From 032d2626cbab151e85ccf668825aedc7410ef9b0 Mon Sep 17 00:00:00 2001 From: huangbo77 Date: Wed, 25 Aug 2021 11:50:05 +0800 Subject: [PATCH] adding cpu mode for crnn --- model_zoo/official/cv/crnn/README.md | 37 +++++++++--- model_zoo/official/cv/crnn/convert_iiit5k.py | 4 +- model_zoo/official/cv/crnn/eval.py | 5 +- model_zoo/official/cv/crnn/export.py | 1 + .../official/cv/crnn/scripts/run_eval_cpu.sh | 58 +++++++++++++++++++ .../crnn/scripts/run_standalone_train_cpu.sh | 54 +++++++++++++++++ model_zoo/official/cv/crnn/src/crnn.py | 51 +++++++++++++++- model_zoo/official/cv/crnn/train.py | 4 +- 8 files changed, 197 insertions(+), 17 deletions(-) create mode 100644 model_zoo/official/cv/crnn/scripts/run_eval_cpu.sh create mode 100644 model_zoo/official/cv/crnn/scripts/run_standalone_train_cpu.sh diff --git a/model_zoo/official/cv/crnn/README.md b/model_zoo/official/cv/crnn/README.md index 048ac41595b..922bd409ac5 100644 --- a/model_zoo/official/cv/crnn/README.md +++ b/model_zoo/official/cv/crnn/README.md @@ -64,7 +64,7 @@ We provide `convert_ic03.py`, `convert_iiit5k.py`, `convert_svt.py` as exmples f ## [Environment Requirements](#contents) - Hardware - - Prepare hardware environment with Ascend processor or GPU. + - Prepare hardware environment with Ascend, GPU or CPU processor. - Framework - [MindSpore](https://gitee.com/mindspore/mindspore) - For more information, please check the resources below： @@ -105,6 +105,16 @@ We provide `convert_ic03.py`, `convert_iiit5k.py`, `convert_svt.py` as exmples f $ bash scripts/run_standalone_train.sh [DATASET_NAME] [DATASET_PATH] GPU ``` + - Running on CPU + + ```shell + # standalone training example in CPU + $ bash scripts/run_standalone_train_cpu.sh [DATASET_NAME] [DATASET_PATH] + + # evaluation example in CPU + $ bash scripts/run_eval_cpu.sh [DATASET_NAME] [DATASET_PATH] [CHECKPOINT_PATH] + ``` + DATASET_NAME is one of `ic03`, `ic13`, `svt`, `iiit5k`, `synth`. For distributed training, a hccl configuration file with JSON format needs to be created in advance. @@ -142,9 +152,11 @@ crnn ├── convert_svt.py # Convert the original SVT dataset ├── requirements.txt # Requirements for this dataset ├── scripts -│ ├── run_distribute_train.sh # Launch distributed training in Ascend(8 pcs) -│ ├── run_eval.sh # Launch evaluation -│ └── run_standalone_train.sh # Launch standalone training(1 pcs) +│ ├── run_standalone_train_cpu.sh # Launch standalone training in CPU +│ ├── run_eval_cpu.sh # Launch evaluation in CPU +│ ├── run_distribute_train.sh # Launch distributed training in Ascend or GPU(8 pcs) +│ ├── run_eval.sh # Launch evaluation in Ascend or GPU +│ └── run_standalone_train.sh # Launch standalone training in Ascend or GPU(1 pcs) ├── src │ ├── model_utils │ ├── config.py # Parameter config @@ -172,11 +184,14 @@ crnn #### Training Script Parameters ```shell -# distributed training +# distributed training in Ascend or GPU Usage: bash scripts/run_distribute_train.sh [DATASET_NAME] [DATASET_PATH] [PLATFORM] [RANK_TABLE_FILE](if Ascend) -# standalone training +# standalone training in Ascend or GPU Usage: bash scripts/run_standalone_train.sh [DATASET_NAME] [DATASET_PATH] [PLATFORM] + +# standalone training in CPU +Usage: bash scripts/run_standalone_train_cpu.sh [DATASET_NAME] [DATASET_PATH] ``` #### Parameters Configuration @@ -220,6 +235,12 @@ max_text_length": 23, # max number of digits in each bash scripts/run_standalone_train.sh [DATASET_NAME] [DATASET_PATH] [PLATFORM](optional) ``` +- Or run `run_standalone_train_cpu.sh` for non-distributed training of CRNN model in CPU. + +``` bash +bash scripts/run_standalone_train_cpu.sh [DATASET_NAME] [DATASET_PATH] +``` + #### [Distributed Training](#contents) - Run `run_distribute_train.sh` for distributed training of CRNN model on Ascend or GPU @@ -296,6 +317,8 @@ Epoch time: 2743.688s, per step time: 0.097s ``` bash bash scripts/run_eval.sh [DATASET_NAME] [DATASET_PATH] [CHECKPOINT_PATH] [PLATFORM](optional) + +bash scripts/run_eval_cpu.sh [DATASET_NAME] [DATASET_PATH] [CHECKPOINT_PATH] ``` Check the `eval/log.txt` and you will get outputs as following: @@ -313,7 +336,7 @@ You can add `run_eval` to start shell and set it True.You need also add `eval_da ### [Export MindIR](#contents) ```shell -python export.py --ckpt_file [CKPT_PATH] --file_name [FILE_NAME] --file_format [FILE_FORMAT] +python export.py --ckpt_file [CKPT_PATH] --file_name [FILE_NAME] --file_format [FILE_FORMAT] --device_target [DEVICE_TARGET] --model_version [MODEL_VERSION](required for cpu) ``` The ckpt_file parameter is required, diff --git a/model_zoo/official/cv/crnn/convert_iiit5k.py b/model_zoo/official/cv/crnn/convert_iiit5k.py index aed709d8236..8234ae6a87b 100644 --- a/model_zoo/official/cv/crnn/convert_iiit5k.py +++ b/model_zoo/official/cv/crnn/convert_iiit5k.py @@ -1,4 +1,4 @@ -# Copyright 2020 Huawei Technologies Co., Ltd +# Copyright 2020-21 Huawei Technologies Co., Ltd # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. @@ -46,7 +46,7 @@ def mat_to_list(mat_file): for elem in testdata: img_name = elem[0] label = elem[1] - ann = img_name+',' +label + ann = img_name[0] + ',' + label[0] ann_output.append(ann) return ann_output diff --git a/model_zoo/official/cv/crnn/eval.py b/model_zoo/official/cv/crnn/eval.py index f06c5551a5c..635ae6124e9 100644 --- a/model_zoo/official/cv/crnn/eval.py +++ b/model_zoo/official/cv/crnn/eval.py @@ -1,4 +1,4 @@ -# Copyright 2020 Huawei Technologies Co., Ltd +# Copyright 2020-21 Huawei Technologies Co., Ltd # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. @@ -51,7 +51,7 @@ def crnn_eval(): loss = CTCLoss(max_sequence_length=config.num_step, max_label_length=max_text_length, batch_size=config.batch_size) - net = crnn(config, full_precision=config.device_target == 'GPU') + net = crnn(config, full_precision=config.device_target != 'Ascend') # load checkpoint param_dict = load_checkpoint(config.checkpoint_path) load_param_into_net(net, param_dict) @@ -62,6 +62,5 @@ def crnn_eval(): res = model.eval(dataset, dataset_sink_mode=config.device_target == 'Ascend') print("result:", res, flush=True) - if __name__ == '__main__': crnn_eval() diff --git a/model_zoo/official/cv/crnn/export.py b/model_zoo/official/cv/crnn/export.py index 324ec090dcb..c3913e81a7e 100644 --- a/model_zoo/official/cv/crnn/export.py +++ b/model_zoo/official/cv/crnn/export.py @@ -23,6 +23,7 @@ from src.model_utils.moxing_adapter import moxing_wrapper from src.model_utils.config import config from src.model_utils.device_adapter import get_device_id +context.set_context(mode=context.GRAPH_MODE, device_target=config.device_target, save_graphs=False) def modelarts_pre_process(): config.file_name = os.path.join(config.output_path, config.file_name) diff --git a/model_zoo/official/cv/crnn/scripts/run_eval_cpu.sh b/model_zoo/official/cv/crnn/scripts/run_eval_cpu.sh new file mode 100644 index 00000000000..44f6302ed08 --- /dev/null +++ b/model_zoo/official/cv/crnn/scripts/run_eval_cpu.sh @@ -0,0 +1,58 @@ +#!/bin/bash +# Copyright 2021 Huawei Technologies Co., Ltd +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# ============================================================================ + +if [ $# != 3 ]; then + echo "Usage: bash run_eval_cpu.sh [DATASET_NAME] [DATASET_PATH] [CHECKPOINT_PATH]" + exit 1 +fi + +get_real_path() { + if [ "${1:0:1}" == "/" ]; then + echo "$1" + else + echo "$(realpath -m $PWD/$1)" + fi +} + +DATASET_NAME=$1 +PATH1=$(get_real_path $2) +PATH2=$(get_real_path $3) + +if [ ! -d $PATH1 ]; then + echo "error: DATASET_PATH=$PATH1 is not a directory" + exit 1 +fi + +if [ ! -f $PATH2 ]; then + echo "error: CHECKPOINT_PATH=$PATH2 is not a file" + exit 1 +fi + +if [ -d "eval" ]; then + rm -rf ./eval +fi +mkdir ./eval +cp ./*.py ./eval +cp -r ./src ./eval +cp ./*yaml ./eval +cd ./eval || exit +env >env.log +python eval.py --eval_dataset=$DATASET_NAME \ + --eval_dataset_path=$PATH1 \ + --checkpoint_path=$PATH2 \ + --model_version="V2" \ + --device_target=CPU > log.txt 2>&1 & +cd .. diff --git a/model_zoo/official/cv/crnn/scripts/run_standalone_train_cpu.sh b/model_zoo/official/cv/crnn/scripts/run_standalone_train_cpu.sh new file mode 100644 index 00000000000..835d98f1b86 --- /dev/null +++ b/model_zoo/official/cv/crnn/scripts/run_standalone_train_cpu.sh @@ -0,0 +1,54 @@ +#!/bin/bash +# Copyright 2021 Huawei Technologies Co., Ltd +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# ============================================================================ + +if [ $# != 2 ]; then + echo "Usage: bash scripts/run_standalone_train_cpu.sh [DATASET_NAME] [DATASET_PATH]" + exit 1 +fi + +get_real_path() { + if [ "${1:0:1}" == "/" ]; then + echo "$1" + else + echo "$(realpath -m $PWD/$1)" + fi +} + +DATASET_NAME=$1 +PATH1=$(get_real_path $2) + +if [ ! -d $PATH1 ]; then + echo "error: DATASET_PATH=$PATH1 is not a directory" + exit 1 +fi + +if [ -d "train" ]; then + rm -rf ./train +fi +WORKDIR=./train_cpu +rm -rf $WORKDIR +mkdir $WORKDIR +cp ./*.py $WORKDIR +cp -r ./src $WORKDIR +cp ./*yaml $WORKDIR +cd $WORKDIR || exit + +env >env.log +python train.py --train_dataset=$DATASET_NAME \ + --train_dataset_path=$PATH1 \ + --device_target=CPU \ + --model_version="V2" > log.txt 2>&1 & +cd .. diff --git a/model_zoo/official/cv/crnn/src/crnn.py b/model_zoo/official/cv/crnn/src/crnn.py index e9954f09476..3726628b652 100644 --- a/model_zoo/official/cv/crnn/src/crnn.py +++ b/model_zoo/official/cv/crnn/src/crnn.py @@ -1,4 +1,4 @@ -# Copyright 2020 Huawei Technologies Co., Ltd +# Copyright 2020-21 Huawei Technologies Co., Ltd # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. @@ -14,14 +14,17 @@ # ============================================================================ """Warpctc network definition.""" +import math import numpy as np import mindspore.nn as nn from mindspore import Tensor +from mindspore import context from mindspore.common import dtype as mstype from mindspore.ops import operations as P from mindspore.ops import functional as F from mindspore.common import Parameter from mindspore.common.initializer import TruncatedNormal +from mindspore.common.initializer import initializer def _bn(channel): return nn.BatchNorm2d(channel, eps=1e-4, momentum=0.9, gamma_init=1, beta_init=0, moving_mean_init=0, @@ -42,6 +45,46 @@ class Conv(nn.Cell): out = self.Relu(out) return out +class LSTMCPU(nn.Cell): + """Stacked LSTM (Long Short-Term Memory) layers for CPU""" + def __init__(self, + input_size, + hidden_size, + num_layers=1, + has_bias=True, + dropout=0, + bidirectional=False): + super(LSTMCPU, self).__init__() + self.transpose = P.Transpose() + self.num_layers = num_layers + self.bidirectional = bidirectional + self.dropout = dropout + self.lstm = P.LSTM(input_size=input_size, + hidden_size=hidden_size, + num_layers=num_layers, + has_bias=has_bias, + bidirectional=bidirectional, + dropout=float(dropout)) + weight_size = 0 + gate_size = 4 * hidden_size + stdv = 1 / math.sqrt(hidden_size) + num_directions = 2 if bidirectional else 1 + + for layer in range(num_layers): + input_layer_size = input_size if layer == 0 else hidden_size * num_directions + increment_size = gate_size * input_layer_size + increment_size += gate_size * hidden_size + if has_bias: + increment_size += gate_size + weight_size += increment_size * num_directions + w_np = np.random.uniform(-stdv, stdv, (weight_size, 1, 1)).astype(np.float32) + self.weight = Parameter(initializer(Tensor(w_np), [weight_size, 1, 1]), name='weight') + + def construct(self, x, hx): + h, c = hx + x, h, c, _, _ = self.lstm(x, h, c, self.weight) + return x, (h, c) + class VGG(nn.Cell): """VGG Network structure""" def __init__(self, is_training=True): @@ -76,8 +119,10 @@ class BidirectionalLSTM(nn.Cell): def __init__(self, nIn, nHidden, nOut, batch_size): super(BidirectionalLSTM, self).__init__() - - self.rnn = nn.LSTM(nIn, nHidden, bidirectional=True) + if context.get_context("device_target") == "CPU": + self.rnn = LSTMCPU(nIn, nHidden, bidirectional=True) + else: + self.rnn = nn.LSTM(nIn, nHidden, bidirectional=True) self.embedding = nn.Dense(in_channels=nHidden * 2, out_channels=nOut) self.h0 = Tensor(np.zeros([1 * 2, batch_size, nHidden]).astype(np.float32)) self.c0 = Tensor(np.zeros([1 * 2, batch_size, nHidden]).astype(np.float32)) diff --git a/model_zoo/official/cv/crnn/train.py b/model_zoo/official/cv/crnn/train.py index ed482ba1e98..c7afc2ebca1 100644 --- a/model_zoo/official/cv/crnn/train.py +++ b/model_zoo/official/cv/crnn/train.py @@ -1,4 +1,4 @@ -# Copyright 2020 Huawei Technologies Co., Ltd +# Copyright 2020-21 Huawei Technologies Co., Ltd # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. @@ -88,7 +88,7 @@ def train(): loss = CTCLoss(max_sequence_length=config.num_step, max_label_length=max_text_length, batch_size=config.batch_size) - net = crnn(config, full_precision=config.device_target == 'GPU') + net = crnn(config, full_precision=config.device_target != 'Ascend') opt = nn.SGD(params=net.trainable_params(), learning_rate=lr, momentum=config.momentum, nesterov=config.nesterov) net_with_loss = WithLossCell(net, loss)