Ahmed Taei 25fd005dd9 Initial implementation of Blockwise Model Update Filtering (BMUF) ...

Summary:
A Single machine multi-GPU version of BMUF algorithm. BMUF is a modification to
model averaging where updates to global model is implemented as a filter:
param_t = param_(t-1) + delta
delta = \beta delta_(t-1) + \alpha average(param_t) - param_(t-1)

Reviewed By: akyrola

Differential Revision: D4995057

fbshipit-source-id: 48176ba66d67eaf3fa4dee16d50d9589825ddba4

2017-05-15 18:18:15 -07:00

..

docs

fixed document generator for github

2017-05-15 11:40:46 -07:00

examples

Add model saving and loading to resnet50_trainer.py

2017-05-05 10:08:37 -07:00

helpers

New cudnn ops

2017-05-08 16:33:21 -07:00

layers

Support un-equal weight training for mtml models

2017-05-15 12:56:11 -07:00

mint

doxygen python block added

2017-03-29 06:46:16 -07:00

mkl

MKL related files with review comments incorporated

2017-04-25 00:31:29 -07:00

models

RNNCell API change: returns states and outputs

2017-05-08 15:19:48 -07:00

operator_test

L1 Operator

2017-05-11 18:03:10 -07:00

predictor

Re-apply #266

2017-04-25 21:17:04 -07:00

_import_c_extension.py

doxygen python block added

2017-03-29 06:46:16 -07:00

attention.py

Drop extra Reshape in attention calculation

2017-05-09 17:16:36 -07:00

brew_test.py

arg scope in ModelHelper

2017-05-12 11:18:59 -07:00

brew.py

arg scope in ModelHelper

2017-05-12 11:18:59 -07:00

caffe_translator_test.py

Allow test discovery in caffe2/python/

2017-03-14 18:16:41 -07:00

caffe_translator.py

adding 3d operator translators

2017-05-11 23:01:44 -07:00

checkpoint_test.py

Adds interfaces to check the existence of a DB

2017-04-11 14:07:49 -07:00

checkpoint.py

Adds interfaces to check the existence of a DB

2017-04-11 14:07:49 -07:00

CMakeLists.txt

…

cnn.py

arg scope in ModelHelper

2017-05-12 11:18:59 -07:00

context_test.py

…

context.py

doxygen python block added

2017-03-29 06:46:16 -07:00

control_test.py

…

control.py

doxygen python block added

2017-03-29 06:46:16 -07:00

convnet_benchmarks_test.py

…

convnet_benchmarks.py

doxygen python block added

2017-03-29 06:46:16 -07:00

core_gradients_test.py

Fix backward pass computation when an input is used in a Fill-op input for shape

2017-04-11 19:32:22 -07:00

core_test.py

fix enternalinputs handling in AppendNet v2

2017-05-05 21:50:57 -07:00

core.py

Relax requirement on token uniqueness

2017-05-09 19:36:00 -07:00

crf.py

Integrate CRF in DeepText + New caffe2 operator for viterbi decode

2017-05-01 20:39:41 -07:00

data_parallel_model_test.py

Initial implementation of Blockwise Model Update Filtering (BMUF)

2017-05-15 18:18:15 -07:00

data_parallel_model.py

Initial implementation of Blockwise Model Update Filtering (BMUF)

2017-05-15 18:18:15 -07:00

data_workers_test.py

Add direct enqueuing to enable RNN input, allow specify batch columns

2017-05-03 14:49:44 -07:00

data_workers.py

Add direct enqueuing to enable RNN input, allow specify batch columns

2017-05-03 14:49:44 -07:00

dataio_test.py

…

dataio.py

doxygen python block added

2017-03-29 06:46:16 -07:00

dataset.py

doxygen python block added

2017-03-29 06:46:16 -07:00

db_test.py

…

device_checker.py

doxygen python block added

2017-03-29 06:46:16 -07:00

dyndep.py

doxygen python block added

2017-03-29 06:46:16 -07:00

empty.so

Adding video data layer for caffe2

2017-05-05 14:16:38 -07:00

experiment_util.py

Port Xray OC workflow to elastic_data_parallel_model

2017-05-01 00:32:47 -07:00

extension_loader.py

Make extension loader properly handle visibility.

2017-03-30 14:38:38 -07:00

gradient_check_test.py

gradient checker for nets

2017-03-28 13:03:14 -07:00

gradient_checker.py

MultiRNNCell (Caffe2)

2017-05-03 10:02:31 -07:00

hsm_util.py

doxygen python block added

2017-03-29 06:46:16 -07:00

hypothesis_test_util.py

Add option to control the size of lengths tensor

2017-04-20 09:53:22 -07:00

hypothesis_test.py

cuda unique op

2017-05-11 21:08:10 -07:00

layer_model_helper.py

gradient impact of task layers on shared is configurable

2017-05-11 20:34:04 -07:00

layer_model_instantiator.py

JoinContext

2017-05-02 17:32:26 -07:00

layer_test_util.py

Re-apply #266

2017-04-25 21:17:04 -07:00

layers_test.py

LastNWindowCollector

2017-05-04 17:32:09 -07:00

load_save_test.py

Allow Load operator to load into overriden names

2017-04-27 01:18:12 -07:00

lstm_benchmark.py

lstm_benchmark use rnn_cell.LSTM multicell + assertion

2017-05-04 17:02:32 -07:00

memonger_test.py

share forward activation blobs + pass unused free blobs down all branches + use shape infernece

2017-04-25 14:23:25 -07:00

memonger.py

share forward activation blobs + pass unused free blobs down all branches + use shape infernece

2017-04-25 14:23:25 -07:00

mkl_test_util.py

doxygen python block added

2017-03-29 06:46:16 -07:00

model_device_test.py

…

model_helper.py

arg scope in ModelHelper

2017-05-12 11:18:59 -07:00

mpi_python.cc

Fix pybind11 module name for MPI helpers

2017-05-02 23:18:50 -07:00

muji_test.py

…

muji.py

doxygen python block added

2017-03-29 06:46:16 -07:00

net_builder_test.py

Allow test discovery in caffe2/python/

2017-03-14 18:16:41 -07:00

net_builder.py

doxygen python block added

2017-03-29 06:46:16 -07:00

net_drawer.py

net_drawer: --input is required

2017-05-04 11:45:57 -07:00

net_printer_test.py

…

net_printer.py

doxygen python block added

2017-03-29 06:46:16 -07:00

optimizer_test_util.py

create_net: explicitly specify if one wants to overwrite the network.

2017-04-17 21:46:53 -07:00

optimizer_test.py

Returns auxiliary parameters in the optimizers.

2017-04-17 10:16:32 -07:00

optimizer.py

Refactor Optimizer to Allow scale_learning_rate

2017-05-09 13:16:21 -07:00

pipeline.py

doxygen python block added

2017-03-29 06:46:16 -07:00

predictor_constants.py

Re-apply #266

2017-04-25 21:17:04 -07:00

pybind_state_gpu.cc

…

pybind_state_mkl.cc

…

pybind_state.cc

Relax requirement on token uniqueness

2017-05-09 19:36:00 -07:00

pybind_state.h

bugfix for Windows, esp. VS 2017

2017-03-21 05:17:59 -07:00

python_op_test.py

…

queue_util.py

doxygen python block added

2017-03-29 06:46:16 -07:00

record_queue.py

doxygen python block added

2017-03-29 06:46:16 -07:00

recurrent.py

fix recompute_blobs_on_backward

2017-05-05 14:01:34 -07:00

rnn_cell.py

Optimize memory usage for MI-LSTM

2017-05-10 16:53:43 -07:00

schema_test.py

Prevent false overwriting of a field

2017-05-02 23:18:49 -07:00

schema.py

Prevent false overwriting of a field

2017-05-02 23:18:49 -07:00

scope_test.py

Fix corruption of NameScope when exception is thrown

2017-04-24 22:46:27 -07:00

scope.py

Fix corruption of NameScope when exception is thrown

2017-04-24 22:46:27 -07:00

session_test.py

Warn on setting blob on Scalar

2017-05-01 20:18:30 -07:00

session.py

doxygen python block added

2017-03-29 06:46:16 -07:00

sparse_to_dense_mask_test.py

…

task.py

doxygen python block added

2017-03-29 06:46:16 -07:00

test_util.py

doxygen python block added

2017-03-29 06:46:16 -07:00

text_file_reader.py

doxygen python block added

2017-03-29 06:46:16 -07:00

timeout_guard.py

doxygen python block added

2017-03-29 06:46:16 -07:00

toy_regression_test.py

…

tt_core_test.py

…

tt_core.py

doxygen python block added

2017-03-29 06:46:16 -07:00

utils.py

Free scratch blobs when data workers exits, add utility function to reset blobs

2017-04-26 13:40:13 -07:00

visualize.py

doxygen python block added

2017-03-29 06:46:16 -07:00

workspace_test.py

create_net: explicitly specify if one wants to overwrite the network.

2017-04-17 21:46:53 -07:00

workspace.py

create_net: explicitly specify if one wants to overwrite the network.

2017-04-17 21:46:53 -07:00