Skip to content

SeetaResearch / Dragon

Go to a project
Commit 5d518b6c by Ting PAN
Options

io refactoring

1 parent 31e02b2b
Inline Side-by-side
Showing with 429 additions and 356 deletions
Dragon/CMakeLists.txt
......@@ -24,7 +24,7 @@ set(3RDPARTY_DIR ${PROJECT_SOURCE_DIR}/../3rdparty)
set(PYTHON_DIR /usr/include/python2.7) # prefer
#set(PYTHON_DIR /usr/include/python3.x) # optional, set specific version
#set(ANACONDA_DIR /xxx/anaconda) # optional, set specific version below if using py3
set(NUMPY_DIR /xxx/numpy) # require， root folder of numpy package
set(NUMPY_DIR /xxx/numpy) # require, root folder of numpy package
# set CUDA compiling architecture
set(CUDA_ARCH -gencode arch=compute_20,code=sm_20
......
Dragon/include/core/context.h
......@@ -52,7 +52,7 @@ class CPUContext{
inline static void Delete(void* data) { free(data); }
template<typename T, class DstContext, class SrcContext>
inline static void Copy(int n, T* dst, const T* src){
inline static void Copy(int n, T* dst, const T* src) {
if (dst == src) return;
// only the basic types(e.g. int/float) can memcpy correctly
if (std::is_fundamental<T>::value)
......
Dragon/include/core/context_cuda.h
......@@ -119,7 +119,7 @@ class CUDAContext {
inline static void Delete(void* data) { cudaFree(data); }
template<typename T, class DstContext, class SrcContext>
static void Copy(int n, T* dst, const T* src){
static void Copy(int n, T* dst, const T* src) {
if (dst == src) return;
Memcpy<SrcContext, DstContext>(n * sizeof(T), (void*)dst, (const void*)src);
}
......@@ -148,7 +148,7 @@ class CUDAContext {
}
#ifdef WITH_CUDNN
cudnnHandle_t cudnn_handle(){
cudnnHandle_t cudnn_handle() {
auto& handle = cuda_object_.cudnn_handle[gpu_id_];
if (handle) {
return handle;
......
Dragon/include/core/tensor.h
......@@ -77,7 +77,7 @@ class Tensor {
inline TIndex offset(const vector<TIndex>& vec) {
CHECK_LE(vec.size(), ndim());
TIndex offset = 0;
for (int i = 0; i < ndim(); i++){
for (int i = 0; i < ndim(); i++) {
offset = offset * dim(i);
if (vec.size() > i) offset += vec[i];
}
......@@ -130,7 +130,7 @@ class Tensor {
}
template <class Context>
void* raw_mutable_data(const TypeMeta& meta){
void* raw_mutable_data(const TypeMeta& meta) {
void* data_ptr;
active_data_ptr<Context>(&data_ptr);
if (meta_ == meta && data_ptr) {
......
Dragon/include/core/typeid.h
......@@ -75,20 +75,20 @@ class TypeMeta {
bool Match() const { return (id_ == Id<T>()); }
template <typename T>
static void Ctor(void* ptr, size_t n){
static void Ctor(void* ptr, size_t n) {
T* typed_ptr = static_cast<T*>(ptr);
for (unsigned int i = 0; i < n; i++) new(typed_ptr + i) T;
}
template <typename T>
static void Copy(const void* src, void* dst, size_t n){
static void Copy(const void* src, void* dst, size_t n) {
const T* typed_src = static_cast<const T*>(src);
T* typed_dst = static_cast<T*>(dst);
for (unsigned int i = 0; i < n; i++) typed_dst[i] = typed_src[i];
}
template <typename T>
static void Dtor(void* ptr, size_t n){
static void Dtor(void* ptr, size_t n) {
T* typed_ptr = static_cast<T*>(ptr);
for (unsigned int i = 0; i < n; i++) typed_ptr[i].~T();
}
......
Dragon/include/core/workspace.h
......@@ -44,7 +44,7 @@ class Workspace{
return tensor_map_.count(query) > 0;
}
inline Tensor* CreateTensor(const string& name){
inline Tensor* CreateTensor(const string& name) {
string query = GetTensorName(name);
if (!HasTensor(query))
tensor_map_[query] = unique_ptr<Tensor>(new Tensor(query));
......@@ -143,7 +143,7 @@ class Workspace{
return graph_map_[graph_name]->Run(include, exclude);
}
inline vector<string> GetGraphs(){
inline vector<string> GetGraphs() {
vector<string> names;
for (auto& it : graph_map_) names.push_back(it.first);
return names;
......
Dragon/include/operators/common/utils_op.h
......@@ -25,9 +25,9 @@ class AccuracyOp final: public Operator<Context> {
public:
AccuracyOp(const OperatorDef& op_def, Workspace* ws)
: Operator<Context>(op_def, ws),
top_k(OperatorBase::GetSingleArg<int>("top_k", 1)){
top_k(OperatorBase::GetSingleArg<int>("top_k", 1)) {
vector<int> args = OperatorBase::GetRepeatedArg<int>("ignore_labels");
if (args.size()){
if (args.size()) {
ignore_labels.Reshape(vector<TIndex>(1, args.size()));
int* ignore_data = ignore_labels.mutable_data<int, CPUContext>();
for (int i = 0; i < args.size(); i++) ignore_data[i] = args[i];
......
Dragon/include/operators/vision/roi_align_op.h
......@@ -39,7 +39,7 @@ class ROIAlignGradientOp : public Operator<Context> {
: Operator<Context>(op_def, ws),
pool_h(OperatorBase::GetSingleArg<int>("pool_h", 0)),
pool_w(OperatorBase::GetSingleArg<int>("pool_w", 0)),
spatial_scale(OperatorBase::GetSingleArg<float>("spatial_scale", 1.0)){
spatial_scale(OperatorBase::GetSingleArg<float>("spatial_scale", 1.0)) {
CHECK_GT(pool_h, 0) << "\npool_h must > 0";
CHECK_GT(pool_w, 0) << "\npool_w must > 0";
}
......
Dragon/include/utils/caffemodel.h
......@@ -54,7 +54,7 @@ inline void LoadCaffeModel(string file, string scope, Workspace* ws) {
ReadProtoFromBinaryFile(file.c_str(), &net_param);
LOG(INFO) << "Restore From Model @: " << file << "......";
LOG(INFO) << "Model Format: CaffeModel";
for (int i = 0; i < net_param.layer_size(); i++){
for (int i = 0; i < net_param.layer_size(); i++) {
const LayerParameter& layer = net_param.layer(i);
const string& layer_name = layer.name();
string prefix = scope + layer_name + "@param";
......
Dragon/include/utils/string.h
......@@ -22,7 +22,7 @@ inline std::vector<std::string> SplitString(const std::string& str,
std::vector<std::string> ret;
std::string temp(str);
size_t pos;
while (pos = temp.find(c), pos != std::string::npos){
while (pos = temp.find(c), pos != std::string::npos) {
ret.push_back(temp.substr(0, pos));
temp.erase(0, pos + 1);
}
......
Dragon/modules/python/dragon.cc
......@@ -31,7 +31,7 @@ const TypeMeta& NumpyTypeToDragon(int numpy_type) {
{ NPY_FLOAT16, TypeMeta::Make<float16>() },
{ NPY_UINT8, TypeMeta::Make<uint8_t>() }};
static TypeMeta unknown_type; // id = 0
static TypeMeta unknown_type;
return dragon_type_map.count(numpy_type) ? dragon_type_map[numpy_type] : unknown_type;
}
......@@ -50,7 +50,7 @@ REGISTER_TENSOR_FETCHER(TypeMeta::Id<NumpyFetcher>(), NumpyFetcher);
REGISTER_TENSOR_FETCHER(TypeMeta::Id<StringFetcher>(), StringFetcher);
REGISTER_TENSOR_FEEDER(TypeMeta::Id<NumpyFeeder>(), NumpyFeeder);
extern "C"{
extern "C" {
PyObject* RegisteredOperatorsCC(PyObject* self, PyObject* args) {
set<string> all_keys;
......@@ -123,7 +123,7 @@ bool SwitchWorkspaceInternal(const string& name, const bool create_if_missing) {
} else if (create_if_missing) {
unique_ptr<Workspace> new_workspace(new Workspace());
g_workspace = new_workspace.get();
g_workspaces[name] = std::move(new_workspace); // ???
g_workspaces[name] = std::move(new_workspace);
g_current_workspace = name;
return true;
} else {
......
Dragon/modules/python/dragon.h
......@@ -33,7 +33,7 @@ inline PyObject* StdStringToPyBytes(const std::string& str) {
return PyBytes_FromStringAndSize(str.c_str(), str.size());
}
template <typename T>
inline void MakeStringInternal(std::stringstream& ss, const T& t){ ss << t; }
inline void MakeStringInternal(std::stringstream& ss, const T& t) { ss << t; }
template <typename T,typename ... Args>
inline void MakeStringInternal(std::stringstream& ss, const T& t, const Args& ... args) {
......@@ -124,7 +124,7 @@ class NumpyFeeder : public TensorFeederBase {
Tensor* tensor) override {
PyArrayObject* array = PyArray_GETCONTIGUOUS(original_array);
const TypeMeta& meta = NumpyTypeToDragon(PyArray_TYPE(array));
if (meta.id() == 0){
if (meta.id() == 0) {
PyErr_SetString(PyExc_TypeError, "numpy data type is not supported.");
return nullptr;
}
......
Dragon/modules/python/py_mpi.h
......@@ -63,11 +63,11 @@ inline PyObject* MPICreateGroupCC(PyObject* self, PyObject* args) {
// check inclue ranks
int size = PyList_Size(incl);
if (size > 0){
if (size > 0) {
all_ranks.clear();
unique_ptr<int> incl_ranks(new int[size]);
int* ranks = incl_ranks.get();
for (int i = 0; i < size; i++){
for (int i = 0; i < size; i++) {
ranks[i] = _PyInt_AsInt(PyList_GetItem(incl, i));
all_ranks.insert(ranks[i]);
}
......@@ -81,7 +81,7 @@ inline PyObject* MPICreateGroupCC(PyObject* self, PyObject* args) {
all_ranks.clear(); Set<int> tmp;
unique_ptr<int> excl_ranks(new int[size]);
int* ranks = excl_ranks.get();
for (int i = 0; i < size; i++){
for (int i = 0; i < size; i++) {
ranks[i] = _PyInt_AsInt(PyList_GetItem(excl, i));
tmp.insert(ranks[i]);
}
......@@ -97,7 +97,7 @@ inline PyObject* MPICreateGroupCC(PyObject* self, PyObject* args) {
if (local_comm != MPI_COMM_NULL) {
int world_rank, local_size;
MPI_Comm_rank(MPI_COMM_WORLD, &world_rank);
if (world_rank == local_root){
if (world_rank == local_root) {
MPI_Comm_size(local_comm, &local_size);
std::stringstream ss;
ss << "Rank[" << world_rank << "]: "
......
Dragon/python/dragon/io/__init__.py 0 → 100644
# --------------------------------------------------------
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
from .data_batch import DataBatch
Dragon/python/dragon/vm/caffe/io/blob_fetcher.py → Dragon/python/dragon/io/blob_fetcher.py
# --------------------------------------------------------
# Caffe for Dragon
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
......@@ -10,7 +10,7 @@ from six.moves import range as xrange
from dragon.config import logger
from .__init__ import GetProperty
from .utils import GetProperty
class BlobFetcher(Process):
def __init__(self, **kwargs):
......@@ -30,16 +30,16 @@ class BlobFetcher(Process):
atexit.register(cleanup)
def im_list_to_blob(self):
datum = self.Q_in.get() # (h, w, BGR)
im = datum[0]; h, w, c = im.shape
im_blob = np.zeros((self._batch_size, h, w, c), dtype=np.float32)
datum = self.Q_in.get()
im_blob = []
label_blob = np.zeros((self._batch_size, len(datum[1])), dtype=np.float32) \
if len(datum) > 1 else None
for i in xrange(0, self._batch_size):
im_blob[i, 0:h, 0:w, :] = datum[0]
im_blob.append(datum[0])
if label_blob is not None: label_blob[i, :] = datum[1]
if i != self._batch_size - 1: datum = self.Q_in.get()
channel_swap = (0, 3, 1, 2)
im_blob = np.array(im_blob, dtype=np.float32)
im_blob = im_blob.transpose(channel_swap)
return (im_blob, label_blob)
......
Dragon/python/dragon/vm/caffe/io/minibatch.py → Dragon/python/dragon/io/data_batch.py
# --------------------------------------------------------
# Caffe for Dragon
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
import sys
import time
import pprint
from multiprocessing import Queue
if sys.version_info >= (3,0):
from queue import Queue as Queue2
else:
from Queue import Queue as Queue2
import threading
from six.moves import range as xrange
import dragon.core.mpi as mpi
......@@ -16,10 +22,11 @@ from .data_reader import DataReader
from .data_transformer import DataTransformer
from .blob_fetcher import BlobFetcher
from .__init__ import GetProperty
from .utils import GetProperty
class DataBatch(object):
class DataBatch(threading.Thread):
def __init__(self, **kwargs):
super(DataBatch, self).__init__()
"""DataBatch use Triple-Buffering to speed up"""
......@@ -35,10 +42,10 @@ class DataBatch(object):
kwargs['group_size'] = group_size
# configuration
self._prefetch = GetProperty(kwargs, 'prefetch', 40)
self._prefetch = GetProperty(kwargs, 'prefetch', 5)
self._num_readers = GetProperty(kwargs, 'num_readers', 1)
self._num_transformers = GetProperty(kwargs, 'num_transformers', -1)
self._num_fetchers = GetProperty(kwargs, 'num_fetchers', 3)
self._num_fetchers = GetProperty(kwargs, 'num_fetchers', 1)
# default policy
if self._num_transformers == -1:
......@@ -60,6 +67,7 @@ class DataBatch(object):
self.Q_level_1 = Queue(self._prefetch * self._num_readers * self._batch_size)
self.Q_level_2 = Queue(self._prefetch * self._num_readers * self._batch_size)
self.Q_level_3 = Queue(self._prefetch * self._num_readers)
self.Q_level_4 = Queue2(self._prefetch * self._num_readers)
# init readers
self._readers = []
......@@ -102,11 +110,16 @@ class DataBatch(object):
self._fetchers.append(fetcher)
time.sleep(0.1)
self.daemon = True
self.start()
#self.echo()
@property
def blobs(self):
return self.Q_level_3.get()
def run(self):
while True:
self.Q_level_4.put(self.Q_level_3.get())
def get(self):
return self.Q_level_4.get()
def echo(self):
logger.info('---------------------------------------------------------')
......
Dragon/python/dragon/vm/caffe/io/data_reader.py → Dragon/python/dragon/io/data_reader.py
# --------------------------------------------------------
# Caffe for Dragon
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
......@@ -12,8 +12,7 @@ import dragon.config as config
from dragon.config import logger
from dragon.tools.db import LMDB
from .__init__ import GetProperty
from .utils import GetProperty
class DataReader(Process):
def __init__(self, **kwargs):
......
Dragon/python/dragon/vm/caffe/io/data_transformer.py → Dragon/python/dragon/io/data_transformer.py
# --------------------------------------------------------
# Caffe for Dragon
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
......@@ -12,7 +12,7 @@ import dragon.config as config
from dragon.config import logger
import dragon.vm.caffe.proto.caffe_pb2 as pb
from .__init__ import GetProperty
from .utils import GetProperty
try:
import cv2
......@@ -131,5 +131,3 @@ class DataTransformer(Process):
while True:
serialized = self.Q_in.get()
self.Q_out.put(self.transform_image_label(serialized))
\ No newline at end of file
Dragon/python/dragon/vm/caffe/io/__init__.py → Dragon/python/dragon/io/utils.py
# --------------------------------------------------------
# Caffe for Dragon
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
......
Dragon/python/dragon/vm/caffe/io/data_layer.py → Dragon/python/dragon/operators/custom/minibatch.py
# --------------------------------------------------------
# Caffe for Dragon
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
import dragon.vm.caffe as caffe
import dragon.core.workspace as ws
from .minibatch import DataBatch
from dragon.io.data_batch import DataBatch
class DataLayer(caffe.Layer):
def setup(self, bottom, top):
class MiniBatchOp(object):
def setup(self, inputs, outputs):
kwargs = eval(self.param_str)
self._data_batch = DataBatch(**kwargs)
def forward(self, bottom, top):
blobs = self._data_batch.blobs
def run(self, inputs, outputs):
blobs = self._data_batch.get()
for idx, blob in enumerate(blobs):
ws.FeedTensor(top[idx], blob)
\ No newline at end of file
ws.FeedTensor(outputs[idx], blob)
\ No newline at end of file
Dragon/python/dragon/operators/data.py
......@@ -10,7 +10,6 @@ from dragon.operators.utils import Run
def LMDBData(**kwargs):
"""
:param kwargs: a dict of imagenet data param
:param --> mean_value: a list of mean values for channles [B-G-R]
:param --> source: a str of the images root directory
:param --> imageset: a str of text file contains image name / label
......@@ -30,8 +29,8 @@ def LMDBData(**kwargs):
args = locals(); kwargs = args['kwargs']
del args['kwargs']; kwargs = dict(args, **kwargs)
kwargs['module'] = 'dragon.vm.caffe.io.data_layer'
kwargs['op'] = 'DataLayer'
kwargs['module'] = 'dragon.operators.custom.minibatch'
kwargs['op'] = 'MiniBatchOp'
return Run([], param_str=str(kwargs), nout=2, **kwargs)
......
Dragon/python/dragon/tools/im2lmdb.py 0 → 100644
# --------------------------------------------------------
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
""" Generate LMDB from images """
import os
import sys
import time
import shutil
import argparse
import cv2
try:
import numpy as np
except: pass
from dragon.tools.db import LMDB
from dragon.vm.caffe.proto import caffe_pb2
def resize_image(im, resize):
if im.shape[0] > im.shape[1]:
newsize = (resize, im.shape[0] * resize / im.shape[1])
else:
newsize = (im.shape[1] * resize / im.shape[0], resize)
im = cv2.resize(im, newsize)
return im
def make_db(args):
if os.path.isfile(args.list) is False:
raise ValueError('the path of image list is invalid.')
if os.path.isdir(args.database) is True:
raise ValueError('the database is already exist or invalid.')
print('start time: ', time.strftime("%a, %d %b %Y %H:%M:%S", time.gmtime()))
db = LMDB(max_commit=10000)
db.open(args.database, mode='w')
total_line = sum(1 for line in open(args.list))
count = 0
zfill_flag = '{0:0%d}' % (args.zfill)
encode_param = [int(cv2.IMWRITE_JPEG_QUALITY), args.quality]
start_time = time.time()
with open(args.list, 'r') as input_file:
records = input_file.readlines()
if args.shuffle:
import random
random.shuffle(records)
for record in records:
count += 1
if count % 10000 == 0:
now_time = time.time()
print('{0} / {1} in {2:.2f} sec'.format(
count, total_line, now_time - start_time))
db.commit()
record = record.split()
path = record[0]
label = record[1]
img = cv2.imread(os.path.join(args.root, path))
if args.resize > 0:
img = resize_image(img, args.resize)
if args.pad > 0:
pad_img = np.zeros((img.shape[0] + 2 * args.pad,
img.shape[1] + 2 * args.pad, 3), dtype=img.dtype)
pad_img[args.pad : args.pad + img.shape[0],
args.pad : args.pad + img.shape[1], :] = img
img = pad_img
result, imgencode = cv2.imencode('.jpg', img, encode_param)
datum = caffe_pb2.Datum()
datum.height, datum.width, datum.channels = img.shape
datum.label = int(label)
datum.encoded = True
datum.data = imgencode.tostring()
db.put(zfill_flag.format(count - 1), datum.SerializeToString())
now_time = time.time()
print('{0} / {1} in {2:.2f} sec'.format(count, total_line, now_time - start_time))
db.put('size', str(count))
db.put('zfill', str(args.zfill))
db.commit()
db.close()
shutil.copy(args.list, args.database + '/image_list.txt')
end_time = time.time()
print('{0} images have been stored in the database.'.format(total_line))
print('This task finishes within {0:.2f} seconds.'.format(end_time - start_time))
print('The size of database is {0} MB.'.
format(float(os.path.getsize(args.database + '/data.mdb') / 1000 / 1000)))
def parse_args():
parser = argparse.ArgumentParser(description='Create LMDB from images for classification.')
parser.add_argument('--root', help='the root folder of raw images')
parser.add_argument('--list', help='the filepath of image list')
parser.add_argument('--database', help='the filepath of database')
parser.add_argument('--zfill', type=int, default=8, help='zfill for the key of database')
parser.add_argument('--resize', type=int, default=0, help='resize the shorter edge of image to the newsize')
parser.add_argument('--pad', type=int, default=0, help='zero-pad the image')
parser.add_argument('--quality', type=int, default=95, help='JPEG quality for encoding, 1-100')
parser.add_argument('--shuffle', type=bool, default=True, help='randomize the order in list file True')
if len(sys.argv) < 4:
parser.print_help()
sys.exit(1)
args = parser.parse_args()
return args
if __name__ == '__main__':
args = parse_args()
make_db(args)
\ No newline at end of file
Dragon/python/dragon/utils.py deleted 100644 → 0
# --------------------------------------------------------
# Dragon
# Copyright(c) 2017 SeetaTech
# Written by Ting Pan
# --------------------------------------------------------
from google.protobuf.message import Message
from protos import dragon_pb2 as pb
import numpy as np
def MakeArgument(key, value):
argument = pb.Argument()
argument.name = key
if type(value) is float: argument.f = value
elif type(value) is int : argument.i = value
elif type(value) is np.int64: argument.i64 = int(value)
elif type(value) is str: argument.s = value
elif type(value) is unicode: argument.s = value
elif type(value) is bool: argument.b = value
elif isinstance(value, Message): argument.s = value.SerializeToString()
elif all(type(v) is float for v in value): argument.floats.extend(value)
elif all(type(v) is int for v in value): argument.ints.extend(value)
elif all(type(v) is str for v in value): argument.strings.extend(value)
elif all(type(v) is unicode or type(v) is str for v in value): argument.strings.extend(value)
elif all(isinstance(v,Message) for v in value):
argument.strings.extend([v.SerializeToString() for v in value])
else: raise ValueError('unknown argument type: key={} value={} value type={}' \
.format(key,value,type(value)))
return argument
def MakeOperatorDef(op_type, inputs, outputs, name='',
device_option=None, arg=None, engine=None, **kwargs):
operator = pb.OperatorDef()
operator.type = op_type
operator.name = name
operator.input.extend([str(tensor) for tensor in inputs])
operator.output.extend([str(tensor) for tensor in outputs])
if device_option is not None:
operator.device_option.CopyFrom(device_option)
if engine is not None:
operator.engine = engine
if 'random_seed' in kwargs:
operator.device_option.random_seed = kwargs['random_seed']
del kwargs['random_seed']
if arg is not None:
operator.arg.extend(arg)
for k,v in kwargs.items():
if v is None: continue
operator.arg.add().CopyFrom(MakeArgument(k,v))
return operator
def MakeDeviceOption(device_type, gpu_id, rng_seed = None):
""" return a DeviceOption """
option = pb.DeviceOption()
option.device_type = device_type
option.gpu_id = gpu_id
if rng_seed is not None: option.random_seed = rng_seed
return option
# fix the python stdout
class Unbuffered(object):
def __init__(self, stream):
self.stream = stream
def write(self, data):
self.stream.write(data)
self.stream.flush()
def __getattr__(self, attr):
return getattr(self.stream, attr)
# clear the stdout buffer for mpi(c++ & python)
import sys
sys.stdout = Unbuffered(sys.stdout)
\ No newline at end of file
Dragon/python/dragon/vm/caffe/proto/caffe.proto
......@@ -221,7 +221,7 @@ message SolverParameter {
// RMSProp decay value
// MeanSquare(t) = rms_decay*MeanSquare(t-1) + (1-rms_decay)*SquareGradient(t)
optional float rms_decay = 38;
optional float rms_decay = 38 [default = 0.99];
// If true, print information about the state of the net that may help with
// debugging learning problems.
......@@ -676,7 +676,7 @@ message DataParameter {
optional bool force_encoded_color = 9 [default = false];
// Prefetch queue (Number of batches to prefetch to host memory, increase if
// data access bandwidth varies).
optional uint32 prefetch = 10 [default = 40];
optional uint32 prefetch = 10 [default = 5];
}
message DropoutParameter {
......
Dragon/python/dragon/vm/caffe/proto/caffe_pb2.py
......@@ -19,7 +19,7 @@ _sym_db = _symbol_database.Default()
DESCRIPTOR = _descriptor.FileDescriptor(
name='caffe.proto',
package='caffe',
serialized_pb=_b('\n\x0b\x63\x61\x66\x66\x65.proto\x12\x05\x63\x61\x66\x66\x65\"\x1c\n\tBlobShape\x12\x0f\n\x03\x64im\x18\x01 \x03(\x03\x42\x02\x10\x01\"\xcc\x01\n\tBlobProto\x12\x1f\n\x05shape\x18\x07 \x01(\x0b\x32\x10.caffe.BlobShape\x12\x10\n\x04\x64\x61ta\x18\x05 \x03(\x02\x42\x02\x10\x01\x12\x10\n\x04\x64iff\x18\x06 \x03(\x02\x42\x02\x10\x01\x12\x17\n\x0b\x64ouble_data\x18\x08 \x03(\x01\x42\x02\x10\x01\x12\x17\n\x0b\x64ouble_diff\x18\t \x03(\x01\x42\x02\x10\x01\x12\x0e\n\x03num\x18\x01 \x01(\x05:\x01\x30\x12\x13\n\x08\x63hannels\x18\x02 \x01(\x05:\x01\x30\x12\x11\n\x06height\x18\x03 \x01(\x05:\x01\x30\x12\x10\n\x05width\x18\x04 \x01(\x05:\x01\x30\"2\n\x0f\x42lobProtoVector\x12\x1f\n\x05\x62lobs\x18\x01 \x03(\x0b\x32\x10.caffe.BlobProto\"\x81\x01\n\x05\x44\x61tum\x12\x10\n\x08\x63hannels\x18\x01 \x01(\x05\x12\x0e\n\x06height\x18\x02 \x01(\x05\x12\r\n\x05width\x18\x03 \x01(\x05\x12\x0c\n\x04\x64\x61ta\x18\x04 \x01(\x0c\x12\r\n\x05label\x18\x05 \x01(\x05\x12\x12\n\nfloat_data\x18\x06 \x03(\x02\x12\x16\n\x07\x65ncoded\x18\x07 \x01(\x08:\x05\x66\x61lse\"\x8a\x02\n\x0f\x46illerParameter\x12\x16\n\x04type\x18\x01 \x01(\t:\x08\x63onstant\x12\x10\n\x05value\x18\x02 \x01(\x02:\x01\x30\x12\x0e\n\x03min\x18\x03 \x01(\x02:\x01\x30\x12\x0e\n\x03max\x18\x04 \x01(\x02:\x01\x31\x12\x0f\n\x04mean\x18\x05 \x01(\x02:\x01\x30\x12\x0e\n\x03std\x18\x06 \x01(\x02:\x01\x31\x12\x12\n\x06sparse\x18\x07 \x01(\x05:\x02-1\x12\x42\n\rvariance_norm\x18\x08 \x01(\x0e\x32#.caffe.FillerParameter.VarianceNorm:\x06\x46\x41N_IN\"4\n\x0cVarianceNorm\x12\n\n\x06\x46\x41N_IN\x10\x00\x12\x0b\n\x07\x46\x41N_OUT\x10\x01\x12\x0b\n\x07\x41VERAGE\x10\x02\"\x8e\x02\n\x0cNetParameter\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\r\n\x05input\x18\x03 \x03(\t\x12%\n\x0binput_shape\x18\x08 \x03(\x0b\x32\x10.caffe.BlobShape\x12\x11\n\tinput_dim\x18\x04 \x03(\x05\x12\x1d\n\x0e\x66orce_backward\x18\x05 \x01(\x08:\x05\x66\x61lse\x12\x1e\n\x05state\x18\x06 \x01(\x0b\x32\x0f.caffe.NetState\x12\x19\n\ndebug_info\x18\x07 \x01(\x08:\x05\x66\x61lse\x12$\n\x05layer\x18\x64 \x03(\x0b\x32\x15.caffe.LayerParameter\x12\'\n\x06layers\x18\x02 \x03(\x0b\x32\x17.caffe.V1LayerParameter\"\xc9\n\n\x0fSolverParameter\x12\x0b\n\x03net\x18\x18 \x01(\t\x12&\n\tnet_param\x18\x19 \x01(\x0b\x32\x13.caffe.NetParameter\x12\x11\n\ttrain_net\x18\x01 \x01(\t\x12\x10\n\x08test_net\x18\x02 \x03(\t\x12,\n\x0ftrain_net_param\x18\x15 \x01(\x0b\x32\x13.caffe.NetParameter\x12+\n\x0etest_net_param\x18\x16 \x03(\x0b\x32\x13.caffe.NetParameter\x12$\n\x0btrain_state\x18\x1a \x01(\x0b\x32\x0f.caffe.NetState\x12#\n\ntest_state\x18\x1b \x03(\x0b\x32\x0f.caffe.NetState\x12\x11\n\ttest_iter\x18\x03 \x03(\x05\x12\x18\n\rtest_interval\x18\x04 \x01(\x05:\x01\x30\x12 \n\x11test_compute_loss\x18\x13 \x01(\x08:\x05\x66\x61lse\x12!\n\x13test_initialization\x18 \x01(\x08:\x04true\x12\x0f\n\x07\x62\x61se_lr\x18\x05 \x01(\x02\x12\x10\n\x08stage_lr\x18\x32 \x03(\x02\x12\x12\n\nstage_iter\x18\x33 \x03(\x05\x12\x0f\n\x07\x64isplay\x18\x06 \x01(\x05\x12\x17\n\x0c\x61verage_loss\x18! \x01(\x05:\x01\x31\x12\x10\n\x08max_iter\x18\x07 \x01(\x05\x12\x14\n\titer_size\x18$ \x01(\x05:\x01\x31\x12\x11\n\tlr_policy\x18\x08 \x01(\t\x12\r\n\x05gamma\x18\t \x01(\x02\x12\r\n\x05power\x18\n \x01(\x02\x12\x10\n\x08momentum\x18\x0b \x01(\x02\x12\x14\n\x0cweight_decay\x18\x0c \x01(\x02\x12\x1f\n\x13regularization_type\x18\x1d \x01(\t:\x02L2\x12\x10\n\x08stepsize\x18\r \x01(\x05\x12\x11\n\tstepvalue\x18\" \x03(\x05\x12\x1a\n\x0e\x63lip_gradients\x18# \x01(\x02:\x02-1\x12\x13\n\x08snapshot\x18\x0e \x01(\x05:\x01\x30\x12\x17\n\x0fsnapshot_prefix\x18\x0f \x01(\t\x12\x1c\n\rsnapshot_diff\x18\x10 \x01(\x08:\x05\x66\x61lse\x12K\n\x0fsnapshot_format\x18% \x01(\x0e\x32%.caffe.SolverParameter.SnapshotFormat:\x0b\x42INARYPROTO\x12;\n\x0bsolver_mode\x18\x11 \x01(\x0e\x32!.caffe.SolverParameter.SolverMode:\x03GPU\x12\x14\n\tdevice_id\x18\x12 \x01(\x05:\x01\x30\x12\x17\n\x0brandom_seed\x18\x14 \x01(\x03:\x02-1\x12\x11\n\x04type\x18( \x01(\t:\x03SGD\x12\x15\n\x05\x64\x65lta\x18\x1f \x01(\x02:\x06\x31\x65-008\x12\x18\n\tmomentum2\x18\' \x01(\x02:\x05\x30.999\x12\x17\n\trms_decay\x18& \x01(\x02:\x04\x30.99\x12\x19\n\ndebug_info\x18\x17 \x01(\x08:\x05\x66\x61lse\x12\"\n\x14snapshot_after_train\x18\x1c \x01(\x08:\x04true\x12;\n\x0bsolver_type\x18\x1e \x01(\x0e\x32!.caffe.SolverParameter.SolverType:\x03SGD\"+\n\x0eSnapshotFormat\x12\x08\n\x04HDF5\x10\x00\x12\x0f\n\x0b\x42INARYPROTO\x10\x01\"\x1e\n\nSolverMode\x12\x07\n\x03\x43PU\x10\x00\x12\x07\n\x03GPU\x10\x01\"U\n\nSolverType\x12\x07\n\x03SGD\x10\x00\x12\x0c\n\x08NESTEROV\x10\x01\x12\x0b\n\x07\x41\x44\x41GRAD\x10\x02\x12\x0b\n\x07RMSPROP\x10\x03\x12\x0c\n\x08\x41\x44\x41\x44\x45LTA\x10\x04\x12\x08\n\x04\x41\x44\x41M\x10\x05\"l\n\x0bSolverState\x12\x0c\n\x04iter\x18\x01 \x01(\x05\x12\x13\n\x0blearned_net\x18\x02 \x01(\t\x12!\n\x07history\x18\x03 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x17\n\x0c\x63urrent_step\x18\x04 \x01(\x05:\x01\x30\"N\n\x08NetState\x12!\n\x05phase\x18\x01 \x01(\x0e\x32\x0c.caffe.Phase:\x04TEST\x12\x10\n\x05level\x18\x02 \x01(\x05:\x01\x30\x12\r\n\x05stage\x18\x03 \x03(\t\"\x85\x01\n\x0cNetStateRule\x12\x1b\n\x05phase\x18\x01 \x01(\x0e\x32\x0c.caffe.Phase\x12\x11\n\tmin_level\x18\x02 \x01(\x05\x12\x11\n\tmax_level\x18\x03 \x01(\x05\x12\r\n\x05stage\x18\x04 \x03(\t\x12\x11\n\tnot_stage\x18\x05 \x03(\t\x12\x10\n\x08mpi_rank\x18\x06 \x03(\r\"\xa3\x01\n\tParamSpec\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x31\n\nshare_mode\x18\x02 \x01(\x0e\x32\x1d.caffe.ParamSpec.DimCheckMode\x12\x12\n\x07lr_mult\x18\x03 \x01(\x02:\x01\x31\x12\x15\n\ndecay_mult\x18\x04 \x01(\x02:\x01\x31\"*\n\x0c\x44imCheckMode\x12\n\n\x06STRICT\x10\x00\x12\x0e\n\nPERMISSIVE\x10\x01\"\x8f\x18\n\x0eLayerParameter\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04type\x18\x02 \x01(\t\x12\x0e\n\x06\x62ottom\x18\x03 \x03(\t\x12\x0b\n\x03top\x18\x04 \x03(\t\x12\x1b\n\x05phase\x18\n \x01(\x0e\x32\x0c.caffe.Phase\x12\x13\n\x0bloss_weight\x18\x05 \x03(\x02\x12\x1f\n\x05param\x18\x06 \x03(\x0b\x32\x10.caffe.ParamSpec\x12\x1f\n\x05\x62lobs\x18\x07 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x16\n\x0epropagate_down\x18\x0b \x03(\x08\x12$\n\x07include\x18\x08 \x03(\x0b\x32\x13.caffe.NetStateRule\x12$\n\x07\x65xclude\x18\t \x03(\x0b\x32\x13.caffe.NetStateRule\x12\x37\n\x0ftransform_param\x18\x64 \x01(\x0b\x32\x1e.caffe.TransformationParameter\x12(\n\nloss_param\x18\x65 \x01(\x0b\x32\x14.caffe.LossParameter\x12\x30\n\x0e\x61\x63\x63uracy_param\x18\x66 \x01(\x0b\x32\x18.caffe.AccuracyParameter\x12,\n\x0c\x61rgmax_param\x18g \x01(\x0b\x32\x16.caffe.ArgMaxParameter\x12\x34\n\x10\x62\x61tch_norm_param\x18\x8b\x01 \x01(\x0b\x32\x19.caffe.BatchNormParameter\x12)\n\nbias_param\x18\x8d\x01 \x01(\x0b\x32\x14.caffe.BiasParameter\x12,\n\x0c\x63oncat_param\x18h \x01(\x0b\x32\x16.caffe.ConcatParameter\x12?\n\x16\x63ontrastive_loss_param\x18i \x01(\x0b\x32\x1f.caffe.ContrastiveLossParameter\x12\x36\n\x11\x63onvolution_param\x18j \x01(\x0b\x32\x1b.caffe.ConvolutionParameter\x12)\n\ncrop_param\x18\x90\x01 \x01(\x0b\x32\x14.caffe.CropParameter\x12(\n\ndata_param\x18k \x01(\x0b\x32\x14.caffe.DataParameter\x12.\n\rdropout_param\x18l \x01(\x0b\x32\x17.caffe.DropoutParameter\x12\x33\n\x10\x64ummy_data_param\x18m \x01(\x0b\x32\x19.caffe.DummyDataParameter\x12.\n\reltwise_param\x18n \x01(\x0b\x32\x17.caffe.EltwiseParameter\x12\'\n\telu_param\x18\x8c\x01 \x01(\x0b\x32\x13.caffe.ELUParameter\x12+\n\x0b\x65mbed_param\x18\x89\x01 \x01(\x0b\x32\x15.caffe.EmbedParameter\x12&\n\texp_param\x18o \x01(\x0b\x32\x13.caffe.ExpParameter\x12/\n\rflatten_param\x18\x87\x01 \x01(\x0b\x32\x17.caffe.FlattenParameter\x12\x31\n\x0fhdf5_data_param\x18p \x01(\x0b\x32\x18.caffe.HDF5DataParameter\x12\x35\n\x11hdf5_output_param\x18q \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\x12\x33\n\x10hinge_loss_param\x18r \x01(\x0b\x32\x19.caffe.HingeLossParameter\x12\x33\n\x10image_data_param\x18s \x01(\x0b\x32\x19.caffe.ImageDataParameter\x12\x39\n\x13infogain_loss_param\x18t \x01(\x0b\x32\x1c.caffe.InfogainLossParameter\x12\x39\n\x13inner_product_param\x18u \x01(\x0b\x32\x1c.caffe.InnerProductParameter\x12+\n\x0binput_param\x18\x8f\x01 \x01(\x0b\x32\x15.caffe.InputParameter\x12\'\n\tlog_param\x18\x86\x01 \x01(\x0b\x32\x13.caffe.LogParameter\x12&\n\tlrn_param\x18v \x01(\x0b\x32\x13.caffe.LRNParameter\x12\x35\n\x11memory_data_param\x18w \x01(\x0b\x32\x1a.caffe.MemoryDataParameter\x12&\n\tmvn_param\x18x \x01(\x0b\x32\x13.caffe.MVNParameter\x12\x33\n\x0fparameter_param\x18\x91\x01 \x01(\x0b\x32\x19.caffe.ParameterParameter\x12.\n\rpooling_param\x18y \x01(\x0b\x32\x17.caffe.PoolingParameter\x12*\n\x0bpower_param\x18z \x01(\x0b\x32\x15.caffe.PowerParameter\x12+\n\x0bprelu_param\x18\x83\x01 \x01(\x0b\x32\x15.caffe.PReLUParameter\x12-\n\x0cpython_param\x18\x82\x01 \x01(\x0b\x32\x16.caffe.PythonParameter\x12\x33\n\x0freduction_param\x18\x88\x01 \x01(\x0b\x32\x19.caffe.ReductionParameter\x12(\n\nrelu_param\x18{ \x01(\x0b\x32\x14.caffe.ReLUParameter\x12/\n\rreshape_param\x18\x85\x01 \x01(\x0b\x32\x17.caffe.ReshapeParameter\x12+\n\x0bscale_param\x18\x8e\x01 \x01(\x0b\x32\x15.caffe.ScaleParameter\x12.\n\rsigmoid_param\x18| \x01(\x0b\x32\x17.caffe.SigmoidParameter\x12.\n\rsoftmax_param\x18} \x01(\x0b\x32\x17.caffe.SoftmaxParameter\x12\'\n\tspp_param\x18\x84\x01 \x01(\x0b\x32\x13.caffe.SPPParameter\x12*\n\x0bslice_param\x18~ \x01(\x0b\x32\x15.caffe.SliceParameter\x12(\n\ntanh_param\x18\x7f \x01(\x0b\x32\x14.caffe.TanHParameter\x12\x33\n\x0fthreshold_param\x18\x80\x01 \x01(\x0b\x32\x19.caffe.ThresholdParameter\x12)\n\ntile_param\x18\x8a\x01 \x01(\x0b\x32\x14.caffe.TileParameter\x12\x36\n\x11window_data_param\x18\x81\x01 \x01(\x0b\x32\x1a.caffe.WindowDataParameter\x12\x31\n\x0eimagenet_param\x18\x96\x01 \x01(\x0b\x32\x18.caffe.ImagenetParameter\x12\x36\n\x11roi_pooling_param\x18\x97\x01 \x01(\x0b\x32\x1a.caffe.ROIPoolingParameter\x12;\n\x14smooth_l1_loss_param\x18\x98\x01 \x01(\x0b\x32\x1c.caffe.SmoothL1LossParameter\x12\'\n\tmpi_param\x18\x99\x01 \x01(\x0b\x32\x13.caffe.MPIParameter\x12/\n\rpermute_param\x18\x9a\x01 \x01(\x0b\x32\x17.caffe.PermuteParameter\x12\x33\n\x0fnormalize_param\x18\x9b\x01 \x01(\x0b\x32\x19.caffe.NormalizeParameter\x12\x31\n\x0eparallel_param\x18\x9d\x01 \x01(\x0b\x32\x18.caffe.ParallelParameter\x12\x31\n\x0ennresize_param\x18\x9e\x01 \x01(\x0b\x32\x18.caffe.NNResizeParameter\x12\x36\n\x11\x65xpand_dims_param\x18\x9f\x01 \x01(\x0b\x32\x1a.caffe.ExpandDimsParameter\x12\x31\n\x0eproposal_param\x18\xa0\x01 \x01(\x0b\x32\x18.caffe.ProposalParameter\x12\x38\n\x12\x62\x61tch_renorm_param\x18\xa1\x01 \x01(\x0b\x32\x1b.caffe.BatchRenormParameter\"\x93\x02\n\x17TransformationParameter\x12\x10\n\x05scale\x18\x01 \x01(\x02:\x01\x31\x12\x15\n\x06mirror\x18\x02 \x01(\x08:\x05\x66\x61lse\x12\x14\n\tcrop_size\x18\x03 \x01(\r:\x01\x30\x12\x11\n\tmean_file\x18\x04 \x01(\t\x12\x12\n\nmean_value\x18\x05 \x03(\x02\x12\x1a\n\x0b\x66orce_color\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x19\n\nforce_gray\x18\x07 \x01(\x08:\x05\x66\x61lse\x12!\n\x12\x63olor_augmentation\x18\x08 \x01(\x08:\x05\x66\x61lse\x12\x1b\n\x10min_random_scale\x18\t \x01(\x02:\x01\x31\x12\x1b\n\x10max_random_scale\x18\n \x01(\x02:\x01\x31\"\xeb\x01\n\rLossParameter\x12\x14\n\x0cignore_label\x18\x01 \x01(\x05\x12\x44\n\rnormalization\x18\x03 \x01(\x0e\x32&.caffe.LossParameter.NormalizationMode:\x05VALID\x12\x11\n\tnormalize\x18\x02 \x01(\x08\x1a\'\n\x13\x45xpandDimsParameter\x12\x10\n\x04\x61xis\x18\x01 \x01(\x05:\x02-1\"B\n\x11NormalizationMode\x12\x08\n\x04\x46ULL\x10\x00\x12\t\n\x05VALID\x10\x01\x12\x0e\n\nBATCH_SIZE\x10\x02\x12\x08\n\x04NONE\x10\x03\"L\n\x11\x41\x63\x63uracyParameter\x12\x10\n\x05top_k\x18\x01 \x01(\r:\x01\x31\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x31\x12\x14\n\x0cignore_label\x18\x03 \x01(\x05\"M\n\x0f\x41rgMaxParameter\x12\x1a\n\x0bout_max_val\x18\x01 \x01(\x08:\x05\x66\x61lse\x12\x10\n\x05top_k\x18\x02 \x01(\r:\x01\x31\x12\x0c\n\x04\x61xis\x18\x03 \x01(\x05\"9\n\x0f\x43oncatParameter\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x31\x12\x15\n\nconcat_dim\x18\x01 \x01(\r:\x01\x31\"h\n\x12\x42\x61tchNormParameter\x12\x18\n\x10use_global_stats\x18\x01 \x01(\x08\x12$\n\x17moving_average_fraction\x18\x02 \x01(\x02:\x03\x30.9\x12\x12\n\x03\x65ps\x18\x03 \x01(\x02:\x05\x30.001\"]\n\rBiasParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\x13\n\x08num_axes\x18\x02 \x01(\x05:\x01\x31\x12&\n\x06\x66iller\x18\x03 \x01(\x0b\x32\x16.caffe.FillerParameter\"L\n\x18\x43ontrastiveLossParameter\x12\x11\n\x06margin\x18\x01 \x01(\x02:\x01\x31\x12\x1d\n\x0elegacy_version\x18\x02 \x01(\x08:\x05\x66\x61lse\"\xfc\x03\n\x14\x43onvolutionParameter\x12\x12\n\nnum_output\x18\x01 \x01(\r\x12\x17\n\tbias_term\x18\x02 \x01(\x08:\x04true\x12\x0b\n\x03pad\x18\x03 \x03(\r\x12\x13\n\x0bkernel_size\x18\x04 \x03(\r\x12\x0e\n\x06stride\x18\x06 \x03(\r\x12\x10\n\x08\x64ilation\x18\x12 \x03(\r\x12\x10\n\x05pad_h\x18\t \x01(\r:\x01\x30\x12\x10\n\x05pad_w\x18\n \x01(\r:\x01\x30\x12\x10\n\x08kernel_h\x18\x0b \x01(\r\x12\x10\n\x08kernel_w\x18\x0c \x01(\r\x12\x10\n\x08stride_h\x18\r \x01(\r\x12\x10\n\x08stride_w\x18\x0e \x01(\r\x12\x10\n\x05group\x18\x05 \x01(\r:\x01\x31\x12-\n\rweight_filler\x18\x07 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x08 \x01(\x0b\x32\x16.caffe.FillerParameter\x12;\n\x06\x65ngine\x18\x0f \x01(\x0e\x32\".caffe.ConvolutionParameter.Engine:\x07\x44\x45\x46\x41ULT\x12\x0f\n\x04\x61xis\x18\x10 \x01(\x05:\x01\x31\x12\x1e\n\x0f\x66orce_nd_im2col\x18\x11 \x01(\x08:\x05\x66\x61lse\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"0\n\rCropParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x32\x12\x0e\n\x06offset\x18\x02 \x03(\r\"\xa5\x02\n\rDataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x12\n\nbatch_size\x18\x04 \x01(\r\x12\x14\n\trand_skip\x18\x07 \x01(\r:\x01\x30\x12\x31\n\x07\x62\x61\x63kend\x18\x08 \x01(\x0e\x32\x17.caffe.DataParameter.DB:\x07LEVELDB\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x11\n\tmean_file\x18\x03 \x01(\t\x12\x14\n\tcrop_size\x18\x05 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\"\n\x13\x66orce_encoded_color\x18\t \x01(\x08:\x05\x66\x61lse\x12\x14\n\x08prefetch\x18\n \x01(\r:\x02\x34\x30\"\x1b\n\x02\x44\x42\x12\x0b\n\x07LEVELDB\x10\x00\x12\x08\n\x04LMDB\x10\x01\"I\n\x10\x44ropoutParameter\x12\x1a\n\rdropout_ratio\x18\x01 \x01(\x02:\x03\x30.5\x12\x19\n\x0bscale_train\x18\x02 \x01(\x08:\x04true\"\xa0\x01\n\x12\x44ummyDataParameter\x12+\n\x0b\x64\x61ta_filler\x18\x01 \x03(\x0b\x32\x16.caffe.FillerParameter\x12\x1f\n\x05shape\x18\x06 \x03(\x0b\x32\x10.caffe.BlobShape\x12\x0b\n\x03num\x18\x02 \x03(\r\x12\x10\n\x08\x63hannels\x18\x03 \x03(\r\x12\x0e\n\x06height\x18\x04 \x03(\r\x12\r\n\x05width\x18\x05 \x03(\r\"\xa5\x01\n\x10\x45ltwiseParameter\x12\x39\n\toperation\x18\x01 \x01(\x0e\x32!.caffe.EltwiseParameter.EltwiseOp:\x03SUM\x12\r\n\x05\x63oeff\x18\x02 \x03(\x02\x12\x1e\n\x10stable_prod_grad\x18\x03 \x01(\x08:\x04true\"\'\n\tEltwiseOp\x12\x08\n\x04PROD\x10\x00\x12\x07\n\x03SUM\x10\x01\x12\x07\n\x03MAX\x10\x02\" \n\x0c\x45LUParameter\x12\x10\n\x05\x61lpha\x18\x01 \x01(\x02:\x01\x31\"\xac\x01\n\x0e\x45mbedParameter\x12\x12\n\nnum_output\x18\x01 \x01(\r\x12\x11\n\tinput_dim\x18\x02 \x01(\r\x12\x17\n\tbias_term\x18\x03 \x01(\x08:\x04true\x12-\n\rweight_filler\x18\x04 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x05 \x01(\x0b\x32\x16.caffe.FillerParameter\"D\n\x0c\x45xpParameter\x12\x10\n\x04\x62\x61se\x18\x01 \x01(\x02:\x02-1\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x10\n\x05shift\x18\x03 \x01(\x02:\x01\x30\"9\n\x10\x46lattenParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\x14\n\x08\x65nd_axis\x18\x02 \x01(\x05:\x02-1\"O\n\x11HDF5DataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x12\n\nbatch_size\x18\x02 \x01(\r\x12\x16\n\x07shuffle\x18\x03 \x01(\x08:\x05\x66\x61lse\"(\n\x13HDF5OutputParameter\x12\x11\n\tfile_name\x18\x01 \x01(\t\"^\n\x12HingeLossParameter\x12\x30\n\x04norm\x18\x01 \x01(\x0e\x32\x1e.caffe.HingeLossParameter.Norm:\x02L1\"\x16\n\x04Norm\x12\x06\n\x02L1\x10\x01\x12\x06\n\x02L2\x10\x02\"\x97\x02\n\x12ImageDataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x15\n\nbatch_size\x18\x04 \x01(\r:\x01\x31\x12\x14\n\trand_skip\x18\x07 \x01(\r:\x01\x30\x12\x16\n\x07shuffle\x18\x08 \x01(\x08:\x05\x66\x61lse\x12\x15\n\nnew_height\x18\t \x01(\r:\x01\x30\x12\x14\n\tnew_width\x18\n \x01(\r:\x01\x30\x12\x16\n\x08is_color\x18\x0b \x01(\x08:\x04true\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x11\n\tmean_file\x18\x03 \x01(\t\x12\x14\n\tcrop_size\x18\x05 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x15\n\x0broot_folder\x18\x0c \x01(\t:\x00\"\'\n\x15InfogainLossParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\"\xcb\x01\n\x15InnerProductParameter\x12\x12\n\nnum_output\x18\x01 \x01(\r\x12\x17\n\tbias_term\x18\x02 \x01(\x08:\x04true\x12-\n\rweight_filler\x18\x03 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x04 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x0f\n\x04\x61xis\x18\x05 \x01(\x05:\x01\x31\x12\x18\n\ttranspose\x18\x06 \x01(\x08:\x05\x66\x61lse\"1\n\x0eInputParameter\x12\x1f\n\x05shape\x18\x01 \x03(\x0b\x32\x10.caffe.BlobShape\"D\n\x0cLogParameter\x12\x10\n\x04\x62\x61se\x18\x01 \x01(\x02:\x02-1\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x10\n\x05shift\x18\x03 \x01(\x02:\x01\x30\"\xb8\x02\n\x0cLRNParameter\x12\x15\n\nlocal_size\x18\x01 \x01(\r:\x01\x35\x12\x10\n\x05\x61lpha\x18\x02 \x01(\x02:\x01\x31\x12\x12\n\x04\x62\x65ta\x18\x03 \x01(\x02:\x04\x30.75\x12\x44\n\x0bnorm_region\x18\x04 \x01(\x0e\x32\x1e.caffe.LRNParameter.NormRegion:\x0f\x41\x43ROSS_CHANNELS\x12\x0c\n\x01k\x18\x05 \x01(\x02:\x01\x31\x12\x33\n\x06\x65ngine\x18\x06 \x01(\x0e\x32\x1a.caffe.LRNParameter.Engine:\x07\x44\x45\x46\x41ULT\"5\n\nNormRegion\x12\x13\n\x0f\x41\x43ROSS_CHANNELS\x10\x00\x12\x12\n\x0eWITHIN_CHANNEL\x10\x01\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"\xbd\x01\n\x13MemoryDataParameter\x12\x12\n\nbatch_size\x18\x01 \x01(\r\x12\x10\n\x08\x63hannels\x18\x02 \x01(\r\x12\x0e\n\x06height\x18\x03 \x01(\r\x12\r\n\x05width\x18\x04 \x01(\r\x12;\n\x05\x64type\x18\x05 \x01(\x0e\x32#.caffe.MemoryDataParameter.DataType:\x07\x46LOAT32\"$\n\x08\x44\x61taType\x12\x0b\n\x07\x46LOAT32\x10\x00\x12\x0b\n\x07\x46LOAT16\x10\x01\"e\n\x0cMVNParameter\x12 \n\x12normalize_variance\x18\x01 \x01(\x08:\x04true\x12\x1e\n\x0f\x61\x63ross_channels\x18\x02 \x01(\x08:\x05\x66\x61lse\x12\x13\n\x03\x65ps\x18\x03 \x01(\x02:\x06\x31\x65-009\"5\n\x12ParameterParameter\x12\x1f\n\x05shape\x18\x01 \x01(\x0b\x32\x10.caffe.BlobShape\"\xa2\x03\n\x10PoolingParameter\x12\x35\n\x04pool\x18\x01 \x01(\x0e\x32\".caffe.PoolingParameter.PoolMethod:\x03MAX\x12\x0e\n\x03pad\x18\x04 \x01(\r:\x01\x30\x12\x10\n\x05pad_h\x18\t \x01(\r:\x01\x30\x12\x10\n\x05pad_w\x18\n \x01(\r:\x01\x30\x12\x13\n\x0bkernel_size\x18\x02 \x01(\r\x12\x10\n\x08kernel_h\x18\x05 \x01(\r\x12\x10\n\x08kernel_w\x18\x06 \x01(\r\x12\x11\n\x06stride\x18\x03 \x01(\r:\x01\x31\x12\x10\n\x08stride_h\x18\x07 \x01(\r\x12\x10\n\x08stride_w\x18\x08 \x01(\r\x12\x37\n\x06\x65ngine\x18\x0b \x01(\x0e\x32\x1e.caffe.PoolingParameter.Engine:\x07\x44\x45\x46\x41ULT\x12\x1d\n\x0eglobal_pooling\x18\x0c \x01(\x08:\x05\x66\x61lse\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"Y\n\x13ROIPoolingParameter\x12\x13\n\x08pooled_h\x18\x01 \x01(\r:\x01\x30\x12\x13\n\x08pooled_w\x18\x02 \x01(\r:\x01\x30\x12\x18\n\rspatial_scale\x18\x03 \x01(\x02:\x01\x31\"F\n\x0ePowerParameter\x12\x10\n\x05power\x18\x01 \x01(\x02:\x01\x31\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x10\n\x05shift\x18\x03 \x01(\x02:\x01\x30\"g\n\x0fPythonParameter\x12\x0e\n\x06module\x18\x01 \x01(\t\x12\r\n\x05layer\x18\x02 \x01(\t\x12\x13\n\tparam_str\x18\x03 \x01(\t:\x00\x12 \n\x11share_in_parallel\x18\x04 \x01(\x08:\x05\x66\x61lse\"\xad\x01\n\x12ReductionParameter\x12=\n\toperation\x18\x01 \x01(\x0e\x32%.caffe.ReductionParameter.ReductionOp:\x03SUM\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x30\x12\x10\n\x05\x63oeff\x18\x03 \x01(\x02:\x01\x31\"5\n\x0bReductionOp\x12\x07\n\x03SUM\x10\x01\x12\x08\n\x04\x41SUM\x10\x02\x12\t\n\x05SUMSQ\x10\x03\x12\x08\n\x04MEAN\x10\x04\"\x8d\x01\n\rReLUParameter\x12\x19\n\x0enegative_slope\x18\x01 \x01(\x02:\x01\x30\x12\x34\n\x06\x65ngine\x18\x02 \x01(\x0e\x32\x1b.caffe.ReLUParameter.Engine:\x07\x44\x45\x46\x41ULT\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"Z\n\x10ReshapeParameter\x12\x1f\n\x05shape\x18\x01 \x01(\x0b\x32\x10.caffe.BlobShape\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x30\x12\x14\n\x08num_axes\x18\x03 \x01(\x05:\x02-1\"\xa5\x01\n\x0eScaleParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\x13\n\x08num_axes\x18\x02 \x01(\x05:\x01\x31\x12&\n\x06\x66iller\x18\x03 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x18\n\tbias_term\x18\x04 \x01(\x08:\x05\x66\x61lse\x12+\n\x0b\x62ias_filler\x18\x05 \x01(\x0b\x32\x16.caffe.FillerParameter\"x\n\x10SigmoidParameter\x12\x37\n\x06\x65ngine\x18\x01 \x01(\x0e\x32\x1e.caffe.SigmoidParameter.Engine:\x07\x44\x45\x46\x41ULT\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"L\n\x0eSliceParameter\x12\x0f\n\x04\x61xis\x18\x03 \x01(\x05:\x01\x31\x12\x13\n\x0bslice_point\x18\x02 \x03(\r\x12\x14\n\tslice_dim\x18\x01 \x01(\r:\x01\x31\"\x89\x01\n\x10SoftmaxParameter\x12\x37\n\x06\x65ngine\x18\x01 \x01(\x0e\x32\x1e.caffe.SoftmaxParameter.Engine:\x07\x44\x45\x46\x41ULT\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x31\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"r\n\rTanHParameter\x12\x34\n\x06\x65ngine\x18\x01 \x01(\x0e\x32\x1b.caffe.TanHParameter.Engine:\x07\x44\x45\x46\x41ULT\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"T\n\rTileParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\r\n\x05tiles\x18\x02 \x01(\x05\x12#\n\tmultiples\x18\x03 \x01(\x0b\x32\x10.caffe.BlobShape\"*\n\x12ThresholdParameter\x12\x14\n\tthreshold\x18\x01 \x01(\x02:\x01\x30\"\xc1\x02\n\x13WindowDataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x11\n\tmean_file\x18\x03 \x01(\t\x12\x12\n\nbatch_size\x18\x04 \x01(\r\x12\x14\n\tcrop_size\x18\x05 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x19\n\x0c\x66g_threshold\x18\x07 \x01(\x02:\x03\x30.5\x12\x19\n\x0c\x62g_threshold\x18\x08 \x01(\x02:\x03\x30.5\x12\x19\n\x0b\x66g_fraction\x18\t \x01(\x02:\x04\x30.25\x12\x16\n\x0b\x63ontext_pad\x18\n \x01(\r:\x01\x30\x12\x17\n\tcrop_mode\x18\x0b \x01(\t:\x04warp\x12\x1b\n\x0c\x63\x61\x63he_images\x18\x0c \x01(\x08:\x05\x66\x61lse\x12\x15\n\x0broot_folder\x18\r \x01(\t:\x00\"\xeb\x01\n\x0cSPPParameter\x12\x16\n\x0epyramid_height\x18\x01 \x01(\r\x12\x31\n\x04pool\x18\x02 \x01(\x0e\x32\x1e.caffe.SPPParameter.PoolMethod:\x03MAX\x12\x33\n\x06\x65ngine\x18\x06 \x01(\x0e\x32\x1a.caffe.SPPParameter.Engine:\x07\x44\x45\x46\x41ULT\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"\xe0\x13\n\x10V1LayerParameter\x12\x0e\n\x06\x62ottom\x18\x02 \x03(\t\x12\x0b\n\x03top\x18\x03 \x03(\t\x12\x0c\n\x04name\x18\x04 \x01(\t\x12$\n\x07include\x18 \x03(\x0b\x32\x13.caffe.NetStateRule\x12$\n\x07\x65xclude\x18! \x03(\x0b\x32\x13.caffe.NetStateRule\x12/\n\x04type\x18\x05 \x01(\x0e\x32!.caffe.V1LayerParameter.LayerType\x12\x1f\n\x05\x62lobs\x18\x06 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x0e\n\x05param\x18\xe9\x07 \x03(\t\x12>\n\x0f\x62lob_share_mode\x18\xea\x07 \x03(\x0e\x32$.caffe.V1LayerParameter.DimCheckMode\x12\x10\n\x08\x62lobs_lr\x18\x07 \x03(\x02\x12\x14\n\x0cweight_decay\x18\x08 \x03(\x02\x12\x13\n\x0bloss_weight\x18# \x03(\x02\x12\x30\n\x0e\x61\x63\x63uracy_param\x18\x1b \x01(\x0b\x32\x18.caffe.AccuracyParameter\x12,\n\x0c\x61rgmax_param\x18\x17 \x01(\x0b\x32\x16.caffe.ArgMaxParameter\x12,\n\x0c\x63oncat_param\x18\t \x01(\x0b\x32\x16.caffe.ConcatParameter\x12?\n\x16\x63ontrastive_loss_param\x18( \x01(\x0b\x32\x1f.caffe.ContrastiveLossParameter\x12\x36\n\x11\x63onvolution_param\x18\n \x01(\x0b\x32\x1b.caffe.ConvolutionParameter\x12(\n\ndata_param\x18\x0b \x01(\x0b\x32\x14.caffe.DataParameter\x12.\n\rdropout_param\x18\x0c \x01(\x0b\x32\x17.caffe.DropoutParameter\x12\x33\n\x10\x64ummy_data_param\x18\x1a \x01(\x0b\x32\x19.caffe.DummyDataParameter\x12.\n\reltwise_param\x18\x18 \x01(\x0b\x32\x17.caffe.EltwiseParameter\x12&\n\texp_param\x18) \x01(\x0b\x32\x13.caffe.ExpParameter\x12\x31\n\x0fhdf5_data_param\x18\r \x01(\x0b\x32\x18.caffe.HDF5DataParameter\x12\x35\n\x11hdf5_output_param\x18\x0e \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\x12\x33\n\x10hinge_loss_param\x18\x1d \x01(\x0b\x32\x19.caffe.HingeLossParameter\x12\x33\n\x10image_data_param\x18\x0f \x01(\x0b\x32\x19.caffe.ImageDataParameter\x12\x39\n\x13infogain_loss_param\x18\x10 \x01(\x0b\x32\x1c.caffe.InfogainLossParameter\x12\x39\n\x13inner_product_param\x18\x11 \x01(\x0b\x32\x1c.caffe.InnerProductParameter\x12&\n\tlrn_param\x18\x12 \x01(\x0b\x32\x13.caffe.LRNParameter\x12\x35\n\x11memory_data_param\x18\x16 \x01(\x0b\x32\x1a.caffe.MemoryDataParameter\x12&\n\tmvn_param\x18\" \x01(\x0b\x32\x13.caffe.MVNParameter\x12.\n\rpooling_param\x18\x13 \x01(\x0b\x32\x17.caffe.PoolingParameter\x12*\n\x0bpower_param\x18\x15 \x01(\x0b\x32\x15.caffe.PowerParameter\x12(\n\nrelu_param\x18\x1e \x01(\x0b\x32\x14.caffe.ReLUParameter\x12.\n\rsigmoid_param\x18& \x01(\x0b\x32\x17.caffe.SigmoidParameter\x12.\n\rsoftmax_param\x18\' \x01(\x0b\x32\x17.caffe.SoftmaxParameter\x12*\n\x0bslice_param\x18\x1f \x01(\x0b\x32\x15.caffe.SliceParameter\x12(\n\ntanh_param\x18% \x01(\x0b\x32\x14.caffe.TanHParameter\x12\x32\n\x0fthreshold_param\x18\x19 \x01(\x0b\x32\x19.caffe.ThresholdParameter\x12\x35\n\x11window_data_param\x18\x14 \x01(\x0b\x32\x1a.caffe.WindowDataParameter\x12\x37\n\x0ftransform_param\x18$ \x01(\x0b\x32\x1e.caffe.TransformationParameter\x12(\n\nloss_param\x18* \x01(\x0b\x32\x14.caffe.LossParameter\x12&\n\x05layer\x18\x01 \x01(\x0b\x32\x17.caffe.V0LayerParameter\"\xd8\x04\n\tLayerType\x12\x08\n\x04NONE\x10\x00\x12\n\n\x06\x41\x42SVAL\x10#\x12\x0c\n\x08\x41\x43\x43URACY\x10\x01\x12\n\n\x06\x41RGMAX\x10\x1e\x12\x08\n\x04\x42NLL\x10\x02\x12\n\n\x06\x43ONCAT\x10\x03\x12\x14\n\x10\x43ONTRASTIVE_LOSS\x10%\x12\x0f\n\x0b\x43ONVOLUTION\x10\x04\x12\x08\n\x04\x44\x41TA\x10\x05\x12\x11\n\rDECONVOLUTION\x10\'\x12\x0b\n\x07\x44ROPOUT\x10\x06\x12\x0e\n\nDUMMY_DATA\x10 \x12\x12\n\x0e\x45UCLIDEAN_LOSS\x10\x07\x12\x0b\n\x07\x45LTWISE\x10\x19\x12\x07\n\x03\x45XP\x10&\x12\x0b\n\x07\x46LATTEN\x10\x08\x12\r\n\tHDF5_DATA\x10\t\x12\x0f\n\x0bHDF5_OUTPUT\x10\n\x12\x0e\n\nHINGE_LOSS\x10\x1c\x12\n\n\x06IM2COL\x10\x0b\x12\x0e\n\nIMAGE_DATA\x10\x0c\x12\x11\n\rINFOGAIN_LOSS\x10\r\x12\x11\n\rINNER_PRODUCT\x10\x0e\x12\x07\n\x03LRN\x10\x0f\x12\x0f\n\x0bMEMORY_DATA\x10\x1d\x12\x1d\n\x19MULTINOMIAL_LOGISTIC_LOSS\x10\x10\x12\x07\n\x03MVN\x10\"\x12\x0b\n\x07POOLING\x10\x11\x12\t\n\x05POWER\x10\x1a\x12\x08\n\x04RELU\x10\x12\x12\x0b\n\x07SIGMOID\x10\x13\x12\x1e\n\x1aSIGMOID_CROSS_ENTROPY_LOSS\x10\x1b\x12\x0b\n\x07SILENCE\x10$\x12\x0b\n\x07SOFTMAX\x10\x14\x12\x10\n\x0cSOFTMAX_LOSS\x10\x15\x12\t\n\x05SPLIT\x10\x16\x12\t\n\x05SLICE\x10!\x12\x08\n\x04TANH\x10\x17\x12\x0f\n\x0bWINDOW_DATA\x10\x18\x12\r\n\tTHRESHOLD\x10\x1f\"*\n\x0c\x44imCheckMode\x12\n\n\x06STRICT\x10\x00\x12\x0e\n\nPERMISSIVE\x10\x01\"\xfd\x07\n\x10V0LayerParameter\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04type\x18\x02 \x01(\t\x12\x12\n\nnum_output\x18\x03 \x01(\r\x12\x16\n\x08\x62iasterm\x18\x04 \x01(\x08:\x04true\x12-\n\rweight_filler\x18\x05 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x06 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x0e\n\x03pad\x18\x07 \x01(\r:\x01\x30\x12\x12\n\nkernelsize\x18\x08 \x01(\r\x12\x10\n\x05group\x18\t \x01(\r:\x01\x31\x12\x11\n\x06stride\x18\n \x01(\r:\x01\x31\x12\x35\n\x04pool\x18\x0b \x01(\x0e\x32\".caffe.V0LayerParameter.PoolMethod:\x03MAX\x12\x1a\n\rdropout_ratio\x18\x0c \x01(\x02:\x03\x30.5\x12\x15\n\nlocal_size\x18\r \x01(\r:\x01\x35\x12\x10\n\x05\x61lpha\x18\x0e \x01(\x02:\x01\x31\x12\x12\n\x04\x62\x65ta\x18\x0f \x01(\x02:\x04\x30.75\x12\x0c\n\x01k\x18\x16 \x01(\x02:\x01\x31\x12\x0e\n\x06source\x18\x10 \x01(\t\x12\x10\n\x05scale\x18\x11 \x01(\x02:\x01\x31\x12\x10\n\x08meanfile\x18\x12 \x01(\t\x12\x11\n\tbatchsize\x18\x13 \x01(\r\x12\x13\n\x08\x63ropsize\x18\x14 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x15 \x01(\x08:\x05\x66\x61lse\x12\x1f\n\x05\x62lobs\x18\x32 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x10\n\x08\x62lobs_lr\x18\x33 \x03(\x02\x12\x14\n\x0cweight_decay\x18\x34 \x03(\x02\x12\x14\n\trand_skip\x18\x35 \x01(\r:\x01\x30\x12\x1d\n\x10\x64\x65t_fg_threshold\x18\x36 \x01(\x02:\x03\x30.5\x12\x1d\n\x10\x64\x65t_bg_threshold\x18\x37 \x01(\x02:\x03\x30.5\x12\x1d\n\x0f\x64\x65t_fg_fraction\x18\x38 \x01(\x02:\x04\x30.25\x12\x1a\n\x0f\x64\x65t_context_pad\x18: \x01(\r:\x01\x30\x12\x1b\n\rdet_crop_mode\x18; \x01(\t:\x04warp\x12\x12\n\x07new_num\x18< \x01(\x05:\x01\x30\x12\x17\n\x0cnew_channels\x18= \x01(\x05:\x01\x30\x12\x15\n\nnew_height\x18> \x01(\x05:\x01\x30\x12\x14\n\tnew_width\x18? \x01(\x05:\x01\x30\x12\x1d\n\x0eshuffle_images\x18@ \x01(\x08:\x05\x66\x61lse\x12\x15\n\nconcat_dim\x18\x41 \x01(\r:\x01\x31\x12\x36\n\x11hdf5_output_param\x18\xe9\x07 \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"W\n\x0ePReLUParameter\x12&\n\x06\x66iller\x18\x01 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x1d\n\x0e\x63hannel_shared\x18\x02 \x01(\x08:\x05\x66\x61lse\"\x87\x01\n\x11ImagenetParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x12\n\nbatch_size\x18\x02 \x01(\r\x12\x13\n\x08prefetch\x18\x04 \x01(\r:\x01\x34\x12\x10\n\x08imageset\x18\x05 \x01(\t\x12\x16\n\x07shuffle\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x0f\n\x04step\x18\x07 \x01(\r:\x01\x31\")\n\x15SmoothL1LossParameter\x12\x10\n\x05sigma\x18\x01 \x01(\x02:\x01\x31\"H\n\x0cMPIParameter\x12\x0f\n\x04root\x18\x01 \x01(\r:\x01\x30\x12\x12\n\x07\x63omm_id\x18\x02 \x01(\x04:\x01\x30\x12\x13\n\x08group_id\x18\x03 \x01(\x04:\x01\x30\"!\n\x10PermuteParameter\x12\r\n\x05order\x18\x01 \x03(\r\"\x93\x01\n\x12NormalizeParameter\x12\x1c\n\x0e\x61\x63ross_spatial\x18\x01 \x01(\x08:\x04true\x12,\n\x0cscale_filler\x18\x02 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x1c\n\x0e\x63hannel_shared\x18\x03 \x01(\x08:\x04true\x12\x13\n\x03\x65ps\x18\x04 \x01(\x02:\x06\x31\x65-010\"_\n\x11ParallelParameter\x12\x16\n\x07shuffle\x18\x01 \x01(\x08:\x05\x66\x61lse\x12\x18\n\tnode_step\x18\x02 \x01(\x08:\x05\x66\x61lse\x12\x18\n\tpartition\x18\x03 \x01(\x08:\x05\x66\x61lse\"T\n\x11NNResizeParameter\x12\x1f\n\x05shape\x18\x01 \x01(\x0b\x32\x10.caffe.BlobShape\x12\x0e\n\x02\x66x\x18\x02 \x01(\x02:\x02-1\x12\x0e\n\x02\x66y\x18\x03 \x01(\x02:\x02-1\"\'\n\x13\x45xpandDimsParameter\x12\x10\n\x04\x61xis\x18\x01 \x01(\x05:\x02-1\"\xc8\x01\n\x11ProposalParameter\x12\x17\n\x0b\x66\x65\x61t_stride\x18\x01 \x01(\r:\x02\x31\x36\x12\x15\n\tbase_size\x18\x02 \x01(\r:\x02\x31\x36\x12\x14\n\x08min_size\x18\x03 \x01(\r:\x02\x31\x36\x12\r\n\x05ratio\x18\x04 \x03(\x02\x12\r\n\x05scale\x18\x05 \x03(\x02\x12\x1a\n\x0cpre_nms_topn\x18\x06 \x01(\r:\x04\x36\x30\x30\x30\x12\x1a\n\rpost_nms_topn\x18\x07 \x01(\r:\x03\x33\x30\x30\x12\x17\n\nnms_thresh\x18\x08 \x01(\x02:\x03\x30.7\"\xa2\x01\n\x14\x42\x61tchRenormParameter\x12\x18\n\x10use_global_stats\x18\x01 \x01(\x08\x12$\n\x17moving_average_fraction\x18\x02 \x01(\x02:\x03\x30.9\x12\x12\n\x03\x65ps\x18\x03 \x01(\x02:\x05\x30.001\x12\x10\n\x05r_max\x18\x04 \x01(\x02:\x01\x33\x12\x10\n\x05\x64_max\x18\x05 \x01(\x02:\x01\x35\x12\x12\n\x07t_delta\x18\x06 \x01(\x02:\x01\x31*\x1c\n\x05Phase\x12\t\n\x05TRAIN\x10\x00\x12\x08\n\x04TEST\x10\x01')
serialized_pb=_b('\n\x0b\x63\x61\x66\x66\x65.proto\x12\x05\x63\x61\x66\x66\x65\"\x1c\n\tBlobShape\x12\x0f\n\x03\x64im\x18\x01 \x03(\x03\x42\x02\x10\x01\"\xcc\x01\n\tBlobProto\x12\x1f\n\x05shape\x18\x07 \x01(\x0b\x32\x10.caffe.BlobShape\x12\x10\n\x04\x64\x61ta\x18\x05 \x03(\x02\x42\x02\x10\x01\x12\x10\n\x04\x64iff\x18\x06 \x03(\x02\x42\x02\x10\x01\x12\x17\n\x0b\x64ouble_data\x18\x08 \x03(\x01\x42\x02\x10\x01\x12\x17\n\x0b\x64ouble_diff\x18\t \x03(\x01\x42\x02\x10\x01\x12\x0e\n\x03num\x18\x01 \x01(\x05:\x01\x30\x12\x13\n\x08\x63hannels\x18\x02 \x01(\x05:\x01\x30\x12\x11\n\x06height\x18\x03 \x01(\x05:\x01\x30\x12\x10\n\x05width\x18\x04 \x01(\x05:\x01\x30\"2\n\x0f\x42lobProtoVector\x12\x1f\n\x05\x62lobs\x18\x01 \x03(\x0b\x32\x10.caffe.BlobProto\"\x81\x01\n\x05\x44\x61tum\x12\x10\n\x08\x63hannels\x18\x01 \x01(\x05\x12\x0e\n\x06height\x18\x02 \x01(\x05\x12\r\n\x05width\x18\x03 \x01(\x05\x12\x0c\n\x04\x64\x61ta\x18\x04 \x01(\x0c\x12\r\n\x05label\x18\x05 \x01(\x05\x12\x12\n\nfloat_data\x18\x06 \x03(\x02\x12\x16\n\x07\x65ncoded\x18\x07 \x01(\x08:\x05\x66\x61lse\"\x8a\x02\n\x0f\x46illerParameter\x12\x16\n\x04type\x18\x01 \x01(\t:\x08\x63onstant\x12\x10\n\x05value\x18\x02 \x01(\x02:\x01\x30\x12\x0e\n\x03min\x18\x03 \x01(\x02:\x01\x30\x12\x0e\n\x03max\x18\x04 \x01(\x02:\x01\x31\x12\x0f\n\x04mean\x18\x05 \x01(\x02:\x01\x30\x12\x0e\n\x03std\x18\x06 \x01(\x02:\x01\x31\x12\x12\n\x06sparse\x18\x07 \x01(\x05:\x02-1\x12\x42\n\rvariance_norm\x18\x08 \x01(\x0e\x32#.caffe.FillerParameter.VarianceNorm:\x06\x46\x41N_IN\"4\n\x0cVarianceNorm\x12\n\n\x06\x46\x41N_IN\x10\x00\x12\x0b\n\x07\x46\x41N_OUT\x10\x01\x12\x0b\n\x07\x41VERAGE\x10\x02\"\x8e\x02\n\x0cNetParameter\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\r\n\x05input\x18\x03 \x03(\t\x12%\n\x0binput_shape\x18\x08 \x03(\x0b\x32\x10.caffe.BlobShape\x12\x11\n\tinput_dim\x18\x04 \x03(\x05\x12\x1d\n\x0e\x66orce_backward\x18\x05 \x01(\x08:\x05\x66\x61lse\x12\x1e\n\x05state\x18\x06 \x01(\x0b\x32\x0f.caffe.NetState\x12\x19\n\ndebug_info\x18\x07 \x01(\x08:\x05\x66\x61lse\x12$\n\x05layer\x18\x64 \x03(\x0b\x32\x15.caffe.LayerParameter\x12\'\n\x06layers\x18\x02 \x03(\x0b\x32\x17.caffe.V1LayerParameter\"\xc9\n\n\x0fSolverParameter\x12\x0b\n\x03net\x18\x18 \x01(\t\x12&\n\tnet_param\x18\x19 \x01(\x0b\x32\x13.caffe.NetParameter\x12\x11\n\ttrain_net\x18\x01 \x01(\t\x12\x10\n\x08test_net\x18\x02 \x03(\t\x12,\n\x0ftrain_net_param\x18\x15 \x01(\x0b\x32\x13.caffe.NetParameter\x12+\n\x0etest_net_param\x18\x16 \x03(\x0b\x32\x13.caffe.NetParameter\x12$\n\x0btrain_state\x18\x1a \x01(\x0b\x32\x0f.caffe.NetState\x12#\n\ntest_state\x18\x1b \x03(\x0b\x32\x0f.caffe.NetState\x12\x11\n\ttest_iter\x18\x03 \x03(\x05\x12\x18\n\rtest_interval\x18\x04 \x01(\x05:\x01\x30\x12 \n\x11test_compute_loss\x18\x13 \x01(\x08:\x05\x66\x61lse\x12!\n\x13test_initialization\x18 \x01(\x08:\x04true\x12\x0f\n\x07\x62\x61se_lr\x18\x05 \x01(\x02\x12\x10\n\x08stage_lr\x18\x32 \x03(\x02\x12\x12\n\nstage_iter\x18\x33 \x03(\x05\x12\x0f\n\x07\x64isplay\x18\x06 \x01(\x05\x12\x17\n\x0c\x61verage_loss\x18! \x01(\x05:\x01\x31\x12\x10\n\x08max_iter\x18\x07 \x01(\x05\x12\x14\n\titer_size\x18$ \x01(\x05:\x01\x31\x12\x11\n\tlr_policy\x18\x08 \x01(\t\x12\r\n\x05gamma\x18\t \x01(\x02\x12\r\n\x05power\x18\n \x01(\x02\x12\x10\n\x08momentum\x18\x0b \x01(\x02\x12\x14\n\x0cweight_decay\x18\x0c \x01(\x02\x12\x1f\n\x13regularization_type\x18\x1d \x01(\t:\x02L2\x12\x10\n\x08stepsize\x18\r \x01(\x05\x12\x11\n\tstepvalue\x18\" \x03(\x05\x12\x1a\n\x0e\x63lip_gradients\x18# \x01(\x02:\x02-1\x12\x13\n\x08snapshot\x18\x0e \x01(\x05:\x01\x30\x12\x17\n\x0fsnapshot_prefix\x18\x0f \x01(\t\x12\x1c\n\rsnapshot_diff\x18\x10 \x01(\x08:\x05\x66\x61lse\x12K\n\x0fsnapshot_format\x18% \x01(\x0e\x32%.caffe.SolverParameter.SnapshotFormat:\x0b\x42INARYPROTO\x12;\n\x0bsolver_mode\x18\x11 \x01(\x0e\x32!.caffe.SolverParameter.SolverMode:\x03GPU\x12\x14\n\tdevice_id\x18\x12 \x01(\x05:\x01\x30\x12\x17\n\x0brandom_seed\x18\x14 \x01(\x03:\x02-1\x12\x11\n\x04type\x18( \x01(\t:\x03SGD\x12\x15\n\x05\x64\x65lta\x18\x1f \x01(\x02:\x06\x31\x65-008\x12\x18\n\tmomentum2\x18\' \x01(\x02:\x05\x30.999\x12\x17\n\trms_decay\x18& \x01(\x02:\x04\x30.99\x12\x19\n\ndebug_info\x18\x17 \x01(\x08:\x05\x66\x61lse\x12\"\n\x14snapshot_after_train\x18\x1c \x01(\x08:\x04true\x12;\n\x0bsolver_type\x18\x1e \x01(\x0e\x32!.caffe.SolverParameter.SolverType:\x03SGD\"+\n\x0eSnapshotFormat\x12\x08\n\x04HDF5\x10\x00\x12\x0f\n\x0b\x42INARYPROTO\x10\x01\"\x1e\n\nSolverMode\x12\x07\n\x03\x43PU\x10\x00\x12\x07\n\x03GPU\x10\x01\"U\n\nSolverType\x12\x07\n\x03SGD\x10\x00\x12\x0c\n\x08NESTEROV\x10\x01\x12\x0b\n\x07\x41\x44\x41GRAD\x10\x02\x12\x0b\n\x07RMSPROP\x10\x03\x12\x0c\n\x08\x41\x44\x41\x44\x45LTA\x10\x04\x12\x08\n\x04\x41\x44\x41M\x10\x05\"l\n\x0bSolverState\x12\x0c\n\x04iter\x18\x01 \x01(\x05\x12\x13\n\x0blearned_net\x18\x02 \x01(\t\x12!\n\x07history\x18\x03 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x17\n\x0c\x63urrent_step\x18\x04 \x01(\x05:\x01\x30\"N\n\x08NetState\x12!\n\x05phase\x18\x01 \x01(\x0e\x32\x0c.caffe.Phase:\x04TEST\x12\x10\n\x05level\x18\x02 \x01(\x05:\x01\x30\x12\r\n\x05stage\x18\x03 \x03(\t\"\x85\x01\n\x0cNetStateRule\x12\x1b\n\x05phase\x18\x01 \x01(\x0e\x32\x0c.caffe.Phase\x12\x11\n\tmin_level\x18\x02 \x01(\x05\x12\x11\n\tmax_level\x18\x03 \x01(\x05\x12\r\n\x05stage\x18\x04 \x03(\t\x12\x11\n\tnot_stage\x18\x05 \x03(\t\x12\x10\n\x08mpi_rank\x18\x06 \x03(\r\"\xa3\x01\n\tParamSpec\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x31\n\nshare_mode\x18\x02 \x01(\x0e\x32\x1d.caffe.ParamSpec.DimCheckMode\x12\x12\n\x07lr_mult\x18\x03 \x01(\x02:\x01\x31\x12\x15\n\ndecay_mult\x18\x04 \x01(\x02:\x01\x31\"*\n\x0c\x44imCheckMode\x12\n\n\x06STRICT\x10\x00\x12\x0e\n\nPERMISSIVE\x10\x01\"\x8f\x18\n\x0eLayerParameter\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04type\x18\x02 \x01(\t\x12\x0e\n\x06\x62ottom\x18\x03 \x03(\t\x12\x0b\n\x03top\x18\x04 \x03(\t\x12\x1b\n\x05phase\x18\n \x01(\x0e\x32\x0c.caffe.Phase\x12\x13\n\x0bloss_weight\x18\x05 \x03(\x02\x12\x1f\n\x05param\x18\x06 \x03(\x0b\x32\x10.caffe.ParamSpec\x12\x1f\n\x05\x62lobs\x18\x07 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x16\n\x0epropagate_down\x18\x0b \x03(\x08\x12$\n\x07include\x18\x08 \x03(\x0b\x32\x13.caffe.NetStateRule\x12$\n\x07\x65xclude\x18\t \x03(\x0b\x32\x13.caffe.NetStateRule\x12\x37\n\x0ftransform_param\x18\x64 \x01(\x0b\x32\x1e.caffe.TransformationParameter\x12(\n\nloss_param\x18\x65 \x01(\x0b\x32\x14.caffe.LossParameter\x12\x30\n\x0e\x61\x63\x63uracy_param\x18\x66 \x01(\x0b\x32\x18.caffe.AccuracyParameter\x12,\n\x0c\x61rgmax_param\x18g \x01(\x0b\x32\x16.caffe.ArgMaxParameter\x12\x34\n\x10\x62\x61tch_norm_param\x18\x8b\x01 \x01(\x0b\x32\x19.caffe.BatchNormParameter\x12)\n\nbias_param\x18\x8d\x01 \x01(\x0b\x32\x14.caffe.BiasParameter\x12,\n\x0c\x63oncat_param\x18h \x01(\x0b\x32\x16.caffe.ConcatParameter\x12?\n\x16\x63ontrastive_loss_param\x18i \x01(\x0b\x32\x1f.caffe.ContrastiveLossParameter\x12\x36\n\x11\x63onvolution_param\x18j \x01(\x0b\x32\x1b.caffe.ConvolutionParameter\x12)\n\ncrop_param\x18\x90\x01 \x01(\x0b\x32\x14.caffe.CropParameter\x12(\n\ndata_param\x18k \x01(\x0b\x32\x14.caffe.DataParameter\x12.\n\rdropout_param\x18l \x01(\x0b\x32\x17.caffe.DropoutParameter\x12\x33\n\x10\x64ummy_data_param\x18m \x01(\x0b\x32\x19.caffe.DummyDataParameter\x12.\n\reltwise_param\x18n \x01(\x0b\x32\x17.caffe.EltwiseParameter\x12\'\n\telu_param\x18\x8c\x01 \x01(\x0b\x32\x13.caffe.ELUParameter\x12+\n\x0b\x65mbed_param\x18\x89\x01 \x01(\x0b\x32\x15.caffe.EmbedParameter\x12&\n\texp_param\x18o \x01(\x0b\x32\x13.caffe.ExpParameter\x12/\n\rflatten_param\x18\x87\x01 \x01(\x0b\x32\x17.caffe.FlattenParameter\x12\x31\n\x0fhdf5_data_param\x18p \x01(\x0b\x32\x18.caffe.HDF5DataParameter\x12\x35\n\x11hdf5_output_param\x18q \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\x12\x33\n\x10hinge_loss_param\x18r \x01(\x0b\x32\x19.caffe.HingeLossParameter\x12\x33\n\x10image_data_param\x18s \x01(\x0b\x32\x19.caffe.ImageDataParameter\x12\x39\n\x13infogain_loss_param\x18t \x01(\x0b\x32\x1c.caffe.InfogainLossParameter\x12\x39\n\x13inner_product_param\x18u \x01(\x0b\x32\x1c.caffe.InnerProductParameter\x12+\n\x0binput_param\x18\x8f\x01 \x01(\x0b\x32\x15.caffe.InputParameter\x12\'\n\tlog_param\x18\x86\x01 \x01(\x0b\x32\x13.caffe.LogParameter\x12&\n\tlrn_param\x18v \x01(\x0b\x32\x13.caffe.LRNParameter\x12\x35\n\x11memory_data_param\x18w \x01(\x0b\x32\x1a.caffe.MemoryDataParameter\x12&\n\tmvn_param\x18x \x01(\x0b\x32\x13.caffe.MVNParameter\x12\x33\n\x0fparameter_param\x18\x91\x01 \x01(\x0b\x32\x19.caffe.ParameterParameter\x12.\n\rpooling_param\x18y \x01(\x0b\x32\x17.caffe.PoolingParameter\x12*\n\x0bpower_param\x18z \x01(\x0b\x32\x15.caffe.PowerParameter\x12+\n\x0bprelu_param\x18\x83\x01 \x01(\x0b\x32\x15.caffe.PReLUParameter\x12-\n\x0cpython_param\x18\x82\x01 \x01(\x0b\x32\x16.caffe.PythonParameter\x12\x33\n\x0freduction_param\x18\x88\x01 \x01(\x0b\x32\x19.caffe.ReductionParameter\x12(\n\nrelu_param\x18{ \x01(\x0b\x32\x14.caffe.ReLUParameter\x12/\n\rreshape_param\x18\x85\x01 \x01(\x0b\x32\x17.caffe.ReshapeParameter\x12+\n\x0bscale_param\x18\x8e\x01 \x01(\x0b\x32\x15.caffe.ScaleParameter\x12.\n\rsigmoid_param\x18| \x01(\x0b\x32\x17.caffe.SigmoidParameter\x12.\n\rsoftmax_param\x18} \x01(\x0b\x32\x17.caffe.SoftmaxParameter\x12\'\n\tspp_param\x18\x84\x01 \x01(\x0b\x32\x13.caffe.SPPParameter\x12*\n\x0bslice_param\x18~ \x01(\x0b\x32\x15.caffe.SliceParameter\x12(\n\ntanh_param\x18\x7f \x01(\x0b\x32\x14.caffe.TanHParameter\x12\x33\n\x0fthreshold_param\x18\x80\x01 \x01(\x0b\x32\x19.caffe.ThresholdParameter\x12)\n\ntile_param\x18\x8a\x01 \x01(\x0b\x32\x14.caffe.TileParameter\x12\x36\n\x11window_data_param\x18\x81\x01 \x01(\x0b\x32\x1a.caffe.WindowDataParameter\x12\x31\n\x0eimagenet_param\x18\x96\x01 \x01(\x0b\x32\x18.caffe.ImagenetParameter\x12\x36\n\x11roi_pooling_param\x18\x97\x01 \x01(\x0b\x32\x1a.caffe.ROIPoolingParameter\x12;\n\x14smooth_l1_loss_param\x18\x98\x01 \x01(\x0b\x32\x1c.caffe.SmoothL1LossParameter\x12\'\n\tmpi_param\x18\x99\x01 \x01(\x0b\x32\x13.caffe.MPIParameter\x12/\n\rpermute_param\x18\x9a\x01 \x01(\x0b\x32\x17.caffe.PermuteParameter\x12\x33\n\x0fnormalize_param\x18\x9b\x01 \x01(\x0b\x32\x19.caffe.NormalizeParameter\x12\x31\n\x0eparallel_param\x18\x9d\x01 \x01(\x0b\x32\x18.caffe.ParallelParameter\x12\x31\n\x0ennresize_param\x18\x9e\x01 \x01(\x0b\x32\x18.caffe.NNResizeParameter\x12\x36\n\x11\x65xpand_dims_param\x18\x9f\x01 \x01(\x0b\x32\x1a.caffe.ExpandDimsParameter\x12\x31\n\x0eproposal_param\x18\xa0\x01 \x01(\x0b\x32\x18.caffe.ProposalParameter\x12\x38\n\x12\x62\x61tch_renorm_param\x18\xa1\x01 \x01(\x0b\x32\x1b.caffe.BatchRenormParameter\"\x93\x02\n\x17TransformationParameter\x12\x10\n\x05scale\x18\x01 \x01(\x02:\x01\x31\x12\x15\n\x06mirror\x18\x02 \x01(\x08:\x05\x66\x61lse\x12\x14\n\tcrop_size\x18\x03 \x01(\r:\x01\x30\x12\x11\n\tmean_file\x18\x04 \x01(\t\x12\x12\n\nmean_value\x18\x05 \x03(\x02\x12\x1a\n\x0b\x66orce_color\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x19\n\nforce_gray\x18\x07 \x01(\x08:\x05\x66\x61lse\x12!\n\x12\x63olor_augmentation\x18\x08 \x01(\x08:\x05\x66\x61lse\x12\x1b\n\x10min_random_scale\x18\t \x01(\x02:\x01\x31\x12\x1b\n\x10max_random_scale\x18\n \x01(\x02:\x01\x31\"\xeb\x01\n\rLossParameter\x12\x14\n\x0cignore_label\x18\x01 \x01(\x05\x12\x44\n\rnormalization\x18\x03 \x01(\x0e\x32&.caffe.LossParameter.NormalizationMode:\x05VALID\x12\x11\n\tnormalize\x18\x02 \x01(\x08\x1a\'\n\x13\x45xpandDimsParameter\x12\x10\n\x04\x61xis\x18\x01 \x01(\x05:\x02-1\"B\n\x11NormalizationMode\x12\x08\n\x04\x46ULL\x10\x00\x12\t\n\x05VALID\x10\x01\x12\x0e\n\nBATCH_SIZE\x10\x02\x12\x08\n\x04NONE\x10\x03\"L\n\x11\x41\x63\x63uracyParameter\x12\x10\n\x05top_k\x18\x01 \x01(\r:\x01\x31\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x31\x12\x14\n\x0cignore_label\x18\x03 \x01(\x05\"M\n\x0f\x41rgMaxParameter\x12\x1a\n\x0bout_max_val\x18\x01 \x01(\x08:\x05\x66\x61lse\x12\x10\n\x05top_k\x18\x02 \x01(\r:\x01\x31\x12\x0c\n\x04\x61xis\x18\x03 \x01(\x05\"9\n\x0f\x43oncatParameter\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x31\x12\x15\n\nconcat_dim\x18\x01 \x01(\r:\x01\x31\"h\n\x12\x42\x61tchNormParameter\x12\x18\n\x10use_global_stats\x18\x01 \x01(\x08\x12$\n\x17moving_average_fraction\x18\x02 \x01(\x02:\x03\x30.9\x12\x12\n\x03\x65ps\x18\x03 \x01(\x02:\x05\x30.001\"]\n\rBiasParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\x13\n\x08num_axes\x18\x02 \x01(\x05:\x01\x31\x12&\n\x06\x66iller\x18\x03 \x01(\x0b\x32\x16.caffe.FillerParameter\"L\n\x18\x43ontrastiveLossParameter\x12\x11\n\x06margin\x18\x01 \x01(\x02:\x01\x31\x12\x1d\n\x0elegacy_version\x18\x02 \x01(\x08:\x05\x66\x61lse\"\xfc\x03\n\x14\x43onvolutionParameter\x12\x12\n\nnum_output\x18\x01 \x01(\r\x12\x17\n\tbias_term\x18\x02 \x01(\x08:\x04true\x12\x0b\n\x03pad\x18\x03 \x03(\r\x12\x13\n\x0bkernel_size\x18\x04 \x03(\r\x12\x0e\n\x06stride\x18\x06 \x03(\r\x12\x10\n\x08\x64ilation\x18\x12 \x03(\r\x12\x10\n\x05pad_h\x18\t \x01(\r:\x01\x30\x12\x10\n\x05pad_w\x18\n \x01(\r:\x01\x30\x12\x10\n\x08kernel_h\x18\x0b \x01(\r\x12\x10\n\x08kernel_w\x18\x0c \x01(\r\x12\x10\n\x08stride_h\x18\r \x01(\r\x12\x10\n\x08stride_w\x18\x0e \x01(\r\x12\x10\n\x05group\x18\x05 \x01(\r:\x01\x31\x12-\n\rweight_filler\x18\x07 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x08 \x01(\x0b\x32\x16.caffe.FillerParameter\x12;\n\x06\x65ngine\x18\x0f \x01(\x0e\x32\".caffe.ConvolutionParameter.Engine:\x07\x44\x45\x46\x41ULT\x12\x0f\n\x04\x61xis\x18\x10 \x01(\x05:\x01\x31\x12\x1e\n\x0f\x66orce_nd_im2col\x18\x11 \x01(\x08:\x05\x66\x61lse\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"0\n\rCropParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x32\x12\x0e\n\x06offset\x18\x02 \x03(\r\"\xa4\x02\n\rDataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x12\n\nbatch_size\x18\x04 \x01(\r\x12\x14\n\trand_skip\x18\x07 \x01(\r:\x01\x30\x12\x31\n\x07\x62\x61\x63kend\x18\x08 \x01(\x0e\x32\x17.caffe.DataParameter.DB:\x07LEVELDB\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x11\n\tmean_file\x18\x03 \x01(\t\x12\x14\n\tcrop_size\x18\x05 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\"\n\x13\x66orce_encoded_color\x18\t \x01(\x08:\x05\x66\x61lse\x12\x13\n\x08prefetch\x18\n \x01(\r:\x01\x35\"\x1b\n\x02\x44\x42\x12\x0b\n\x07LEVELDB\x10\x00\x12\x08\n\x04LMDB\x10\x01\"I\n\x10\x44ropoutParameter\x12\x1a\n\rdropout_ratio\x18\x01 \x01(\x02:\x03\x30.5\x12\x19\n\x0bscale_train\x18\x02 \x01(\x08:\x04true\"\xa0\x01\n\x12\x44ummyDataParameter\x12+\n\x0b\x64\x61ta_filler\x18\x01 \x03(\x0b\x32\x16.caffe.FillerParameter\x12\x1f\n\x05shape\x18\x06 \x03(\x0b\x32\x10.caffe.BlobShape\x12\x0b\n\x03num\x18\x02 \x03(\r\x12\x10\n\x08\x63hannels\x18\x03 \x03(\r\x12\x0e\n\x06height\x18\x04 \x03(\r\x12\r\n\x05width\x18\x05 \x03(\r\"\xa5\x01\n\x10\x45ltwiseParameter\x12\x39\n\toperation\x18\x01 \x01(\x0e\x32!.caffe.EltwiseParameter.EltwiseOp:\x03SUM\x12\r\n\x05\x63oeff\x18\x02 \x03(\x02\x12\x1e\n\x10stable_prod_grad\x18\x03 \x01(\x08:\x04true\"\'\n\tEltwiseOp\x12\x08\n\x04PROD\x10\x00\x12\x07\n\x03SUM\x10\x01\x12\x07\n\x03MAX\x10\x02\" \n\x0c\x45LUParameter\x12\x10\n\x05\x61lpha\x18\x01 \x01(\x02:\x01\x31\"\xac\x01\n\x0e\x45mbedParameter\x12\x12\n\nnum_output\x18\x01 \x01(\r\x12\x11\n\tinput_dim\x18\x02 \x01(\r\x12\x17\n\tbias_term\x18\x03 \x01(\x08:\x04true\x12-\n\rweight_filler\x18\x04 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x05 \x01(\x0b\x32\x16.caffe.FillerParameter\"D\n\x0c\x45xpParameter\x12\x10\n\x04\x62\x61se\x18\x01 \x01(\x02:\x02-1\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x10\n\x05shift\x18\x03 \x01(\x02:\x01\x30\"9\n\x10\x46lattenParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\x14\n\x08\x65nd_axis\x18\x02 \x01(\x05:\x02-1\"O\n\x11HDF5DataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x12\n\nbatch_size\x18\x02 \x01(\r\x12\x16\n\x07shuffle\x18\x03 \x01(\x08:\x05\x66\x61lse\"(\n\x13HDF5OutputParameter\x12\x11\n\tfile_name\x18\x01 \x01(\t\"^\n\x12HingeLossParameter\x12\x30\n\x04norm\x18\x01 \x01(\x0e\x32\x1e.caffe.HingeLossParameter.Norm:\x02L1\"\x16\n\x04Norm\x12\x06\n\x02L1\x10\x01\x12\x06\n\x02L2\x10\x02\"\x97\x02\n\x12ImageDataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x15\n\nbatch_size\x18\x04 \x01(\r:\x01\x31\x12\x14\n\trand_skip\x18\x07 \x01(\r:\x01\x30\x12\x16\n\x07shuffle\x18\x08 \x01(\x08:\x05\x66\x61lse\x12\x15\n\nnew_height\x18\t \x01(\r:\x01\x30\x12\x14\n\tnew_width\x18\n \x01(\r:\x01\x30\x12\x16\n\x08is_color\x18\x0b \x01(\x08:\x04true\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x11\n\tmean_file\x18\x03 \x01(\t\x12\x14\n\tcrop_size\x18\x05 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x15\n\x0broot_folder\x18\x0c \x01(\t:\x00\"\'\n\x15InfogainLossParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\"\xcb\x01\n\x15InnerProductParameter\x12\x12\n\nnum_output\x18\x01 \x01(\r\x12\x17\n\tbias_term\x18\x02 \x01(\x08:\x04true\x12-\n\rweight_filler\x18\x03 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x04 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x0f\n\x04\x61xis\x18\x05 \x01(\x05:\x01\x31\x12\x18\n\ttranspose\x18\x06 \x01(\x08:\x05\x66\x61lse\"1\n\x0eInputParameter\x12\x1f\n\x05shape\x18\x01 \x03(\x0b\x32\x10.caffe.BlobShape\"D\n\x0cLogParameter\x12\x10\n\x04\x62\x61se\x18\x01 \x01(\x02:\x02-1\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x10\n\x05shift\x18\x03 \x01(\x02:\x01\x30\"\xb8\x02\n\x0cLRNParameter\x12\x15\n\nlocal_size\x18\x01 \x01(\r:\x01\x35\x12\x10\n\x05\x61lpha\x18\x02 \x01(\x02:\x01\x31\x12\x12\n\x04\x62\x65ta\x18\x03 \x01(\x02:\x04\x30.75\x12\x44\n\x0bnorm_region\x18\x04 \x01(\x0e\x32\x1e.caffe.LRNParameter.NormRegion:\x0f\x41\x43ROSS_CHANNELS\x12\x0c\n\x01k\x18\x05 \x01(\x02:\x01\x31\x12\x33\n\x06\x65ngine\x18\x06 \x01(\x0e\x32\x1a.caffe.LRNParameter.Engine:\x07\x44\x45\x46\x41ULT\"5\n\nNormRegion\x12\x13\n\x0f\x41\x43ROSS_CHANNELS\x10\x00\x12\x12\n\x0eWITHIN_CHANNEL\x10\x01\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"\xbd\x01\n\x13MemoryDataParameter\x12\x12\n\nbatch_size\x18\x01 \x01(\r\x12\x10\n\x08\x63hannels\x18\x02 \x01(\r\x12\x0e\n\x06height\x18\x03 \x01(\r\x12\r\n\x05width\x18\x04 \x01(\r\x12;\n\x05\x64type\x18\x05 \x01(\x0e\x32#.caffe.MemoryDataParameter.DataType:\x07\x46LOAT32\"$\n\x08\x44\x61taType\x12\x0b\n\x07\x46LOAT32\x10\x00\x12\x0b\n\x07\x46LOAT16\x10\x01\"e\n\x0cMVNParameter\x12 \n\x12normalize_variance\x18\x01 \x01(\x08:\x04true\x12\x1e\n\x0f\x61\x63ross_channels\x18\x02 \x01(\x08:\x05\x66\x61lse\x12\x13\n\x03\x65ps\x18\x03 \x01(\x02:\x06\x31\x65-009\"5\n\x12ParameterParameter\x12\x1f\n\x05shape\x18\x01 \x01(\x0b\x32\x10.caffe.BlobShape\"\xa2\x03\n\x10PoolingParameter\x12\x35\n\x04pool\x18\x01 \x01(\x0e\x32\".caffe.PoolingParameter.PoolMethod:\x03MAX\x12\x0e\n\x03pad\x18\x04 \x01(\r:\x01\x30\x12\x10\n\x05pad_h\x18\t \x01(\r:\x01\x30\x12\x10\n\x05pad_w\x18\n \x01(\r:\x01\x30\x12\x13\n\x0bkernel_size\x18\x02 \x01(\r\x12\x10\n\x08kernel_h\x18\x05 \x01(\r\x12\x10\n\x08kernel_w\x18\x06 \x01(\r\x12\x11\n\x06stride\x18\x03 \x01(\r:\x01\x31\x12\x10\n\x08stride_h\x18\x07 \x01(\r\x12\x10\n\x08stride_w\x18\x08 \x01(\r\x12\x37\n\x06\x65ngine\x18\x0b \x01(\x0e\x32\x1e.caffe.PoolingParameter.Engine:\x07\x44\x45\x46\x41ULT\x12\x1d\n\x0eglobal_pooling\x18\x0c \x01(\x08:\x05\x66\x61lse\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"Y\n\x13ROIPoolingParameter\x12\x13\n\x08pooled_h\x18\x01 \x01(\r:\x01\x30\x12\x13\n\x08pooled_w\x18\x02 \x01(\r:\x01\x30\x12\x18\n\rspatial_scale\x18\x03 \x01(\x02:\x01\x31\"F\n\x0ePowerParameter\x12\x10\n\x05power\x18\x01 \x01(\x02:\x01\x31\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x10\n\x05shift\x18\x03 \x01(\x02:\x01\x30\"g\n\x0fPythonParameter\x12\x0e\n\x06module\x18\x01 \x01(\t\x12\r\n\x05layer\x18\x02 \x01(\t\x12\x13\n\tparam_str\x18\x03 \x01(\t:\x00\x12 \n\x11share_in_parallel\x18\x04 \x01(\x08:\x05\x66\x61lse\"\xad\x01\n\x12ReductionParameter\x12=\n\toperation\x18\x01 \x01(\x0e\x32%.caffe.ReductionParameter.ReductionOp:\x03SUM\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x30\x12\x10\n\x05\x63oeff\x18\x03 \x01(\x02:\x01\x31\"5\n\x0bReductionOp\x12\x07\n\x03SUM\x10\x01\x12\x08\n\x04\x41SUM\x10\x02\x12\t\n\x05SUMSQ\x10\x03\x12\x08\n\x04MEAN\x10\x04\"\x8d\x01\n\rReLUParameter\x12\x19\n\x0enegative_slope\x18\x01 \x01(\x02:\x01\x30\x12\x34\n\x06\x65ngine\x18\x02 \x01(\x0e\x32\x1b.caffe.ReLUParameter.Engine:\x07\x44\x45\x46\x41ULT\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"Z\n\x10ReshapeParameter\x12\x1f\n\x05shape\x18\x01 \x01(\x0b\x32\x10.caffe.BlobShape\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x30\x12\x14\n\x08num_axes\x18\x03 \x01(\x05:\x02-1\"\xa5\x01\n\x0eScaleParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\x13\n\x08num_axes\x18\x02 \x01(\x05:\x01\x31\x12&\n\x06\x66iller\x18\x03 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x18\n\tbias_term\x18\x04 \x01(\x08:\x05\x66\x61lse\x12+\n\x0b\x62ias_filler\x18\x05 \x01(\x0b\x32\x16.caffe.FillerParameter\"x\n\x10SigmoidParameter\x12\x37\n\x06\x65ngine\x18\x01 \x01(\x0e\x32\x1e.caffe.SigmoidParameter.Engine:\x07\x44\x45\x46\x41ULT\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"L\n\x0eSliceParameter\x12\x0f\n\x04\x61xis\x18\x03 \x01(\x05:\x01\x31\x12\x13\n\x0bslice_point\x18\x02 \x03(\r\x12\x14\n\tslice_dim\x18\x01 \x01(\r:\x01\x31\"\x89\x01\n\x10SoftmaxParameter\x12\x37\n\x06\x65ngine\x18\x01 \x01(\x0e\x32\x1e.caffe.SoftmaxParameter.Engine:\x07\x44\x45\x46\x41ULT\x12\x0f\n\x04\x61xis\x18\x02 \x01(\x05:\x01\x31\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"r\n\rTanHParameter\x12\x34\n\x06\x65ngine\x18\x01 \x01(\x0e\x32\x1b.caffe.TanHParameter.Engine:\x07\x44\x45\x46\x41ULT\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"T\n\rTileParameter\x12\x0f\n\x04\x61xis\x18\x01 \x01(\x05:\x01\x31\x12\r\n\x05tiles\x18\x02 \x01(\x05\x12#\n\tmultiples\x18\x03 \x01(\x0b\x32\x10.caffe.BlobShape\"*\n\x12ThresholdParameter\x12\x14\n\tthreshold\x18\x01 \x01(\x02:\x01\x30\"\xc1\x02\n\x13WindowDataParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x10\n\x05scale\x18\x02 \x01(\x02:\x01\x31\x12\x11\n\tmean_file\x18\x03 \x01(\t\x12\x12\n\nbatch_size\x18\x04 \x01(\r\x12\x14\n\tcrop_size\x18\x05 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x19\n\x0c\x66g_threshold\x18\x07 \x01(\x02:\x03\x30.5\x12\x19\n\x0c\x62g_threshold\x18\x08 \x01(\x02:\x03\x30.5\x12\x19\n\x0b\x66g_fraction\x18\t \x01(\x02:\x04\x30.25\x12\x16\n\x0b\x63ontext_pad\x18\n \x01(\r:\x01\x30\x12\x17\n\tcrop_mode\x18\x0b \x01(\t:\x04warp\x12\x1b\n\x0c\x63\x61\x63he_images\x18\x0c \x01(\x08:\x05\x66\x61lse\x12\x15\n\x0broot_folder\x18\r \x01(\t:\x00\"\xeb\x01\n\x0cSPPParameter\x12\x16\n\x0epyramid_height\x18\x01 \x01(\r\x12\x31\n\x04pool\x18\x02 \x01(\x0e\x32\x1e.caffe.SPPParameter.PoolMethod:\x03MAX\x12\x33\n\x06\x65ngine\x18\x06 \x01(\x0e\x32\x1a.caffe.SPPParameter.Engine:\x07\x44\x45\x46\x41ULT\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"+\n\x06\x45ngine\x12\x0b\n\x07\x44\x45\x46\x41ULT\x10\x00\x12\t\n\x05\x43\x41\x46\x46\x45\x10\x01\x12\t\n\x05\x43UDNN\x10\x02\"\xe0\x13\n\x10V1LayerParameter\x12\x0e\n\x06\x62ottom\x18\x02 \x03(\t\x12\x0b\n\x03top\x18\x03 \x03(\t\x12\x0c\n\x04name\x18\x04 \x01(\t\x12$\n\x07include\x18 \x03(\x0b\x32\x13.caffe.NetStateRule\x12$\n\x07\x65xclude\x18! \x03(\x0b\x32\x13.caffe.NetStateRule\x12/\n\x04type\x18\x05 \x01(\x0e\x32!.caffe.V1LayerParameter.LayerType\x12\x1f\n\x05\x62lobs\x18\x06 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x0e\n\x05param\x18\xe9\x07 \x03(\t\x12>\n\x0f\x62lob_share_mode\x18\xea\x07 \x03(\x0e\x32$.caffe.V1LayerParameter.DimCheckMode\x12\x10\n\x08\x62lobs_lr\x18\x07 \x03(\x02\x12\x14\n\x0cweight_decay\x18\x08 \x03(\x02\x12\x13\n\x0bloss_weight\x18# \x03(\x02\x12\x30\n\x0e\x61\x63\x63uracy_param\x18\x1b \x01(\x0b\x32\x18.caffe.AccuracyParameter\x12,\n\x0c\x61rgmax_param\x18\x17 \x01(\x0b\x32\x16.caffe.ArgMaxParameter\x12,\n\x0c\x63oncat_param\x18\t \x01(\x0b\x32\x16.caffe.ConcatParameter\x12?\n\x16\x63ontrastive_loss_param\x18( \x01(\x0b\x32\x1f.caffe.ContrastiveLossParameter\x12\x36\n\x11\x63onvolution_param\x18\n \x01(\x0b\x32\x1b.caffe.ConvolutionParameter\x12(\n\ndata_param\x18\x0b \x01(\x0b\x32\x14.caffe.DataParameter\x12.\n\rdropout_param\x18\x0c \x01(\x0b\x32\x17.caffe.DropoutParameter\x12\x33\n\x10\x64ummy_data_param\x18\x1a \x01(\x0b\x32\x19.caffe.DummyDataParameter\x12.\n\reltwise_param\x18\x18 \x01(\x0b\x32\x17.caffe.EltwiseParameter\x12&\n\texp_param\x18) \x01(\x0b\x32\x13.caffe.ExpParameter\x12\x31\n\x0fhdf5_data_param\x18\r \x01(\x0b\x32\x18.caffe.HDF5DataParameter\x12\x35\n\x11hdf5_output_param\x18\x0e \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\x12\x33\n\x10hinge_loss_param\x18\x1d \x01(\x0b\x32\x19.caffe.HingeLossParameter\x12\x33\n\x10image_data_param\x18\x0f \x01(\x0b\x32\x19.caffe.ImageDataParameter\x12\x39\n\x13infogain_loss_param\x18\x10 \x01(\x0b\x32\x1c.caffe.InfogainLossParameter\x12\x39\n\x13inner_product_param\x18\x11 \x01(\x0b\x32\x1c.caffe.InnerProductParameter\x12&\n\tlrn_param\x18\x12 \x01(\x0b\x32\x13.caffe.LRNParameter\x12\x35\n\x11memory_data_param\x18\x16 \x01(\x0b\x32\x1a.caffe.MemoryDataParameter\x12&\n\tmvn_param\x18\" \x01(\x0b\x32\x13.caffe.MVNParameter\x12.\n\rpooling_param\x18\x13 \x01(\x0b\x32\x17.caffe.PoolingParameter\x12*\n\x0bpower_param\x18\x15 \x01(\x0b\x32\x15.caffe.PowerParameter\x12(\n\nrelu_param\x18\x1e \x01(\x0b\x32\x14.caffe.ReLUParameter\x12.\n\rsigmoid_param\x18& \x01(\x0b\x32\x17.caffe.SigmoidParameter\x12.\n\rsoftmax_param\x18\' \x01(\x0b\x32\x17.caffe.SoftmaxParameter\x12*\n\x0bslice_param\x18\x1f \x01(\x0b\x32\x15.caffe.SliceParameter\x12(\n\ntanh_param\x18% \x01(\x0b\x32\x14.caffe.TanHParameter\x12\x32\n\x0fthreshold_param\x18\x19 \x01(\x0b\x32\x19.caffe.ThresholdParameter\x12\x35\n\x11window_data_param\x18\x14 \x01(\x0b\x32\x1a.caffe.WindowDataParameter\x12\x37\n\x0ftransform_param\x18$ \x01(\x0b\x32\x1e.caffe.TransformationParameter\x12(\n\nloss_param\x18* \x01(\x0b\x32\x14.caffe.LossParameter\x12&\n\x05layer\x18\x01 \x01(\x0b\x32\x17.caffe.V0LayerParameter\"\xd8\x04\n\tLayerType\x12\x08\n\x04NONE\x10\x00\x12\n\n\x06\x41\x42SVAL\x10#\x12\x0c\n\x08\x41\x43\x43URACY\x10\x01\x12\n\n\x06\x41RGMAX\x10\x1e\x12\x08\n\x04\x42NLL\x10\x02\x12\n\n\x06\x43ONCAT\x10\x03\x12\x14\n\x10\x43ONTRASTIVE_LOSS\x10%\x12\x0f\n\x0b\x43ONVOLUTION\x10\x04\x12\x08\n\x04\x44\x41TA\x10\x05\x12\x11\n\rDECONVOLUTION\x10\'\x12\x0b\n\x07\x44ROPOUT\x10\x06\x12\x0e\n\nDUMMY_DATA\x10 \x12\x12\n\x0e\x45UCLIDEAN_LOSS\x10\x07\x12\x0b\n\x07\x45LTWISE\x10\x19\x12\x07\n\x03\x45XP\x10&\x12\x0b\n\x07\x46LATTEN\x10\x08\x12\r\n\tHDF5_DATA\x10\t\x12\x0f\n\x0bHDF5_OUTPUT\x10\n\x12\x0e\n\nHINGE_LOSS\x10\x1c\x12\n\n\x06IM2COL\x10\x0b\x12\x0e\n\nIMAGE_DATA\x10\x0c\x12\x11\n\rINFOGAIN_LOSS\x10\r\x12\x11\n\rINNER_PRODUCT\x10\x0e\x12\x07\n\x03LRN\x10\x0f\x12\x0f\n\x0bMEMORY_DATA\x10\x1d\x12\x1d\n\x19MULTINOMIAL_LOGISTIC_LOSS\x10\x10\x12\x07\n\x03MVN\x10\"\x12\x0b\n\x07POOLING\x10\x11\x12\t\n\x05POWER\x10\x1a\x12\x08\n\x04RELU\x10\x12\x12\x0b\n\x07SIGMOID\x10\x13\x12\x1e\n\x1aSIGMOID_CROSS_ENTROPY_LOSS\x10\x1b\x12\x0b\n\x07SILENCE\x10$\x12\x0b\n\x07SOFTMAX\x10\x14\x12\x10\n\x0cSOFTMAX_LOSS\x10\x15\x12\t\n\x05SPLIT\x10\x16\x12\t\n\x05SLICE\x10!\x12\x08\n\x04TANH\x10\x17\x12\x0f\n\x0bWINDOW_DATA\x10\x18\x12\r\n\tTHRESHOLD\x10\x1f\"*\n\x0c\x44imCheckMode\x12\n\n\x06STRICT\x10\x00\x12\x0e\n\nPERMISSIVE\x10\x01\"\xfd\x07\n\x10V0LayerParameter\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04type\x18\x02 \x01(\t\x12\x12\n\nnum_output\x18\x03 \x01(\r\x12\x16\n\x08\x62iasterm\x18\x04 \x01(\x08:\x04true\x12-\n\rweight_filler\x18\x05 \x01(\x0b\x32\x16.caffe.FillerParameter\x12+\n\x0b\x62ias_filler\x18\x06 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x0e\n\x03pad\x18\x07 \x01(\r:\x01\x30\x12\x12\n\nkernelsize\x18\x08 \x01(\r\x12\x10\n\x05group\x18\t \x01(\r:\x01\x31\x12\x11\n\x06stride\x18\n \x01(\r:\x01\x31\x12\x35\n\x04pool\x18\x0b \x01(\x0e\x32\".caffe.V0LayerParameter.PoolMethod:\x03MAX\x12\x1a\n\rdropout_ratio\x18\x0c \x01(\x02:\x03\x30.5\x12\x15\n\nlocal_size\x18\r \x01(\r:\x01\x35\x12\x10\n\x05\x61lpha\x18\x0e \x01(\x02:\x01\x31\x12\x12\n\x04\x62\x65ta\x18\x0f \x01(\x02:\x04\x30.75\x12\x0c\n\x01k\x18\x16 \x01(\x02:\x01\x31\x12\x0e\n\x06source\x18\x10 \x01(\t\x12\x10\n\x05scale\x18\x11 \x01(\x02:\x01\x31\x12\x10\n\x08meanfile\x18\x12 \x01(\t\x12\x11\n\tbatchsize\x18\x13 \x01(\r\x12\x13\n\x08\x63ropsize\x18\x14 \x01(\r:\x01\x30\x12\x15\n\x06mirror\x18\x15 \x01(\x08:\x05\x66\x61lse\x12\x1f\n\x05\x62lobs\x18\x32 \x03(\x0b\x32\x10.caffe.BlobProto\x12\x10\n\x08\x62lobs_lr\x18\x33 \x03(\x02\x12\x14\n\x0cweight_decay\x18\x34 \x03(\x02\x12\x14\n\trand_skip\x18\x35 \x01(\r:\x01\x30\x12\x1d\n\x10\x64\x65t_fg_threshold\x18\x36 \x01(\x02:\x03\x30.5\x12\x1d\n\x10\x64\x65t_bg_threshold\x18\x37 \x01(\x02:\x03\x30.5\x12\x1d\n\x0f\x64\x65t_fg_fraction\x18\x38 \x01(\x02:\x04\x30.25\x12\x1a\n\x0f\x64\x65t_context_pad\x18: \x01(\r:\x01\x30\x12\x1b\n\rdet_crop_mode\x18; \x01(\t:\x04warp\x12\x12\n\x07new_num\x18< \x01(\x05:\x01\x30\x12\x17\n\x0cnew_channels\x18= \x01(\x05:\x01\x30\x12\x15\n\nnew_height\x18> \x01(\x05:\x01\x30\x12\x14\n\tnew_width\x18? \x01(\x05:\x01\x30\x12\x1d\n\x0eshuffle_images\x18@ \x01(\x08:\x05\x66\x61lse\x12\x15\n\nconcat_dim\x18\x41 \x01(\r:\x01\x31\x12\x36\n\x11hdf5_output_param\x18\xe9\x07 \x01(\x0b\x32\x1a.caffe.HDF5OutputParameter\".\n\nPoolMethod\x12\x07\n\x03MAX\x10\x00\x12\x07\n\x03\x41VE\x10\x01\x12\x0e\n\nSTOCHASTIC\x10\x02\"W\n\x0ePReLUParameter\x12&\n\x06\x66iller\x18\x01 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x1d\n\x0e\x63hannel_shared\x18\x02 \x01(\x08:\x05\x66\x61lse\"\x87\x01\n\x11ImagenetParameter\x12\x0e\n\x06source\x18\x01 \x01(\t\x12\x12\n\nbatch_size\x18\x02 \x01(\r\x12\x13\n\x08prefetch\x18\x04 \x01(\r:\x01\x34\x12\x10\n\x08imageset\x18\x05 \x01(\t\x12\x16\n\x07shuffle\x18\x06 \x01(\x08:\x05\x66\x61lse\x12\x0f\n\x04step\x18\x07 \x01(\r:\x01\x31\")\n\x15SmoothL1LossParameter\x12\x10\n\x05sigma\x18\x01 \x01(\x02:\x01\x31\"H\n\x0cMPIParameter\x12\x0f\n\x04root\x18\x01 \x01(\r:\x01\x30\x12\x12\n\x07\x63omm_id\x18\x02 \x01(\x04:\x01\x30\x12\x13\n\x08group_id\x18\x03 \x01(\x04:\x01\x30\"!\n\x10PermuteParameter\x12\r\n\x05order\x18\x01 \x03(\r\"\x93\x01\n\x12NormalizeParameter\x12\x1c\n\x0e\x61\x63ross_spatial\x18\x01 \x01(\x08:\x04true\x12,\n\x0cscale_filler\x18\x02 \x01(\x0b\x32\x16.caffe.FillerParameter\x12\x1c\n\x0e\x63hannel_shared\x18\x03 \x01(\x08:\x04true\x12\x13\n\x03\x65ps\x18\x04 \x01(\x02:\x06\x31\x65-010\"_\n\x11ParallelParameter\x12\x16\n\x07shuffle\x18\x01 \x01(\x08:\x05\x66\x61lse\x12\x18\n\tnode_step\x18\x02 \x01(\x08:\x05\x66\x61lse\x12\x18\n\tpartition\x18\x03 \x01(\x08:\x05\x66\x61lse\"T\n\x11NNResizeParameter\x12\x1f\n\x05shape\x18\x01 \x01(\x0b\x32\x10.caffe.BlobShape\x12\x0e\n\x02\x66x\x18\x02 \x01(\x02:\x02-1\x12\x0e\n\x02\x66y\x18\x03 \x01(\x02:\x02-1\"\'\n\x13\x45xpandDimsParameter\x12\x10\n\x04\x61xis\x18\x01 \x01(\x05:\x02-1\"\xc8\x01\n\x11ProposalParameter\x12\x17\n\x0b\x66\x65\x61t_stride\x18\x01 \x01(\r:\x02\x31\x36\x12\x15\n\tbase_size\x18\x02 \x01(\r:\x02\x31\x36\x12\x14\n\x08min_size\x18\x03 \x01(\r:\x02\x31\x36\x12\r\n\x05ratio\x18\x04 \x03(\x02\x12\r\n\x05scale\x18\x05 \x03(\x02\x12\x1a\n\x0cpre_nms_topn\x18\x06 \x01(\r:\x04\x36\x30\x30\x30\x12\x1a\n\rpost_nms_topn\x18\x07 \x01(\r:\x03\x33\x30\x30\x12\x17\n\nnms_thresh\x18\x08 \x01(\x02:\x03\x30.7\"\xa2\x01\n\x14\x42\x61tchRenormParameter\x12\x18\n\x10use_global_stats\x18\x01 \x01(\x08\x12$\n\x17moving_average_fraction\x18\x02 \x01(\x02:\x03\x30.9\x12\x12\n\x03\x65ps\x18\x03 \x01(\x02:\x05\x30.001\x12\x10\n\x05r_max\x18\x04 \x01(\x02:\x01\x33\x12\x10\n\x05\x64_max\x18\x05 \x01(\x02:\x01\x35\x12\x12\n\x07t_delta\x18\x06 \x01(\x02:\x01\x31*\x1c\n\x05Phase\x12\t\n\x05TRAIN\x10\x00\x12\x08\n\x04TEST\x10\x01')
)
_sym_db.RegisterFileDescriptor(DESCRIPTOR)
......@@ -40,8 +40,8 @@ _PHASE = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=17162,
serialized_end=17190,
serialized_start=17161,
serialized_end=17189,
)
_sym_db.RegisterEnumDescriptor(_PHASE)
......@@ -253,8 +253,8 @@ _DATAPARAMETER_DB = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=7762,
serialized_end=7789,
serialized_start=7761,
serialized_end=7788,
)
_sym_db.RegisterEnumDescriptor(_DATAPARAMETER_DB)
......@@ -279,8 +279,8 @@ _ELTWISEPARAMETER_ELTWISEOP = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=8156,
serialized_end=8195,
serialized_start=8155,
serialized_end=8194,
)
_sym_db.RegisterEnumDescriptor(_ELTWISEPARAMETER_ELTWISEOP)
......@@ -301,8 +301,8 @@ _HINGELOSSPARAMETER_NORM = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=8730,
serialized_end=8752,
serialized_start=8729,
serialized_end=8751,
)
_sym_db.RegisterEnumDescriptor(_HINGELOSSPARAMETER_NORM)
......@@ -323,8 +323,8 @@ _LRNPARAMETER_NORMREGION = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=9619,
serialized_end=9672,
serialized_start=9618,
serialized_end=9671,
)
_sym_db.RegisterEnumDescriptor(_LRNPARAMETER_NORMREGION)
......@@ -371,8 +371,8 @@ _MEMORYDATAPARAMETER_DATATYPE = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=9873,
serialized_end=9909,
serialized_start=9872,
serialized_end=9908,
)
_sym_db.RegisterEnumDescriptor(_MEMORYDATAPARAMETER_DATATYPE)
......@@ -397,8 +397,8 @@ _POOLINGPARAMETER_POOLMETHOD = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=10397,
serialized_end=10443,
serialized_start=10396,
serialized_end=10442,
)
_sym_db.RegisterEnumDescriptor(_POOLINGPARAMETER_POOLMETHOD)
......@@ -453,8 +453,8 @@ _REDUCTIONPARAMETER_REDUCTIONOP = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=10879,
serialized_end=10932,
serialized_start=10878,
serialized_end=10931,
)
_sym_db.RegisterEnumDescriptor(_REDUCTIONPARAMETER_REDUCTIONOP)
......@@ -583,8 +583,8 @@ _SPPPARAMETER_POOLMETHOD = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=10397,
serialized_end=10443,
serialized_start=10396,
serialized_end=10442,
)
_sym_db.RegisterEnumDescriptor(_SPPPARAMETER_POOLMETHOD)
......@@ -783,8 +783,8 @@ _V1LAYERPARAMETER_LAYERTYPE = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=14371,
serialized_end=14971,
serialized_start=14370,
serialized_end=14970,
)
_sym_db.RegisterEnumDescriptor(_V1LAYERPARAMETER_LAYERTYPE)
......@@ -831,8 +831,8 @@ _V0LAYERPARAMETER_POOLMETHOD = _descriptor.EnumDescriptor(
],
containing_type=None,
options=None,
serialized_start=10397,
serialized_end=10443,
serialized_start=10396,
serialized_end=10442,
)
_sym_db.RegisterEnumDescriptor(_V0LAYERPARAMETER_POOLMETHOD)
......@@ -2929,7 +2929,7 @@ _DATAPARAMETER = _descriptor.Descriptor(
_descriptor.FieldDescriptor(
name='prefetch', full_name='caffe.DataParameter.prefetch', index=9,
number=10, type=13, cpp_type=3, label=1,
has_default_value=True, default_value=40,
has_default_value=True, default_value=5,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
......@@ -2946,7 +2946,7 @@ _DATAPARAMETER = _descriptor.Descriptor(
oneofs=[
],
serialized_start=7496,
serialized_end=7789,
serialized_end=7788,
)
......@@ -2982,8 +2982,8 @@ _DROPOUTPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=7791,
serialized_end=7864,
serialized_start=7790,
serialized_end=7863,
)
......@@ -3047,8 +3047,8 @@ _DUMMYDATAPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=7867,
serialized_end=8027,
serialized_start=7866,
serialized_end=8026,
)
......@@ -3092,8 +3092,8 @@ _ELTWISEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8030,
serialized_end=8195,
serialized_start=8029,
serialized_end=8194,
)
......@@ -3122,8 +3122,8 @@ _ELUPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8197,
serialized_end=8229,
serialized_start=8196,
serialized_end=8228,
)
......@@ -3180,8 +3180,8 @@ _EMBEDPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8232,
serialized_end=8404,
serialized_start=8231,
serialized_end=8403,
)
......@@ -3224,8 +3224,8 @@ _EXPPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8406,
serialized_end=8474,
serialized_start=8405,
serialized_end=8473,
)
......@@ -3261,8 +3261,8 @@ _FLATTENPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8476,
serialized_end=8533,
serialized_start=8475,
serialized_end=8532,
)
......@@ -3305,8 +3305,8 @@ _HDF5DATAPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8535,
serialized_end=8614,
serialized_start=8534,
serialized_end=8613,
)
......@@ -3335,8 +3335,8 @@ _HDF5OUTPUTPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8616,
serialized_end=8656,
serialized_start=8615,
serialized_end=8655,
)
......@@ -3366,8 +3366,8 @@ _HINGELOSSPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8658,
serialized_end=8752,
serialized_start=8657,
serialized_end=8751,
)
......@@ -3473,8 +3473,8 @@ _IMAGEDATAPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=8755,
serialized_end=9034,
serialized_start=8754,
serialized_end=9033,
)
......@@ -3503,8 +3503,8 @@ _INFOGAINLOSSPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9036,
serialized_end=9075,
serialized_start=9035,
serialized_end=9074,
)
......@@ -3568,8 +3568,8 @@ _INNERPRODUCTPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9078,
serialized_end=9281,
serialized_start=9077,
serialized_end=9280,
)
......@@ -3598,8 +3598,8 @@ _INPUTPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9283,
serialized_end=9332,
serialized_start=9282,
serialized_end=9331,
)
......@@ -3642,8 +3642,8 @@ _LOGPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9334,
serialized_end=9402,
serialized_start=9333,
serialized_end=9401,
)
......@@ -3709,8 +3709,8 @@ _LRNPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9405,
serialized_end=9717,
serialized_start=9404,
serialized_end=9716,
)
......@@ -3768,8 +3768,8 @@ _MEMORYDATAPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9720,
serialized_end=9909,
serialized_start=9719,
serialized_end=9908,
)
......@@ -3812,8 +3812,8 @@ _MVNPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=9911,
serialized_end=10012,
serialized_start=9910,
serialized_end=10011,
)
......@@ -3842,8 +3842,8 @@ _PARAMETERPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10014,
serialized_end=10067,
serialized_start=10013,
serialized_end=10066,
)
......@@ -3951,8 +3951,8 @@ _POOLINGPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10070,
serialized_end=10488,
serialized_start=10069,
serialized_end=10487,
)
......@@ -3995,8 +3995,8 @@ _ROIPOOLINGPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10490,
serialized_end=10579,
serialized_start=10489,
serialized_end=10578,
)
......@@ -4039,8 +4039,8 @@ _POWERPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10581,
serialized_end=10651,
serialized_start=10580,
serialized_end=10650,
)
......@@ -4090,8 +4090,8 @@ _PYTHONPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10653,
serialized_end=10756,
serialized_start=10652,
serialized_end=10755,
)
......@@ -4135,8 +4135,8 @@ _REDUCTIONPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10759,
serialized_end=10932,
serialized_start=10758,
serialized_end=10931,
)
......@@ -4173,8 +4173,8 @@ _RELUPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=10935,
serialized_end=11076,
serialized_start=10934,
serialized_end=11075,
)
......@@ -4217,8 +4217,8 @@ _RESHAPEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11078,
serialized_end=11168,
serialized_start=11077,
serialized_end=11167,
)
......@@ -4275,8 +4275,8 @@ _SCALEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11171,
serialized_end=11336,
serialized_start=11170,
serialized_end=11335,
)
......@@ -4306,8 +4306,8 @@ _SIGMOIDPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11338,
serialized_end=11458,
serialized_start=11337,
serialized_end=11457,
)
......@@ -4350,8 +4350,8 @@ _SLICEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11460,
serialized_end=11536,
serialized_start=11459,
serialized_end=11535,
)
......@@ -4388,8 +4388,8 @@ _SOFTMAXPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11539,
serialized_end=11676,
serialized_start=11538,
serialized_end=11675,
)
......@@ -4419,8 +4419,8 @@ _TANHPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11678,
serialized_end=11792,
serialized_start=11677,
serialized_end=11791,
)
......@@ -4463,8 +4463,8 @@ _TILEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11794,
serialized_end=11878,
serialized_start=11793,
serialized_end=11877,
)
......@@ -4493,8 +4493,8 @@ _THRESHOLDPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11880,
serialized_end=11922,
serialized_start=11879,
serialized_end=11921,
)
......@@ -4607,8 +4607,8 @@ _WINDOWDATAPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=11925,
serialized_end=12246,
serialized_start=11924,
serialized_end=12245,
)
......@@ -4653,8 +4653,8 @@ _SPPPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=12249,
serialized_end=12484,
serialized_start=12248,
serialized_end=12483,
)
......@@ -4979,8 +4979,8 @@ _V1LAYERPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=12487,
serialized_end=15015,
serialized_start=12486,
serialized_end=15014,
)
......@@ -5269,8 +5269,8 @@ _V0LAYERPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=15018,
serialized_end=16039,
serialized_start=15017,
serialized_end=16038,
)
......@@ -5306,8 +5306,8 @@ _PRELUPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16041,
serialized_end=16128,
serialized_start=16040,
serialized_end=16127,
)
......@@ -5371,8 +5371,8 @@ _IMAGENETPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16131,
serialized_end=16266,
serialized_start=16130,
serialized_end=16265,
)
......@@ -5401,8 +5401,8 @@ _SMOOTHL1LOSSPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16268,
serialized_end=16309,
serialized_start=16267,
serialized_end=16308,
)
......@@ -5445,8 +5445,8 @@ _MPIPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16311,
serialized_end=16383,
serialized_start=16310,
serialized_end=16382,
)
......@@ -5475,8 +5475,8 @@ _PERMUTEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16385,
serialized_end=16418,
serialized_start=16384,
serialized_end=16417,
)
......@@ -5526,8 +5526,8 @@ _NORMALIZEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16421,
serialized_end=16568,
serialized_start=16420,
serialized_end=16567,
)
......@@ -5570,8 +5570,8 @@ _PARALLELPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16570,
serialized_end=16665,
serialized_start=16569,
serialized_end=16664,
)
......@@ -5614,8 +5614,8 @@ _NNRESIZEPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16667,
serialized_end=16751,
serialized_start=16666,
serialized_end=16750,
)
......@@ -5723,8 +5723,8 @@ _PROPOSALPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16795,
serialized_end=16995,
serialized_start=16794,
serialized_end=16994,
)
......@@ -5788,8 +5788,8 @@ _BATCHRENORMPARAMETER = _descriptor.Descriptor(
extension_ranges=[],
oneofs=[
],
serialized_start=16998,
serialized_end=17160,
serialized_start=16997,
serialized_end=17159,
)
_BLOBPROTO.fields_by_name['shape'].message_type = _BLOBSHAPE
......
Dragon/src/core/context.cc
......@@ -10,7 +10,7 @@ CUDAObject CUDAContext::cuda_object_;
// cpu <- gpu
template<> void CPUContext::Memcpy<CPUContext, CUDAContext>(
size_t nbytes, void* dst, const void* src){
size_t nbytes, void* dst, const void* src) {
#ifdef WITH_CUDA
CUDAContext ctx(POINTER_DEVICE(src));
ctx.Memcpy<CPUContext, CUDAContext>(nbytes, dst, src);
......@@ -21,7 +21,7 @@ template<> void CPUContext::Memcpy<CPUContext, CUDAContext>(
// gpu <- cpu
template<> void CPUContext::Memcpy<CUDAContext, CPUContext>(
size_t nbytes, void* dst, const void* src){
size_t nbytes, void* dst, const void* src) {
#ifdef WITH_CUDA
CUDAContext ctx(POINTER_DEVICE(dst));
ctx.Memcpy<CUDAContext, CPUContext>(nbytes, dst, src);
......
Dragon/src/core/graph.cc
......@@ -151,13 +151,13 @@ GraphDef Graph::Prune(const GraphDef& graph_def) {
OperatorDef op_def;
op_def.CopyFrom(graph_def.op(it));
// handle inputs
for (int i = 0; i < graph_def.op(it).input_size(); i++){
for (int i = 0; i < graph_def.op(it).input_size(); i++) {
string input = graph_def.op(it).input(i);
if (!colored_[input] || !outputs.count(input))
*op_def.mutable_input(i) = "ignore";
}
// handle outputs
for (int i = 0; i < graph_def.op(it).output_size(); i++){
for (int i = 0; i < graph_def.op(it).output_size(); i++) {
string output = graph_def.op(it).output(i);
if (!colored_[output]) *op_def.mutable_output(i) = "ignore";
else outputs.insert(op_def.output(i));
......
Dragon/src/core/graph_gradient.cc
......@@ -23,7 +23,7 @@ CheckTuple GraphGradientMaker::CheckMissingGrad(OperatorDef* forward_op) {
inputs_to_grads_[output] = g_output;
// consider generate virtual grad
else if (targets_set_.count(output) && g_output != "ignore"){
else if (targets_set_.count(output) && g_output != "ignore") {
gen_grads.push_back({ output, idx });
inputs_to_grads_[output] = g_output;
}
......@@ -50,7 +50,7 @@ GraphDef GraphGradientMaker::Make() {
Set<string> all_split_grads;
// PLAY for the forward
for (auto& op : forward_def_.op()){
for (auto& op : forward_def_.op()) {
if (NoGradientRegistry()->Has(op.type())) continue;
for (auto& input : op.input()) inputs_count[input]++;
}
......@@ -73,17 +73,17 @@ GraphDef GraphGradientMaker::Make() {
Gradient grad = MakeGradientForOp(*op, g_outputs);
// replace terms
for (auto& g_op : grad.ops){
for (auto& g_op : grad.ops) {
g_op.set_name(GetOperatorName());
for (int i = 0; i < g_op.input_size(); i++){
for (int i = 0; i < g_op.input_size(); i++) {
string* input = g_op.mutable_input(i);
if (terms_.count(*input)) *input = terms_[*input];
}
for (int i = 0; i < g_op.output_size(); i++){
for (int i = 0; i < g_op.output_size(); i++) {
string* output = g_op.mutable_output(i);
if (terms_.count(*output)) *output = terms_[*output];
}
for (int i = 0; i < grad.g_inputs.size(); i++){
for (int i = 0; i < grad.g_inputs.size(); i++) {
if (terms_.count(grad.g_inputs[i]))
grad.g_inputs[i] = terms_[grad.g_inputs[i]];
}
......@@ -106,14 +106,14 @@ GraphDef GraphGradientMaker::Make() {
string split_name = *output + "_autosplit_" + str(grads_count[*output]++);
if (!is_skip) all_split_grads.insert(split_name);
// gather
if (grads_count[*output] == inputs_count[original_name]){
if (grads_count[*output] == inputs_count[original_name]) {
gather_op = new OperatorDef();
gather_op->set_name(GetOperatorName());
gather_op->set_type("GradientGather");
gather_op->add_output(*output);
if (g_op.has_device_option())
gather_op->mutable_device_option()->CopyFrom(g_op.device_option());
for (int j = 0; j < grads_count[*output]; j++){
for (int j = 0; j < grads_count[*output]; j++) {
string key = *output + "_autosplit_" + str(j);
if (all_split_grads.count(key)) gather_op->add_input(key);
}
......
Dragon/src/core/operator.cc
......@@ -66,7 +66,7 @@ DEFINE_REGISTRY(GradientRegistry, GradientMakerBase, const OperatorDef&, const v
DEFINE_REGISTRY(NoGradientRegistry, GradientMakerBase, const OperatorDef&, const vector<string>&);
#define INSTANTIATE_GET_SINGLE_ARGUMENT(T, fieldname) \
template <> T OperatorBase::GetSingleArg(const string& name, const T& default_value){ \
template <> T OperatorBase::GetSingleArg(const string& name, const T& default_value) { \
if(args_.count(name) == 0) { \
return default_value; \
} \
......@@ -82,7 +82,7 @@ INSTANTIATE_GET_SINGLE_ARGUMENT(int64_t, i64);
#define INSTANTIATE_GET_REPEATED_ARGUMENT(T, fieldname) \
template<> vector<T> OperatorBase::GetRepeatedArg<T>(const string& name){ \
template<> vector<T> OperatorBase::GetRepeatedArg<T>(const string& name) { \
if(args_.count(name) == 0) return vector<T>(); \
vector<T> values; \
for(const auto& v : args_[name]->fieldname()) values.push_back(v); \
......
Dragon/src/core/operator_schema.cc
......@@ -17,7 +17,7 @@ bool OpSchema::Verify(const OperatorDef& def) const {
}
for (int in = 0; in < def.input_size(); in++) {
if (def.input(in) == "ignore") continue;
for (int out = 0; out < def.output_size(); out++){
for (int out = 0; out < def.output_size(); out++) {
if (def.output(out) == "ignore") continue;
if (def.input(in) == def.output(out) && (!CheckInplace(in, out)))
LOG(FATAL) << "[" << def.name() << "] input("
......
Dragon/src/operators/activation/dropout_op.cc
......@@ -71,10 +71,9 @@ void DropoutGradientOp<Context>::RunOnDevice() {
template <class Context>
void DropoutGradientOp<Context>::ClearAfterRun() {
ws()->ReleaseBuffer(mask);
ws()->ReleaseBuffer(mask, true);
}
DEPLOY_CPU(DropoutGradient);
#ifdef WITH_CUDA
DEPLOY_CUDA(DropoutGradient);
......
Dragon/src/operators/arithmetic/dot_op.cc
......@@ -125,7 +125,7 @@ void DotGradientOp<Context>::GemvRunWithType() {
}
template <class Context>
void DotGradientOp<Context>::RunOnDevice(){
void DotGradientOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(0));
output(1)->ReshapeLike(input(1));
......
Dragon/src/operators/arithmetic/eltwise_op.cc
......@@ -65,7 +65,7 @@ void EltwiseGradientOp<Context>::SumRunWithType() {
auto* dYdata = input(-1).template data<T, Context>();
TIndex count = input(-1).count();
for (int i = 0; i < OutputSize(); i++){
for (int i = 0; i < OutputSize(); i++) {
if (output(i)->name() == "ignore") continue;
auto* dXdata = output(i)->template mutable_data<T, Context>();
if (coeffs[i] == float(1)) {
......
Dragon/src/operators/arithmetic/gram_matrix_op.cc
......@@ -8,7 +8,7 @@ template <class Context> template <typename T>
void GramMatrixOp<Context>::RunWithType() {
auto* Xdata = input(0).template data<T, Context>();
auto* Ydata = output(0)->template mutable_data<T, Context>();
for (int i = 0; i < outer_dim; i++){
for (int i = 0; i < outer_dim; i++) {
math::Gemm<T, Context>(CblasNoTrans, CblasTrans,
dim, dim, inner_dim, 1.0, Xdata, Xdata, 0.0, Ydata);
Xdata += x_offset;
......@@ -40,7 +40,7 @@ void GramMatrixGradientOp<Context>::RunWithType() {
auto* dYdata = input(-1).template data<T, Context>();
auto* Xdata = input(0).template data<T, Context>();
auto* dXdata = output(0)->template mutable_data<T, Context>();
for (int i = 0; i < outer_dim; i++){
for (int i = 0; i < outer_dim; i++) {
math::Gemm<T, Context>(CblasNoTrans, CblasNoTrans,
dim, inner_dim, dim, 2.0, dYdata, Xdata, 0.0, dXdata);
dYdata += y_offset;
......@@ -49,7 +49,7 @@ void GramMatrixGradientOp<Context>::RunWithType() {
}
template <class Context>
void GramMatrixGradientOp<Context>::RunOnDevice(){
void GramMatrixGradientOp<Context>::RunOnDevice() {
outer_dim = input(0).count(0, axis);
dim = input(0).dim(axis);
inner_dim = input(0).count(axis + 1);
......
Dragon/src/operators/arithmetic/matmul_op.cc
......@@ -81,7 +81,7 @@ void MatmulGradientOp<Context>::RunWithType() {
}
template <class Context>
void MatmulGradientOp<Context>::RunOnDevice(){
void MatmulGradientOp<Context>::RunOnDevice() {
CHECK(input(0).ndim() == input(1).ndim())
<< "both matrices must have the same number of dimensions.";
CHECK_GE(input(0).ndim(), 2)
......
Dragon/src/operators/arithmetic/pow_op.cc
......@@ -9,7 +9,7 @@ void PowOp<Context>::RunWithType() {
TIndex count = input(0).count();
auto* Ydata = output(0)->template mutable_data<T, Context>();
if (power_scale == float(0)){
if (power_scale == float(0)) {
float value = (power == float(0)) ? float(1) : pow(shift, power);
math::Set<T, Context>(count, dragon_cast<T, float>(value), Ydata);
return;
......
Dragon/src/operators/arithmetic/square_op.cc
......@@ -12,7 +12,7 @@ void SquareOp<Context>::RunWithType() {
}
template <class Context>
void SquareOp<Context>::RunOnDevice(){
void SquareOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(0));
if (input(0).template IsType<float>()) RunWithType<float>();
......
Dragon/src/operators/arithmetic/sub_op.cc
......@@ -44,7 +44,7 @@ void SubOp<Context>::BroadcastRunWithType(int type) {
}
template <class Context>
void SubOp<Context>::RunOnDevice(){
void SubOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(0));
if (input(0).dims() == input(1).dims()) {
......
Dragon/src/operators/common/concat_op.cc
......@@ -7,7 +7,7 @@ namespace dragon {
template <class Context> template <typename T>
void ConcatOp<Context>::RunWithType() {
auto* Ydata = output(0)->template mutable_data<T, Context>();
for (int i = 0; i < nin; i++){
for (int i = 0; i < nin; i++) {
auto* Xdata = input(i).template data<T, Context>();
TIndex count = input(i).count();
x_concat_dim = input(i).dim(axis);
......@@ -25,12 +25,12 @@ void ConcatOp<Context>::RunWithType() {
}
template <class Context>
void ConcatOp<Context>::RunOnDevice(){
void ConcatOp<Context>::RunOnDevice() {
concat_dims = input(0).dims();
for (int i = 1; i < nin; i++) {
CHECK_EQ(concat_dims.size(), input(i).ndim())
<< "\nall inputs must have the same ndim.";
for (int j = 0; j < concat_dims.size(); j++){
for (int j = 0; j < concat_dims.size(); j++) {
if (j == axis) continue;
CHECK_EQ(concat_dims[j], input(i).dim(j))
<< "\nall inputs must have the same dims"
......@@ -82,7 +82,7 @@ void ConcatGradientOp<Context>::RunWithType() {
}
template <class Context>
void ConcatGradientOp<Context>::RunOnDevice(){
void ConcatGradientOp<Context>::RunOnDevice() {
if (input(-1).name() == "ignore") return;
concat_dims = input(-1).dims();
y_concat_dim = concat_dims[axis];
......
Dragon/src/operators/common/flatten_op.cc
......@@ -27,7 +27,7 @@ OPERATOR_SCHEMA(Flatten).NumInputs(1).NumOutputs(1);
template <class Context>
void FlattenGradientOp<Context>::RunOnDevice(){
void FlattenGradientOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(0));
output(0)->Share(input(1));
}
......
Dragon/src/operators/common/reduce_op.cc
......@@ -78,7 +78,7 @@ void ReduceGradientOp<Context>::SumRunWithType() {
template <class Context> template <typename T>
void ReduceGradientOp<Context>::MeanRunWithType() {
auto* dXdata = output(0)->template mutable_data<T, Context>();
if (axis == -1){
if (axis == -1) {
auto* dYdata = input(-1).template data<T, CPUContext>();
math::Set<T, Context>(output(0)->count(), dYdata[0] / input(0).count(), dXdata);
} else {
......
Dragon/src/operators/common/scan_op.cc
......@@ -150,13 +150,11 @@ void ScanOp<Context>::UnrollTemplate() {
template <class Context>
void ScanOp<Context>::RunOnDevice() {
// unroll
UnrollTemplate();
if (!graphs.count(nsteps))
if (!graphs.count(nsteps)) {
graphs[nsteps].reset(new Graph(new_def, ws()));
}
cur_graph = graphs[nsteps].get();
// forward
cur_graph->Run("", "");
}
......@@ -201,13 +199,11 @@ void ScanGradientOp<Context>::MakeGradientOps() {
template <class Context>
void ScanGradientOp<Context>::RunOnDevice() {
// make graph
MakeGradientOps();
if (!graphs.count(nsteps))
if (!graphs.count(nsteps)) {
graphs[nsteps].reset(new Graph(new_def, ws()));
}
cur_graph = graphs[nsteps].get();
// backward
cur_graph->Run("Gradient", "");
}
......
Dragon/src/operators/common/slice_op.cc
......@@ -7,7 +7,7 @@ namespace dragon {
template <class Context> template <typename T>
void SliceOp<Context>::RunWithType() {
auto* Xdata = input(0).template data<T, Context>();
for (int i = 0; i < nout; i++){
for (int i = 0; i < nout; i++) {
auto* Ydata = output(i)->template mutable_data<T, Context>();
TIndex count = output(i)->count();
kernel::Slice<T, Context>(count, outer_dim, inner_dim,
......@@ -46,7 +46,7 @@ OPERATOR_SCHEMA(Slice).NumInputs(1).NumOutputs(1, INT_MAX);
template <class Context> template <typename T>
void SliceGradientOp<Context>::RunWithType() {
auto* dXdata = output(0)->template mutable_data<T, Context>();
for (int i = 0; i < nout; i++){
for (int i = 0; i < nout; i++) {
if (input(i + 1).name() == "ignore") continue;
auto* dYdata = input(i + 1).template data<T, Context>();
TIndex count = input(i + 1).count();
......
Dragon/src/operators/common/utils_op.cc
......@@ -64,7 +64,7 @@ void AccuracyOp<Context>::RunWithType() {
} // end outer_num
output(0)->template mutable_data<T, CPUContext>()[0] = acc / count;
if (OutputSize() > 1){
if (OutputSize() > 1) {
auto* acc_per_class = output(1)->template mutable_data<T, CPUContext>();
for (int i = 0; i < classes; i++)
acc_per_class[i] = num_per_class[i] == 0 ? 0 : acc_per_class[i] / acc_per_class[i];
......
Dragon/src/operators/loss/l1_loss_op.cc
......@@ -28,7 +28,7 @@ void L1LossOp<Context>::RunWithType() {
}
template <class Context>
void L1LossOp<Context>::RunOnDevice(){
void L1LossOp<Context>::RunOnDevice() {
CHECK_EQ(input(0).count(), input(1).count());
output(0)->Reshape(vector<TIndex>(1, 1));
diff = ws()->CreateTensor("_t_" + anchor() + "_l1_loss_diff");
......@@ -53,7 +53,7 @@ void L1LossGradientOp<Context>::RunWithType() {
else if (normalization == "FULL") normalizer = input(0).count();
else if (normalization == "NONE") normalizer = 1;
alpha = alpha / normalizer;
for (int i = 0; i < 2; i++){
for (int i = 0; i < 2; i++) {
if (output(i)->name() == "ignore") continue;
output(i)->ReshapeLike(input(i));
auto* dXdata = output(i)->template mutable_data<T, Context>();
......
Dragon/src/operators/loss/l2_loss_op.cc
......@@ -11,7 +11,7 @@ void L2LossOp<Context>::RunWithType() {
auto* diff_data = diff->template mutable_data<T, Context>();
auto* Ydata = output(0)->template mutable_data<T, CPUContext>();
math::Sub<T, Context>(diff->count(), X0data, X1data, diff_data);
if (InputSize() > 2){
if (InputSize() > 2) {
CHECK_EQ(input(0).count(), input(2).count());
auto* Wdata = input(2).template data<T, Context>();
math::Mul<T, Context>(diff->count(), Wdata, diff_data, diff_data);
......
Dragon/src/operators/loss/smooth_l1_loss_op.cc
......@@ -14,12 +14,12 @@ void SmoothL1LossOp<Context>::RunWithType() {
auto* Ydata = output(0)->template mutable_data<T, CPUContext>();
math::Sub<T, Context>(diff->count(), X0data, X1data, diff_data);
if (InputSize() > 2){
if (InputSize() > 2) {
auto* inside_w_data = input(2).template data<T, Context>();
math::Mul<T, Context>(diff->count(), inside_w_data, diff_data, diff_data);
}
kernel::SmoothL1<T, Context>(diff->count(), sigma2, diff_data, error_data);
if (InputSize() > 3){
if (InputSize() > 3) {
auto* outside_w_data = input(3).template data<T, Context>();
math::Mul<T, Context>(diff->count(), outside_w_data, error_data, error_data);
}
......
Dragon/src/operators/mpi/mpi_broadcast_op.cc
......@@ -26,14 +26,14 @@ void MPIBroadcastOp<Context>::RunWithType() {
}
template <class Context>
void MPIBroadcastOp<Context>::RunOnDevice(){
void MPIBroadcastOp<Context>::RunOnDevice() {
CHECK(this->comm != MPI_COMM_NULL)
<< "\nMPIBroadcastOp, name: " << name()
<< ", does not belong to any group, can't run.";
size_t ndim[1];
TIndex* dims = nullptr;
if (this->comm_rank == this->comm_root){
if (this->comm_rank == this->comm_root) {
ndim[0] = input(0).ndim();
dims = new TIndex[ndim[0]];
for (int i = 0; i < input(0).ndim(); i++)
......@@ -90,7 +90,7 @@ void MPIBroadcastGradientOp<Context>::RunWithType() {
}
template <class Context>
void MPIBroadcastGradientOp<Context>::RunOnDevice(){
void MPIBroadcastGradientOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(-1));
if (input(0).template IsType<float>()) RunWithType<float>();
......
Dragon/src/operators/mpi/mpi_gather_op.cc
......@@ -51,7 +51,7 @@ template <class Context> template <typename T>
void MPIGatherGradientOp<Context>::RunWithType() {
if (this->comm_rank == this->comm_root) {
output(0)->Share(input(this->comm_rank + 1));
for (int i = 0; i < this->comm_size; i++){
for (int i = 0; i < this->comm_size; i++) {
if (i == this->comm_root) continue;
#ifdef WITH_CUDA_AWARE
auto* dYdata = input(this->comm_rank + 1).template data<T, Context>();
......
Dragon/src/operators/norm/instance_norm_op.cc
......@@ -60,7 +60,7 @@ void InstanceNormOp<Context>::RunWithType() {
}
template <class Context>
void InstanceNormOp<Context>::RunOnDevice(){
void InstanceNormOp<Context>::RunOnDevice() {
num = input(0).dim(0); channels = input(0).dim(1);
spatial_dim = input(0).count(2); nbychans = num * channels;
vector<TIndex> dims({ num, channels });
......
Dragon/src/operators/norm/l2_norm_op.cc
......@@ -109,7 +109,7 @@ void L2NormGradientOp<Context>::RunWithType() {
auto* Bdata = buffer->template mutable_data<T, Context>();
auto* BInnerdata = buffer_inner->template mutable_data<T, Context>();
for (int n = 0; n < outer_dim; n++){
for (int n = 0; n < outer_dim; n++) {
if (across_inner) {
Ndata = norm->template data<T, CPUContext>();
T sum_of_x_mul_dy = math::Dot<T, Context>(buffer->count(), Xdata, dYdata);
......
Dragon/src/operators/update/async_update_op.cc
......@@ -21,7 +21,7 @@ void AsyncUpdateOp<Context>::UpdateTimestamp(int tag) {
}
template <class Context>
int AsyncUpdateOp<Context>::GetDelay(int tag){
int AsyncUpdateOp<Context>::GetDelay(int tag) {
Tensor* t = ws()->GetTensor("_t_" + this->domain + "async_timestamp");
int* global_timestamp = t->template mutable_data<int, CPUContext>();
return global_timestamp[tag] - local_timestamp[tag] + 1;
......@@ -71,7 +71,7 @@ AsyncUpdateOp<Context>::AsyncUpdateOp(const OperatorDef& op_def, Workspace* ws)
template <class Context> template <typename T>
void AsyncUpdateOp<Context>::RootRunWithType() {
for (int i = 0; i < InputSize(); i++){
for (int i = 0; i < InputSize(); i++) {
auto* dXdata = input(i).template mutable_data<T, Context>();
auto* Xdata = output(i)->template mutable_data<T, Context>();
......@@ -88,7 +88,7 @@ void AsyncUpdateOp<Context>::RootRunWithType() {
}
template <class Context>
void AsyncUpdateOp<Context>::RunOnDevice(){
void AsyncUpdateOp<Context>::RunOnDevice() {
if (this->comm_rank != this->comm_root) return;
if (input(0).template IsType<float>()) {
......
Dragon/src/operators/update/update_op_base.cc
......@@ -98,7 +98,7 @@ void UpdateOpBase<Context>::ReduceRunWithType() {
}
// ave-normalize
if (comm_size > 1){
if (comm_size > 1) {
#ifdef WITH_CUDA_AWARE
math::Scal<T, Context>(count, T(1.0 / comm_size), dXdata);
#else
......@@ -112,7 +112,7 @@ template <class Context> template <typename T>
void UpdateOpBase<Context>::PreprocessRunWithType() {
// scale
scale_factor = param("scale_gradient");
if (scale_factor != 1){
if (scale_factor != 1) {
auto* dXdata = input(0).template mutable_data<T, Context>();
math::Scal<T, Context>(input(0).count(), scale_factor, dXdata);
}
......@@ -128,11 +128,11 @@ void UpdateOpBase<Context>::PreprocessRunWithType() {
}
}
// decay
l2_decay = param("l2_decay");
if (l2_decay > 0){
l2_decay = param("l2_decay") * decay_mult;
if (l2_decay > 0) {
auto* dXdata = input(0).template mutable_data<T, Context>();
auto* Xdata = output(0)->template data<T, Context>();
math::Axpy<T, Context>(input(0).count(), l2_decay * decay_mult, Xdata, dXdata);
math::Axpy<T, Context>(input(0).count(), l2_decay, Xdata, dXdata);
}
}
......
Dragon/src/operators/utils/compare_op.cc
......@@ -12,7 +12,7 @@ void CompareOp<Context>::EqualRunWithType() {
}
template <class Context>
void CompareOp<Context>::RunOnDevice(){
void CompareOp<Context>::RunOnDevice() {
CHECK_EQ(input(0).count(), input(1).count())
<< "both conditioned tensor must have same elements.";
output(0)->ReshapeLike(input(0));
......
Dragon/src/operators/utils/gradient_op.cc
......@@ -6,7 +6,7 @@ namespace dragon {
template <class Context> template <typename T>
void GradientGenerateOp<Context>::RunWithType() {
for (int i = 0; i < OutputSize(); i++){
for (int i = 0; i < OutputSize(); i++) {
if (output(i)->name() == "ignore") continue;
output(i)->ReshapeLike(input(i));
auto* dXdata = output(0)->template mutable_data<T, Context>();
......
Dragon/src/operators/vision/conv_op.cc
......@@ -7,7 +7,7 @@ namespace dragon {
template <class Context>
void ConvOp<Context>::ComputeOutputShape() {
this->output_shape.clear();
for (int i = 0; i < this->num_spatial_axes; i++){
for (int i = 0; i < this->num_spatial_axes; i++) {
const int input_dim = this->bottom_shape[this->channel_axis + i + 1];
const int dilated_kernel = this->dilation[i] * (this->kernel_size[i] - 1) + 1;
const int output_dim = (input_dim + 2 * this->pad[i] - dilated_kernel) / this->stride[i] + 1;
......@@ -43,7 +43,7 @@ void ConvOp<Context>::RunWithType() {
}
template <class Context>
void ConvOp<Context>::RunOnDevice(){
void ConvOp<Context>::RunOnDevice() {
Reshape();
if (input(0).template IsType<float>()) RunWithType<float>();
......@@ -71,7 +71,7 @@ void ConvGradientOp<Context>::RunWithType() {
Db(dYdata + n * this->y_offset, dBdata);
}
for (int n = 0; n < input(2).dim(0); n++){
for (int n = 0; n < input(2).dim(0); n++) {
if (output(1)->name() != "ignore") {
auto* Xdata = input(0).template data<T, Context>();
auto* dWdata = output(1)->template mutable_data<T, Context>();
......
Dragon/src/operators/vision/conv_op_base.cc
......@@ -161,7 +161,7 @@ void ConvOpBase<Context>::GradientReshape() {
// compute input shape
input_shape.clear();
for (int i = 0; i < num_spatial_axes; i++){
for (int i = 0; i < num_spatial_axes; i++) {
if (ReverseDimensions()) {
input_shape.push_back(input(2).dim(channel_axis + i + 1));
} else {
......@@ -181,7 +181,7 @@ void ConvOpBase<Context>::GradientReshape() {
// compute col buffer shape
col_buffer_shape.clear();
col_buffer_shape.push_back(kernel_dim * group);
for (int i = 0; i < num_spatial_axes; i++){
for (int i = 0; i < num_spatial_axes; i++) {
if (ReverseDimensions()) {
col_buffer_shape.push_back(bottom_shape[channel_axis + i + 1]);
} else {
......
Dragon/src/operators/vision/cudnn_conv_op.cc
......@@ -228,13 +228,13 @@ void CuDNNConvGradientOp<Context>::RunWithType() {
const T* dYdata = input(2).template data<T, Context>();
for (int g = 0; g < this->group; g++) {
if (output(2)->name() != "ignore"){
if (output(2)->name() != "ignore") {
T* dBdata = output(2)->template mutable_data<T, Context>();
CUDNN_CHECK(cudnnConvolutionBackwardBias(handle[g],
CUDNNType<T>::one, input_desc, dYdata + this->y_offset * g,
CUDNNType<T>::one, bias_desc, dBdata + bias_offset * g));
}
if (output(1)->name() != "ignore"){
if (output(1)->name() != "ignore") {
auto* Xdata = input(0).template data<T, Context>();
auto* dWdata = output(1)->template mutable_data<T, Context>();
auto* workspace = buffer2->mutable_data<char, Context>();
......@@ -246,7 +246,7 @@ void CuDNNConvGradientOp<Context>::RunWithType() {
workspace + g * workspace_bwd_filter_size, workspace_bwd_filter_size,
CUDNNType<T>::one, filter_desc, dWdata + this->weight_offset * g));
}
if (output(0)->name() != "ignore"){
if (output(0)->name() != "ignore") {
auto* Wdata = input(1).template data<T, Context>();
auto* dXdata = output(0)->template mutable_data<T, Context>();
auto* workspace = buffer1->mutable_data<char, Context>();
......
Dragon/src/operators/vision/cudnn_lrn_op.cc
......@@ -54,7 +54,7 @@ void CuDNNLRNGradientOp<Context>::RunWithType() {
}
template <class Context>
void CuDNNLRNGradientOp<Context>::RunOnDevice(){
void CuDNNLRNGradientOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(0));
if (this->mode == ACROSS_CHANNELS) {
......
Dragon/src/operators/vision/deconv_op.cc
......@@ -43,7 +43,7 @@ void DeConvOp<Context>::RunWithType() {
}
template <class Context>
void DeConvOp<Context>::RunOnDevice(){
void DeConvOp<Context>::RunOnDevice() {
Reshape();
if (input(0).template IsType<float>()) RunWithType<float>();
......
Dragon/src/operators/vision/lrn_op.cc
......@@ -98,7 +98,7 @@ void LRNOp<Context>::ProdRunWithType() {
}
template <class Context>
void LRNOp<Context>::RunOnDevice(){
void LRNOp<Context>::RunOnDevice() {
if (mode == ACROSS_CHANNELS) {
if (input(0).template IsType<float>()) {
AcrossRunWithType<float>();
......@@ -223,7 +223,7 @@ void LRNGradientOp<Context>::SplitRunWithType() {
}
template <class Context>
void LRNGradientOp<Context>::RunOnDevice(){
void LRNGradientOp<Context>::RunOnDevice() {
if (mode == ACROSS_CHANNELS) {
if (input(0).template IsType<float>()) {
AcrossRunWithType<float>();
......
Dragon/src/operators/vision/nn_resize_op.cc
......@@ -16,7 +16,7 @@ void NNResizeOp<Context>::RunWithType() {
template <class Context>
void NNResizeOp<Context>::RunOnDevice() {
dims = input(0).dims();
if (dsize.size() == 0){
if (dsize.size() == 0) {
CHECK(fy != -1.0 && fx != -1.0);
dims[2] = int(dims[2] * fy);
dims[3] = int(dims[3] * fx);
......@@ -47,7 +47,7 @@ void NNResizeGradientOp<Context>::RunWithType() {
}
template <class Context>
void NNResizeGradientOp<Context>::RunOnDevice(){
void NNResizeGradientOp<Context>::RunOnDevice() {
output(0)->ReshapeLike(input(0));
if (input(0).template IsType<float>()) return RunWithType<float>();
......
Dragon/src/utils/cudnn_device.cc
......@@ -37,7 +37,7 @@ void cudnnSetTensorDesc(cudnnTensorDescriptor_t* desc, const vector<TIndex>& dim
int* dimA = new int[ndim];
int* strideA = new int[ndim];
TIndex stride = 1;
for (int i = ndim - 1; i >= 0; i--){
for (int i = ndim - 1; i >= 0; i--) {
strideA[i] = stride;
dimA[i] = dims[i];
stride *= dimA[i];
......@@ -55,7 +55,7 @@ void cudnnSetTensorDesc(cudnnTensorDescriptor_t* desc,
int ndim = (int)dims.size();
int* dimA = new int[ndim];
int* strideA = new int[ndim];
for (int i = ndim - 1; i >= 0; i--){
for (int i = ndim - 1; i >= 0; i--) {
strideA[i] = strides[i];
dimA[i] = dims[i];
}
......@@ -66,10 +66,10 @@ void cudnnSetTensorDesc(cudnnTensorDescriptor_t* desc,
template <typename T>
void cudnnSetTensorDesc(cudnnTensorDescriptor_t* desc, Tensor* tensor) {
// CUDNN only support ndim from 3 to 8
// cuDNN requires ndim from 3 to 8
// we fake a reshaped dims to pass check
vector<TIndex> fake_dims(tensor->dims());
if (fake_dims.size() < 3 || fake_dims.size() > 8){
if (fake_dims.size() < 3 || fake_dims.size() > 8) {
fake_dims.assign({ 1, 1 });
fake_dims.push_back(tensor->count());
}
......
Dragon/src/utils/math_functions.cc
......@@ -142,7 +142,7 @@ template <> void Add<float, CPUContext>(const int n,
template <> void Sub<float, CPUContext>(const int n,
const float* a,
const float* b,
float* y){
float* y) {
#ifdef WITH_SSE
sse::Sub<float>(n, a, b, y);
#else // naive implement
......
Dragon/src/utils/op_kernel.cc
......@@ -98,7 +98,7 @@ template<> void Softmax<float, CPUContext>(const int count,
const int dim = count / outer_dim;
for (int i = 0; i < outer_dim; ++i) {
context->Copy<float, CPUContext, CPUContext>(inner_dim, scale, x + i*dim);
for (int j = 0; j < classes; ++j){
for (int j = 0; j < classes; ++j) {
for (int k = 0; k < inner_dim; k++)
scale[k] = std::max(scale[k], x[i * dim + j * inner_dim + k]);
}
......@@ -1362,7 +1362,7 @@ template<> void ROIPooling<float, CPUContext>(const float spatial_scale,
Ydata[pool_idx] = Idata[idx];
Mdata[pool_idx] = idx;
}
} //end w
} // end w
} // end h
} // end pw
} // end ph
......@@ -1373,7 +1373,7 @@ template<> void ROIPooling<float, CPUContext>(const float spatial_scale,
} // end c
// offset roi region
Rdata += roi->offset(1);
} //end n
} // end n
}
template<> void ROIPoolingGrad<float, CPUContext>(const float spatial_scale,
......
Dragon/src/utils/op_kernel.cu
......@@ -130,7 +130,7 @@ __global__ void _ReluGrad(const int count,
const T* y,
const float slope,
T* dx) {
CUDA_KERNEL_LOOP(i, count){
CUDA_KERNEL_LOOP(i, count) {
dx[i] = dy[i] * ((y[i] > 0) + slope * (y[i] <= 0));
}
}
......@@ -912,7 +912,7 @@ __global__ void _Sum(const int count,
template<> void Sum<float, CUDAContext>(
const int count, const int axis_dim,
const int inner_dim, const float* x, float* y){
const int inner_dim, const float* x, float* y) {
_Sum<float> << <GET_BLOCKS(count), CUDA_NUM_THREADS >> >(count,
axis_dim,
inner_dim,
......@@ -954,7 +954,7 @@ template<> void SumGrad<float, CUDAContext>(const int count,
template <typename T>
__global__ void _Slice(const int count, const int outer_dim, const int inner_dim,
const int x_slice_dim, const int y_slice_dim, const int slice_offset, const T* x, T* y){
const int x_slice_dim, const int y_slice_dim, const int slice_offset, const T* x, T* y) {
CUDA_KERNEL_LOOP(idx, count) {
const int tmp = y_slice_dim * inner_dim;
const int outer_idx = idx / tmp;
......
Dragon/src/utils/sse_alternative.cc
......@@ -9,20 +9,20 @@ namespace dragon {
namespace sse {
template<> void Set(const int n, const float alpha, float* x){
template<> void Set(const int n, const float alpha, float* x) {
__m128 scalar = SSE_FP32_SCALAR(alpha);
SSE_LOOP1(i, n) SSE_FP32_STORE(x + i, scalar);
SSE_LOOP2(i, n) x[i] = alpha;
}
template<> void Set(const int n, const int alpha, int* x){
template<> void Set(const int n, const int alpha, int* x) {
__m128i scalar = SSE_INT32_SCALAR(alpha);
__m128i* x1 = reinterpret_cast<__m128i*>(x);
SSE_LOOP1(i, n) SSE_INT128_STORE(x1++, scalar);
SSE_LOOP2(i, n) x[i] = alpha;
}
template<> void Add(const int n, const float* a, const float* b, float* y){
template<> void Add(const int n, const float* a, const float* b, float* y) {
__m128 x1, y1, z1;
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(a + i);
......@@ -33,7 +33,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] = a[i] + b[i];
}
template<> void Sub(const int n, const float* a, const float* b, float* y){
template<> void Sub(const int n, const float* a, const float* b, float* y) {
__m128 x1, y1, z1;
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(a + i);
......@@ -44,7 +44,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] = a[i] - b[i];
}
template<> void Mul(const int n, const float* a, const float* b, float* y){
template<> void Mul(const int n, const float* a, const float* b, float* y) {
__m128 x1, y1, z1;
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(a + i);
......@@ -55,7 +55,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] = a[i] * b[i];
}
template<> void Div(const int n, const float* a, const float* b, float* y){
template<> void Div(const int n, const float* a, const float* b, float* y) {
__m128 x1, y1, z1;
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(a + i);
......@@ -66,7 +66,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] = a[i] / b[i];
}
template<> void Scal(const int n, const float alpha, float* y){
template<> void Scal(const int n, const float alpha, float* y) {
__m128 y1, scalar = SSE_FP32_SCALAR(alpha);
SSE_LOOP1(i, n) {
y1 = SSE_FP32_LOAD(y + i);
......@@ -76,7 +76,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] *= alpha;
}
template<> void Scale(const int n, const float alpha, const float* x, float* y){
template<> void Scale(const int n, const float alpha, const float* x, float* y) {
__m128 x1, scalar = SSE_FP32_SCALAR(alpha);
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(x + i);
......@@ -87,7 +87,7 @@ namespace sse {
}
template<> void Axpy(const int n, float alpha, const float* x, float *y){
template<> void Axpy(const int n, float alpha, const float* x, float *y) {
__m128 x1, y1, scalar = SSE_FP32_SCALAR(alpha);
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(x + i);
......@@ -100,7 +100,7 @@ namespace sse {
}
template<> void Axpby(const int n, float alpha, const float* x,
const float beta, float *y){
const float beta, float *y) {
__m128 x1, y1, z1;
__m128 scalar1 = SSE_FP32_SCALAR(alpha);
__m128 scalar2 = SSE_FP32_SCALAR(beta);
......@@ -115,7 +115,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] = alpha * x[i] + beta* y[i];
}
template<> float ASum(const int n, const float *x){
template<> float ASum(const int n, const float *x) {
__m128 x1, sum = SSE_FP32_ZERO;
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(x + i);
......@@ -128,7 +128,7 @@ namespace sse {
return ret;
}
template<> void AddScalar(const int n, const float alpha, float* y){
template<> void AddScalar(const int n, const float alpha, float* y) {
__m128 y1, scalar = SSE_FP32_SCALAR(alpha);
SSE_LOOP1(i, n) {
y1 = SSE_FP32_LOAD(y + i);
......@@ -138,7 +138,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] += alpha;
}
template<> void MulScalar(const int n, const float alpha, float* y){
template<> void MulScalar(const int n, const float alpha, float* y) {
__m128 y1, scalar = SSE_FP32_SCALAR(alpha);
SSE_LOOP1(i, n) {
y1 = SSE_FP32_LOAD(y + i);
......@@ -148,7 +148,7 @@ namespace sse {
SSE_LOOP2(i, n) y[i] *= alpha;
}
template <> float Dot(const int n, const float* a, const float* b){
template <> float Dot(const int n, const float* a, const float* b) {
__m128 x1, y1, sum = SSE_FP32_ZERO;
SSE_LOOP1(i, n) {
x1 = SSE_FP32_LOAD(a + i);
......
examples/cifar10/gen_lmdb.py
......@@ -11,6 +11,7 @@ import sys
import time
import shutil
import tarfile
import numpy as np
from six.moves import range as xrange
import cv2
......@@ -78,7 +79,7 @@ def extract_images():
f.write(item)
def make_db(image_path, label_path, database_path):
def make_db(image_path, label_path, database_path, pad=0):
if os.path.isfile(label_path) is False:
raise ValueError('input path is empty or wrong.')
if os.path.isdir(database_path) is True:
......@@ -111,6 +112,12 @@ def make_db(image_path, label_path, database_path):
label = record[1]
img = cv2.imread(os.path.join(image_path ,path))
if pad > 0:
pad_img = np.zeros((img.shape[0] + 2 * pad,
img.shape[1] + 2 * pad, 3), dtype=np.uint8)
pad_img[pad : pad + img.shape[0],
pad : pad + img.shape[1], :] = img
img = pad_img
result, imgencode = cv2.imencode('.jpg', img, encode_param)
datum = caffe_pb2.Datum()
......
Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!