool_op_cudnn.cc
- resize_op.cc
- resize_op.h
- resize_op.mm
- resize_op_base.h
- resize_op_cnnl.cc
- roi_align_op.cc
- roi_align_op.h
- roi_align_op_cnnl.cc
- roi_pool_op.cc
- roi_pool_op.h
- space_to_depth_op.cc
- space_to_depth_op.h
- space_to_depth_op_cnnl.cc
- CMakeLists.txt
- CMakeLists.txt
- dragon.proto
- onnx.proto
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- backprop.py
- config.py
- context.py
- function_lib.py
- graph_lib.py
- op_args.py
- op_lib.py
- op_schema.py
- op_spec.py
- __init__.py
- cuda.py
- mlu.py
- mps.py
- __init__.py
- backend.py
- __init__.py
- backend.py
- config.py
- context.py
- device_spec.py
- dlpack.py
- mapping.py
- proto_util.py
- sysconfig.py
- tapes.py
- tensor.py
- types.py
- workspace.py
- __init__.py
- activation_ops.py
- array_ops.py
- constant_ops.py
- distributed_ops.py
- framework_ops.py
- init_ops.py
- loss_ops.py
- math_ops.py
- normalization_ops.py
- random_ops.py
- rnn_ops.py
- sort_ops.py
- tensor_ops.py
- vision_ops.py
- __init__.py
- dragon.proto
- __init__.py
- common_utils.py
- __init__.py
- __init__.py
- adam.py
- optimizer.py
- rmsprop.py
- sgd.py
- __init__.py
- decorator.py
- deprecation.py
- inspect.py
- logging.py
- nest.py
- registry.py
- serialization.py
- string.py
- tls.py
- __init__.py
- __init__.py
- cpp_extension.py
- tensorboard.py
- __init__.py
- native.py
- tensorrt.py
- __init__.py
- activation.py
- array.py
- math.py
- normalization.py
- utils.py
- vision.py
- __init__.py
- native.py
- torch.py
- __init__.py
- helper.py
- __init__.py
- __init__.py
- __init__.py
- CMakeLists.txt
- common_cub.h
- common_cuda.h
- common_cudnn.cc
- common_cudnn.h
- common_dlpack.h
- common_eigen.h
- common_mlu.cc
- common_mlu.h
- common_mpi.h
- common_mps.h
- common_mps.mm
- common_nccl.h
- common_openmp.h
- common_thrust.h
- blas.cc
- broadcast.cc
- cast.cc
- elementwise.cc
- random.cc
- reduce.cc
- transform.cc
- transpose.cc
- blas.cu
- broadcast.cu
- cast.cu
- elementwise.cu
- random.cu
- reduce.cu
- transform.cu
- transpose.cu
- blas.mlu
- broadcast.mlu
- cast.mlu
- elementwise.mlu
- random.mlu
- transform.mlu
- blas.mm
- broadcast.mm
- cast.mm
- elementwise.mm
- reduce.mm
- transform.mm
- transpose.mm
- atomics.h
- blas.h
- broadcast.h
- CMakeLists.txt
- elementwise.h
- functional.h
- random.h
- reduce.h
- transform.h
- transpose.h
- types.h
- utils.h
- CMakeLists.txt
- conversions.h
- logging.cc
- logging.h
- macros.h
- map_utils.h
- math_functions.h
- proto_utils.cc
- proto_utils.h
- string_utils.h
- thread.h
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- base_layer.py
- input_layer.py
- input_spec.py
- sequential.py
- __init__.py
- activation.py
- convolutional.py
- core.py
- merging.py
- normalization.py
- pooling.py
- reshaping.py
- __init__.py
- adam.py
- gradient_descent.py
- optimizer.py
- rmsprop.py
- __init__.py
- pickle_format.py
- __init__.py
- conv_utils.py
- generic_utils.py
- losses_utils.py
- __init__.py
- activations.py
- initializers.py
- losses.py
- regularizers.py
- __init__.py
- build_linux.sh
- build_macos.sh
- build_windows_vc142.bat
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- backprop.py
- def_function.py
- __init__.py
- constant_op.py
- dtypes.py
- ops.py
- tensor_shape.py
- tensor_spec.py
- __init__.py
- module.py
- __init__.py
- array_ops.py
- bitwise_ops.py
- clip_ops.py
- init_ops.py
- linalg_ops.py
- math_ops.py
- nn_ops.py
- random_ops.py
- sort_ops.py
- variables.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- engine.py
- __init__.py
- test_autograph.py
- test_device.py
- test_distributed.py
- test_framework.py
- test_ops.py
- test_util.py
- test_engine.py
- test_layers.py
- test_nn.py
- test_optimizer.py
- test_eager.py
- test_framework.py
- test_module.py
- test_ops.py
- test_amp.py
- test_autograd.py
- test_backends.py
- test_jit.py
- test_nn.py
- test_ops.py
- test_optim.py
- test_torch.py
- run_test.py
- build.sh
- README.md
- README.md
- cub
- eigen
- pybind11
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- grad_scaler.py
- __init__.py
- function.py
- functional.py
- grad_mode.py
- __init__.py
- cnnl.py
- cuda.py
- cudnn.py
- mlu.py
- mps.py
- openmp.py
- __init__.py
- device.py
- graphs.py
- memory.py
- random.py
- __init__.py
- trace.py
- __init__.py
- activation.py
- batchnorm.py
- channelshuffle.py
- container.py
- conv.py
- distance.py
- dropout.py
- flatten.py
- fold.py
- linear.py
- loss.py
- module.py
- normalization.py
- padding.py
- pixelshuffle.py
- pooling.py
- rnn.py
- sparse.py
- transformer.py
- upsampling.py
- utils.py
- __init__.py
- functional.py
- init.py
- parameter.py
- __init__.py
- array_ops.py
- constant_ops.py
- distributed_ops.py
- math_ops.py
- random_ops.py
- sort_ops.py
- tensor_ops.py
- __init__.py
- adam.py
- optimizer.py
- rmsprop.py
- sgd.py
- zero.py
- __init__.py
- checkpoint.py
- dlpack.py
- hooks.py
- __init__.py
- cpp.py
- random.py
- serialization.py
- tensor.py
- __init__.py
- __init__.py
- __init__.py
- __init__.py
- roi_align.py
- roi_pool.py
- __init__.py
- __init__.py
- .clang-format
- .flake8
- .gitignore
- .gitmodules
- CMakeLists.txt
- LICENSE
- pyproject.toml
- README.md
- requirements.txt
- setup.py
- version.txt
# Installation Guide
1. Get the code
git clone https://github.com/seetaresearch/dragon
Downloads the entire project code from GitHub to your computer.
cd dragon
Moves into the project folder you just downloaded.
2. Docker
Easy RecommendedPrerequisites
- Git Needed to download the project code from GitHub.
- Docker Desktop Needed to build and run containers. Install it and keep it running in the background.
β οΈ This is a large repository, so this method may point to an internal sub-package rather than the actual core product. Check the full README as well.
docker build -f docker/ubuntu-20.04-cpu/Dockerfile -t dragon .
Builds a runnable image based on the Dockerfile.
docker run -p 8080:80 dragon
Runs the built image as an actual container.
Run docker compose ps to check the containers are Up. If the README mentions a port, open http://localhost:PORT in your browser.
3. CMake
MediumPrerequisites
mkdir build && cd build
Creates a folder to hold the build output and moves into it.
cmake ..
Analyzes the source code and generates build configuration files (must be run inside the build folder).
make
Compiles the code based on the generated build configuration to produce an executable.
Check that an executable was created inside the build folder, then run it directly (e.g. ./build/app_name).
4. Python
EasyPrerequisites
pip install -r requirements.txt
Installs the Python libraries listed in requirements.txt (or similar).
python <μ€νν νμΌλͺ
>.py # READMEμμ μ νν μ€ν νμΌλͺ
μ νμΈνμΈμ
Runs the Python script (or module).
If it runs without errors and prints output in the terminal, it worked.
// repository documentation
Was this content helpful?
(0 ratings)