未验证 提交 ff185e60 编写于 作者: D Double_V 提交者: GitHub

Merge pull request #4270 from LDOUBLEV/sdmgr

[SDMGR]add kie
Global:
use_gpu: True
epoch_num: 300
log_smooth_window: 20
print_batch_step: 50
save_model_dir: ./output/kie_5/
save_epoch_step: 50
# evaluation is run every 5000 iterations after the 4000th iteration
eval_batch_step: [ 0, 80 ]
# 1. If pretrained_model is saved in static mode, such as classification pretrained model
# from static branch, load_static_weights must be set as True.
# 2. If you want to finetune the pretrained models we provide in the docs,
# you should set load_static_weights as False.
load_static_weights: False
cal_metric_during_train: False
pretrained_model: ./output/kie_4/best_accuracy
checkpoints:
save_inference_dir:
use_visualdl: False
class_path: ./train_data/wildreceipt/class_list.txt
infer_img: ./train_data/wildreceipt/1.txt
save_res_path: ./output/sdmgr_kie/predicts_kie.txt
img_scale: [ 1024, 512 ]
Architecture:
model_type: kie
algorithm: SDMGR
Transform:
Backbone:
name: Kie_backbone
Head:
name: SDMGRHead
Loss:
name: SDMGRLoss
Optimizer:
name: Adam
beta1: 0.9
beta2: 0.999
lr:
name: Piecewise
learning_rate: 0.001
decay_epochs: [ 60, 80, 100]
values: [ 0.001, 0.0001, 0.00001]
warmup_epoch: 2
regularizer:
name: 'L2'
factor: 0.00005
PostProcess:
name: None
Metric:
name: KIEMetric
main_indicator: hmean
Train:
dataset:
name: SimpleDataSet
data_dir: ./train_data/wildreceipt/
label_file_list: [ './train_data/wildreceipt/wildreceipt_train.txt' ]
ratio_list: [ 1.0 ]
transforms:
- DecodeImage: # load image
img_mode: RGB
channel_first: False
- NormalizeImage:
scale: 1
mean: [ 123.675, 116.28, 103.53 ]
std: [ 58.395, 57.12, 57.375 ]
order: 'hwc'
- KieLabelEncode: # Class handling label
character_dict_path: ./train_data/wildreceipt/dict.txt
- KieResize:
- ToCHWImage:
- KeepKeys:
keep_keys: [ 'image', 'relations', 'texts', 'points', 'labels', 'tag', 'shape'] # dataloader will return list in this order
loader:
shuffle: True
drop_last: False
batch_size_per_card: 4
num_workers: 4
Eval:
dataset:
name: SimpleDataSet
data_dir: ./train_data/wildreceipt
label_file_list:
- ./train_data/wildreceipt/wildreceipt_test.txt
# - /paddle/data/PaddleOCR/train_data/wildreceipt/1.txt
transforms:
- DecodeImage: # load image
img_mode: RGB
channel_first: False
- KieLabelEncode: # Class handling label
character_dict_path: ./train_data/wildreceipt/dict.txt
- KieResize:
- NormalizeImage:
scale: 1
mean: [ 123.675, 116.28, 103.53 ]
std: [ 58.395, 57.12, 57.375 ]
order: 'hwc'
- ToCHWImage:
- KeepKeys:
keep_keys: [ 'image', 'relations', 'texts', 'points', 'labels', 'tag', 'ori_image', 'ori_boxes', 'shape']
loader:
shuffle: False
drop_last: False
batch_size_per_card: 1 # must be 1
num_workers: 4
\ No newline at end of file
......@@ -21,6 +21,7 @@ PaddleOCR开源的文本检测算法列表:
- [x] EAST([paper](https://arxiv.org/abs/1704.03155))[1]
- [x] SAST([paper](https://arxiv.org/abs/1908.05498))[4]
- [x] PSENet([paper](https://arxiv.org/abs/1903.12473v2)
- [x] SDMGR([paper](https://arxiv.org/pdf/2103.14470.pdf))
在ICDAR2015文本检测公开数据集上,算法效果如下:
|模型|骨干网络|precision|recall|Hmean|下载链接|
......@@ -32,6 +33,7 @@ PaddleOCR开源的文本检测算法列表:
|SAST|ResNet50_vd|91.39%|83.77%|87.42%|[训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_sast_icdar15_v2.0_train.tar)|
|PSE|ResNet50_vd|85.81%|79.53%|82.55%|[训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.1/en_det/det_r50_vd_pse_v2.0_train.tar)|
|PSE|MobileNetV3|82.20%|70.48%|75.89%|[训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.1/en_det/det_mv3_pse_v2.0_train.tar)|
|SDMGR|VGG16|-|-|87.11%|[训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.1/kie/kie_vgg16.tar)|
在Total-text文本检测公开数据集上,算法效果如下:
......
......@@ -19,6 +19,7 @@ from __future__ import unicode_literals
import numpy as np
import string
from shapely.geometry import LineString, Point, Polygon
import json
from ppocr.utils.logging import get_logger
......@@ -286,6 +287,168 @@ class E2ELabelEncodeTrain(object):
return data
class KieLabelEncode(object):
def __init__(self, character_dict_path, norm=10, directed=False, **kwargs):
super(KieLabelEncode, self).__init__()
self.dict = dict({'': 0})
with open(character_dict_path, 'r') as fr:
idx = 1
for line in fr:
char = line.strip()
self.dict[char] = idx
idx += 1
self.norm = norm
self.directed = directed
def compute_relation(self, boxes):
"""Compute relation between every two boxes."""
x1s, y1s = boxes[:, 0:1], boxes[:, 1:2]
x2s, y2s = boxes[:, 4:5], boxes[:, 5:6]
ws, hs = x2s - x1s + 1, np.maximum(y2s - y1s + 1, 1)
dxs = (x1s[:, 0][None] - x1s) / self.norm
dys = (y1s[:, 0][None] - y1s) / self.norm
xhhs, xwhs = hs[:, 0][None] / hs, ws[:, 0][None] / hs
whs = ws / hs + np.zeros_like(xhhs)
relations = np.stack([dxs, dys, whs, xhhs, xwhs], -1)
bboxes = np.concatenate([x1s, y1s, x2s, y2s], -1).astype(np.float32)
return relations, bboxes
def pad_text_indices(self, text_inds):
"""Pad text index to same length."""
max_len = 300
recoder_len = max([len(text_ind) for text_ind in text_inds])
padded_text_inds = -np.ones((len(text_inds), max_len), np.int32)
for idx, text_ind in enumerate(text_inds):
padded_text_inds[idx, :len(text_ind)] = np.array(text_ind)
return padded_text_inds, recoder_len
def list_to_numpy(self, ann_infos):
"""Convert bboxes, relations, texts and labels to ndarray."""
boxes, text_inds = ann_infos['points'], ann_infos['text_inds']
boxes = np.array(boxes, np.int32)
relations, bboxes = self.compute_relation(boxes)
labels = ann_infos.get('labels', None)
if labels is not None:
labels = np.array(labels, np.int32)
edges = ann_infos.get('edges', None)
if edges is not None:
labels = labels[:, None]
edges = np.array(edges)
edges = (edges[:, None] == edges[None, :]).astype(np.int32)
if self.directed:
edges = (edges & labels == 1).astype(np.int32)
np.fill_diagonal(edges, -1)
labels = np.concatenate([labels, edges], -1)
padded_text_inds, recoder_len = self.pad_text_indices(text_inds)
max_num = 300
temp_bboxes = np.zeros([max_num, 4])
h, _ = bboxes.shape
temp_bboxes[:h, :h] = bboxes
temp_relations = np.zeros([max_num, max_num, 5])
temp_relations[:h, :h, :] = relations
temp_padded_text_inds = np.zeros([max_num, max_num])
temp_padded_text_inds[:h, :] = padded_text_inds
temp_labels = np.zeros([max_num, max_num])
temp_labels[:h, :h + 1] = labels
tag = np.array([h, recoder_len])
return dict(
image=ann_infos['image'],
points=temp_bboxes,
relations=temp_relations,
texts=temp_padded_text_inds,
labels=temp_labels,
tag=tag)
def convert_canonical(self, points_x, points_y):
assert len(points_x) == 4
assert len(points_y) == 4
points = [Point(points_x[i], points_y[i]) for i in range(4)]
polygon = Polygon([(p.x, p.y) for p in points])
min_x, min_y, _, _ = polygon.bounds
points_to_lefttop = [
LineString([points[i], Point(min_x, min_y)]) for i in range(4)
]
distances = np.array([line.length for line in points_to_lefttop])
sort_dist_idx = np.argsort(distances)
lefttop_idx = sort_dist_idx[0]
if lefttop_idx == 0:
point_orders = [0, 1, 2, 3]
elif lefttop_idx == 1:
point_orders = [1, 2, 3, 0]
elif lefttop_idx == 2:
point_orders = [2, 3, 0, 1]
else:
point_orders = [3, 0, 1, 2]
sorted_points_x = [points_x[i] for i in point_orders]
sorted_points_y = [points_y[j] for j in point_orders]
return sorted_points_x, sorted_points_y
def sort_vertex(self, points_x, points_y):
assert len(points_x) == 4
assert len(points_y) == 4
x = np.array(points_x)
y = np.array(points_y)
center_x = np.sum(x) * 0.25
center_y = np.sum(y) * 0.25
x_arr = np.array(x - center_x)
y_arr = np.array(y - center_y)
angle = np.arctan2(y_arr, x_arr) * 180.0 / np.pi
sort_idx = np.argsort(angle)
sorted_points_x, sorted_points_y = [], []
for i in range(4):
sorted_points_x.append(points_x[sort_idx[i]])
sorted_points_y.append(points_y[sort_idx[i]])
return self.convert_canonical(sorted_points_x, sorted_points_y)
def __call__(self, data):
import json
label = data['label']
annotations = json.loads(label)
boxes, texts, text_inds, labels, edges = [], [], [], [], []
for ann in annotations:
box = ann['points']
x_list = [box[i][0] for i in range(4)]
y_list = [box[i][1] for i in range(4)]
sorted_x_list, sorted_y_list = self.sort_vertex(x_list, y_list)
sorted_box = []
for x, y in zip(sorted_x_list, sorted_y_list):
sorted_box.append(x)
sorted_box.append(y)
boxes.append(sorted_box)
text = ann['transcription']
texts.append(ann['transcription'])
text_ind = [self.dict[c] for c in text if c in self.dict]
text_inds.append(text_ind)
labels.append(ann['label'])
edges.append(ann.get('edge', 0))
ann_infos = dict(
image=data['image'],
points=boxes,
texts=texts,
text_inds=text_inds,
edges=edges,
labels=labels)
return self.list_to_numpy(ann_infos)
class AttnLabelEncode(BaseRecLabelEncode):
""" Convert between text-label and text-index """
......
......@@ -111,7 +111,6 @@ class NormalizeImage(object):
from PIL import Image
if isinstance(img, Image.Image):
img = np.array(img)
assert isinstance(img,
np.ndarray), "invalid input 'img' in NormalizeImage"
data['image'] = (
......@@ -367,3 +366,53 @@ class E2EResizeForTest(object):
ratio_w = resize_w / float(w)
return im, (ratio_h, ratio_w)
class KieResize(object):
def __init__(self, **kwargs):
super(KieResize, self).__init__()
self.max_side, self.min_side = kwargs['img_scale'][0], kwargs[
'img_scale'][1]
def __call__(self, data):
img = data['image']
points = data['points']
src_h, src_w, _ = img.shape
im_resized, scale_factor, [ratio_h, ratio_w
], [new_h, new_w] = self.resize_image(img)
resize_points = self.resize_boxes(img, points, scale_factor)
data['ori_image'] = img
data['ori_boxes'] = points
data['points'] = resize_points
data['image'] = im_resized
data['shape'] = np.array([new_h, new_w])
return data
def resize_image(self, img):
norm_img = np.zeros([1024, 1024, 3], dtype='float32')
scale = [512, 1024]
h, w = img.shape[:2]
max_long_edge = max(scale)
max_short_edge = min(scale)
scale_factor = min(max_long_edge / max(h, w),
max_short_edge / min(h, w))
resize_w, resize_h = int(w * float(scale_factor) + 0.5), int(h * float(
scale_factor) + 0.5)
max_stride = 32
resize_h = (resize_h + max_stride - 1) // max_stride * max_stride
resize_w = (resize_w + max_stride - 1) // max_stride * max_stride
im = cv2.resize(img, (resize_w, resize_h))
new_h, new_w = im.shape[:2]
w_scale = new_w / w
h_scale = new_h / h
scale_factor = np.array(
[w_scale, h_scale, w_scale, h_scale], dtype=np.float32)
norm_img[:new_h, :new_w, :] = im
return norm_img, scale_factor, [h_scale, w_scale], [new_h, new_w]
def resize_boxes(self, im, points, scale_factor):
points = points * scale_factor
img_shape = im.shape[:2]
points[:, 0::2] = np.clip(points[:, 0::2], 0, img_shape[1])
points[:, 1::2] = np.clip(points[:, 1::2], 0, img_shape[0])
return points
......@@ -35,6 +35,7 @@ from .cls_loss import ClsLoss
# e2e loss
from .e2e_pg_loss import PGLoss
from .kie_sdmgr_loss import SDMGRLoss
# basic loss function
from .basic_loss import DistanceLoss
......@@ -50,7 +51,7 @@ def build_loss(config):
support_dict = [
'DBLoss', 'PSELoss', 'EASTLoss', 'SASTLoss', 'CTCLoss', 'ClsLoss',
'AttentionLoss', 'SRNLoss', 'PGLoss', 'CombinedLoss', 'NRTRLoss',
'TableAttentionLoss', 'SARLoss', 'AsterLoss'
'TableAttentionLoss', 'SARLoss', 'AsterLoss', 'SDMGRLoss'
]
config = copy.deepcopy(config)
module_name = config.pop('name')
......
# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
#
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from paddle import nn
import paddle
class SDMGRLoss(nn.Layer):
def __init__(self, node_weight=1.0, edge_weight=1.0, ignore=0):
super().__init__()
self.loss_node = nn.CrossEntropyLoss(ignore_index=ignore)
self.loss_edge = nn.CrossEntropyLoss(ignore_index=-1)
self.node_weight = node_weight
self.edge_weight = edge_weight
self.ignore = ignore
def pre_process(self, gts, tag):
gts, tag = gts.numpy(), tag.numpy().tolist()
temp_gts = []
batch = len(tag)
for i in range(batch):
num, recoder_len = tag[i][0], tag[i][1]
temp_gts.append(
paddle.to_tensor(
gts[i, :num, :num + 1], dtype='int64'))
return temp_gts
def accuracy(self, pred, target, topk=1, thresh=None):
"""Calculate accuracy according to the prediction and target.
Args:
pred (torch.Tensor): The model prediction, shape (N, num_class)
target (torch.Tensor): The target of each prediction, shape (N, )
topk (int | tuple[int], optional): If the predictions in ``topk``
matches the target, the predictions will be regarded as
correct ones. Defaults to 1.
thresh (float, optional): If not None, predictions with scores under
this threshold are considered incorrect. Default to None.
Returns:
float | tuple[float]: If the input ``topk`` is a single integer,
the function will return a single float as accuracy. If
``topk`` is a tuple containing multiple integers, the
function will return a tuple containing accuracies of
each ``topk`` number.
"""
assert isinstance(topk, (int, tuple))
if isinstance(topk, int):
topk = (topk, )
return_single = True
else:
return_single = False
maxk = max(topk)
if pred.shape[0] == 0:
accu = [pred.new_tensor(0.) for i in range(len(topk))]
return accu[0] if return_single else accu
pred_value, pred_label = paddle.topk(pred, maxk, axis=1)
pred_label = pred_label.transpose(
[1, 0]) # transpose to shape (maxk, N)
correct = paddle.equal(pred_label,
(target.reshape([1, -1]).expand_as(pred_label)))
res = []
for k in topk:
correct_k = paddle.sum(correct[:k].reshape([-1]).astype('float32'),
axis=0,
keepdim=True)
res.append(
paddle.multiply(correct_k,
paddle.to_tensor(100.0 / pred.shape[0])))
return res[0] if return_single else res
def forward(self, pred, batch):
node_preds, edge_preds = pred
gts, tag = batch[4], batch[5]
gts = self.pre_process(gts, tag)
node_gts, edge_gts = [], []
for gt in gts:
node_gts.append(gt[:, 0])
edge_gts.append(gt[:, 1:].reshape([-1]))
node_gts = paddle.concat(node_gts)
edge_gts = paddle.concat(edge_gts)
node_valids = paddle.nonzero(node_gts != self.ignore).reshape([-1])
edge_valids = paddle.nonzero(edge_gts != -1).reshape([-1])
loss_node = self.loss_node(node_preds, node_gts)
loss_edge = self.loss_edge(edge_preds, edge_gts)
loss = self.node_weight * loss_node + self.edge_weight * loss_edge
return dict(
loss=loss,
loss_node=loss_node,
loss_edge=loss_edge,
acc_node=self.accuracy(
paddle.gather(node_preds, node_valids),
paddle.gather(node_gts, node_valids)),
acc_edge=self.accuracy(
paddle.gather(edge_preds, edge_valids),
paddle.gather(edge_gts, edge_valids)))
......@@ -27,10 +27,13 @@ from .cls_metric import ClsMetric
from .e2e_metric import E2EMetric
from .distillation_metric import DistillationMetric
from .table_metric import TableMetric
from .kie_metric import KIEMetric
def build_metric(config):
support_dict = [
"DetMetric", "RecMetric", "ClsMetric", "E2EMetric", "DistillationMetric", "TableMetric"
"DetMetric", "RecMetric", "ClsMetric", "E2EMetric",
"DistillationMetric", "TableMetric", 'KIEMetric'
]
config = copy.deepcopy(config)
......
# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
#
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import numpy as np
import paddle
__all__ = ['KIEMetric']
class KIEMetric(object):
def __init__(self, main_indicator='hmean', **kwargs):
self.main_indicator = main_indicator
self.reset()
self.node = []
self.gt = []
def __call__(self, preds, batch, **kwargs):
nodes, _ = preds
gts, tag = batch[4].squeeze(0), batch[5].tolist()[0]
gts = gts[:tag[0], :1].reshape([-1])
self.node.append(nodes.numpy())
self.gt.append(gts)
# result = self.compute_f1_score(nodes, gts)
# self.results.append(result)
def compute_f1_score(self, preds, gts):
ignores = [0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 25]
C = preds.shape[1]
classes = np.array(sorted(set(range(C)) - set(ignores)))
hist = np.bincount(
(gts * C).astype('int64') + preds.argmax(1), minlength=C
**2).reshape([C, C]).astype('float32')
diag = np.diag(hist)
recalls = diag / hist.sum(1).clip(min=1)
precisions = diag / hist.sum(0).clip(min=1)
f1 = 2 * recalls * precisions / (recalls + precisions).clip(min=1e-8)
return f1[classes]
def combine_results(self, results):
node = np.concatenate(self.node, 0)
gts = np.concatenate(self.gt, 0)
results = self.compute_f1_score(node, gts)
data = {'hmean': results.mean()}
return data
def get_metric(self):
metircs = self.combine_results(self.results)
self.reset()
return metircs
def reset(self):
self.results = [] # clear results
self.node = []
self.gt = []
......@@ -35,7 +35,14 @@ def build_backbone(config, model_type):
]
elif model_type == "e2e":
from .e2e_resnet_vd_pg import ResNet
support_dict = ["ResNet"]
support_dict = ['ResNet']
elif model_type == 'kie':
from .kie_unet_sdmgr import Kie_backbone
support_dict = ['Kie_backbone']
elif model_type == "table":
from .table_resnet_vd import ResNet
from .table_mobilenet_v3 import MobileNetV3
support_dict = ["ResNet", "MobileNetV3"]
else:
raise NotImplementedError
......
# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
#
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import paddle
from paddle import nn
import numpy as np
import cv2
__all__ = ["Kie_backbone"]
class Encoder(nn.Layer):
def __init__(self, num_channels, num_filters):
super(Encoder, self).__init__()
self.conv1 = nn.Conv2D(
num_channels,
num_filters,
kernel_size=3,
stride=1,
padding=1,
bias_attr=False)
self.bn1 = nn.BatchNorm(num_filters, act='relu')
self.conv2 = nn.Conv2D(
num_filters,
num_filters,
kernel_size=3,
stride=1,
padding=1,
bias_attr=False)
self.bn2 = nn.BatchNorm(num_filters, act='relu')
self.pool = nn.MaxPool2D(kernel_size=3, stride=2, padding=1)
def forward(self, inputs):
x = self.conv1(inputs)
x = self.bn1(x)
x = self.conv2(x)
x = self.bn2(x)
x_pooled = self.pool(x)
return x, x_pooled
class Decoder(nn.Layer):
def __init__(self, num_channels, num_filters):
super(Decoder, self).__init__()
self.conv1 = nn.Conv2D(
num_channels,
num_filters,
kernel_size=3,
stride=1,
padding=1,
bias_attr=False)
self.bn1 = nn.BatchNorm(num_filters, act='relu')
self.conv2 = nn.Conv2D(
num_filters,
num_filters,
kernel_size=3,
stride=1,
padding=1,
bias_attr=False)
self.bn2 = nn.BatchNorm(num_filters, act='relu')
self.conv0 = nn.Conv2D(
num_channels,
num_filters,
kernel_size=1,
stride=1,
padding=0,
bias_attr=False)
self.bn0 = nn.BatchNorm(num_filters, act='relu')
def forward(self, inputs_prev, inputs):
x = self.conv0(inputs)
x = self.bn0(x)
x = paddle.nn.functional.interpolate(
x, scale_factor=2, mode='bilinear', align_corners=False)
x = paddle.concat([inputs_prev, x], axis=1)
x = self.conv1(x)
x = self.bn1(x)
x = self.conv2(x)
x = self.bn2(x)
return x
class UNet(nn.Layer):
def __init__(self):
super(UNet, self).__init__()
self.down1 = Encoder(num_channels=3, num_filters=16)
self.down2 = Encoder(num_channels=16, num_filters=32)
self.down3 = Encoder(num_channels=32, num_filters=64)
self.down4 = Encoder(num_channels=64, num_filters=128)
self.down5 = Encoder(num_channels=128, num_filters=256)
self.up1 = Decoder(32, 16)
self.up2 = Decoder(64, 32)
self.up3 = Decoder(128, 64)
self.up4 = Decoder(256, 128)
self.out_channels = 16
def forward(self, inputs):
x1, _ = self.down1(inputs)
_, x2 = self.down2(x1)
_, x3 = self.down3(x2)
_, x4 = self.down4(x3)
_, x5 = self.down5(x4)
x = self.up4(x4, x5)
x = self.up3(x3, x)
x = self.up2(x2, x)
x = self.up1(x1, x)
return x
class Kie_backbone(nn.Layer):
def __init__(self, in_channels, **kwargs):
super(Kie_backbone, self).__init__()
self.out_channels = 16
self.img_feat = UNet()
self.maxpool = nn.MaxPool2D(kernel_size=7)
def bbox2roi(self, bbox_list):
rois_list = []
rois_num = []
for img_id, bboxes in enumerate(bbox_list):
rois_num.append(bboxes.shape[0])
rois_list.append(bboxes)
rois = paddle.concat(rois_list, 0)
rois_num = paddle.to_tensor(rois_num, dtype='int32')
return rois, rois_num
def pre_process(self, img, relations, texts, gt_bboxes, tag, img_size):
img, relations, texts, gt_bboxes, tag, img_size = img.numpy(
), relations.numpy(), texts.numpy(), gt_bboxes.numpy(), tag.numpy(
).tolist(), img_size.numpy()
temp_relations, temp_texts, temp_gt_bboxes = [], [], []
h, w = int(np.max(img_size[:, 0])), int(np.max(img_size[:, 1]))
img = paddle.to_tensor(img[:, :, :h, :w])
batch = len(tag)
for i in range(batch):
num, recoder_len = tag[i][0], tag[i][1]
temp_relations.append(
paddle.to_tensor(
relations[i, :num, :num, :], dtype='float32'))
temp_texts.append(
paddle.to_tensor(
texts[i, :num, :recoder_len], dtype='float32'))
temp_gt_bboxes.append(
paddle.to_tensor(
gt_bboxes[i, :num, ...], dtype='float32'))
return img, temp_relations, temp_texts, temp_gt_bboxes
def forward(self, inputs):
img = inputs[0]
relations, texts, gt_bboxes, tag, img_size = inputs[1], inputs[
2], inputs[3], inputs[5], inputs[-1]
img, relations, texts, gt_bboxes = self.pre_process(
img, relations, texts, gt_bboxes, tag, img_size)
x = self.img_feat(img)
boxes, rois_num = self.bbox2roi(gt_bboxes)
feats = paddle.fluid.layers.roi_align(
x,
boxes,
spatial_scale=1.0,
pooled_height=7,
pooled_width=7,
rois_num=rois_num)
feats = self.maxpool(feats).squeeze(-1).squeeze(-1)
return [relations, texts, feats]
......@@ -33,14 +33,19 @@ def build_head(config):
# cls head
from .cls_head import ClsHead
#kie head
from .kie_sdmgr_head import SDMGRHead
from .table_att_head import TableAttentionHead
support_dict = [
'DBHead', 'PSEHead', 'EASTHead', 'SASTHead', 'CTCHead', 'ClsHead',
'AttentionHead', 'SRNHead', 'PGHead', 'Transformer',
'TableAttentionHead', 'SARHead', 'AsterHead'
'TableAttentionHead', 'SARHead', 'AsterHead', 'SDMGRHead'
]
#table head
from .table_att_head import TableAttentionHead
module_name = config.pop('name')
assert module_name in support_dict, Exception('head only support {}'.format(
......
# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
#
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import math
import paddle
from paddle import nn
import paddle.nn.functional as F
from paddle import ParamAttr
class SDMGRHead(nn.Layer):
def __init__(self,
in_channels,
num_chars=92,
visual_dim=16,
fusion_dim=1024,
node_input=32,
node_embed=256,
edge_input=5,
edge_embed=256,
num_gnn=2,
num_classes=26,
bidirectional=False):
super().__init__()
self.fusion = Block([visual_dim, node_embed], node_embed, fusion_dim)
self.node_embed = nn.Embedding(num_chars, node_input, 0)
hidden = node_embed // 2 if bidirectional else node_embed
self.rnn = nn.LSTM(
input_size=node_input, hidden_size=hidden, num_layers=1)
self.edge_embed = nn.Linear(edge_input, edge_embed)
self.gnn_layers = nn.LayerList(
[GNNLayer(node_embed, edge_embed) for _ in range(num_gnn)])
self.node_cls = nn.Linear(node_embed, num_classes)
self.edge_cls = nn.Linear(edge_embed, 2)
def forward(self, input, targets):
relations, texts, x = input
node_nums, char_nums = [], []
for text in texts:
node_nums.append(text.shape[0])
char_nums.append(paddle.sum((text > -1).astype(int), axis=-1))
max_num = max([char_num.max() for char_num in char_nums])
all_nodes = paddle.concat([
paddle.concat(
[text, paddle.zeros(
(text.shape[0], max_num - text.shape[1]))], -1)
for text in texts
])
temp = paddle.clip(all_nodes, min=0).astype(int)
embed_nodes = self.node_embed(temp)
rnn_nodes, _ = self.rnn(embed_nodes)
b, h, w = rnn_nodes.shape
nodes = paddle.zeros([b, w])
all_nums = paddle.concat(char_nums)
valid = paddle.nonzero((all_nums > 0).astype(int))
temp_all_nums = (
paddle.gather(all_nums, valid) - 1).unsqueeze(-1).unsqueeze(-1)
temp_all_nums = paddle.expand(temp_all_nums, [
temp_all_nums.shape[0], temp_all_nums.shape[1], rnn_nodes.shape[-1]
])
temp_all_nodes = paddle.gather(rnn_nodes, valid)
N, C, A = temp_all_nodes.shape
one_hot = F.one_hot(
temp_all_nums[:, 0, :], num_classes=C).transpose([0, 2, 1])
one_hot = paddle.multiply(
temp_all_nodes, one_hot.astype("float32")).sum(axis=1, keepdim=True)
t = one_hot.expand([N, 1, A]).squeeze(1)
nodes = paddle.scatter(nodes, valid.squeeze(1), t)
if x is not None:
nodes = self.fusion([x, nodes])
all_edges = paddle.concat(
[rel.reshape([-1, rel.shape[-1]]) for rel in relations])
embed_edges = self.edge_embed(all_edges.astype('float32'))
embed_edges = F.normalize(embed_edges)
for gnn_layer in self.gnn_layers:
nodes, cat_nodes = gnn_layer(nodes, embed_edges, node_nums)
node_cls, edge_cls = self.node_cls(nodes), self.edge_cls(cat_nodes)
return node_cls, edge_cls
class GNNLayer(nn.Layer):
def __init__(self, node_dim=256, edge_dim=256):
super().__init__()
self.in_fc = nn.Linear(node_dim * 2 + edge_dim, node_dim)
self.coef_fc = nn.Linear(node_dim, 1)
self.out_fc = nn.Linear(node_dim, node_dim)
self.relu = nn.ReLU()
def forward(self, nodes, edges, nums):
start, cat_nodes = 0, []
for num in nums:
sample_nodes = nodes[start:start + num]
cat_nodes.append(
paddle.concat([
paddle.expand(sample_nodes.unsqueeze(1), [-1, num, -1]),
paddle.expand(sample_nodes.unsqueeze(0), [num, -1, -1])
], -1).reshape([num**2, -1]))
start += num
cat_nodes = paddle.concat([paddle.concat(cat_nodes), edges], -1)
cat_nodes = self.relu(self.in_fc(cat_nodes))
coefs = self.coef_fc(cat_nodes)
start, residuals = 0, []
for num in nums:
residual = F.softmax(
-paddle.eye(num).unsqueeze(-1) * 1e9 +
coefs[start:start + num**2].reshape([num, num, -1]), 1)
residuals.append((residual * cat_nodes[start:start + num**2]
.reshape([num, num, -1])).sum(1))
start += num**2
nodes += self.relu(self.out_fc(paddle.concat(residuals)))
return [nodes, cat_nodes]
class Block(nn.Layer):
def __init__(self,
input_dims,
output_dim,
mm_dim=1600,
chunks=20,
rank=15,
shared=False,
dropout_input=0.,
dropout_pre_lin=0.,
dropout_output=0.,
pos_norm='before_cat'):
super().__init__()
self.rank = rank
self.dropout_input = dropout_input
self.dropout_pre_lin = dropout_pre_lin
self.dropout_output = dropout_output
assert (pos_norm in ['before_cat', 'after_cat'])
self.pos_norm = pos_norm
# Modules
self.linear0 = nn.Linear(input_dims[0], mm_dim)
self.linear1 = (self.linear0
if shared else nn.Linear(input_dims[1], mm_dim))
self.merge_linears0 = nn.LayerList()
self.merge_linears1 = nn.LayerList()
self.chunks = self.chunk_sizes(mm_dim, chunks)
for size in self.chunks:
ml0 = nn.Linear(size, size * rank)
self.merge_linears0.append(ml0)
ml1 = ml0 if shared else nn.Linear(size, size * rank)
self.merge_linears1.append(ml1)
self.linear_out = nn.Linear(mm_dim, output_dim)
def forward(self, x):
x0 = self.linear0(x[0])
x1 = self.linear1(x[1])
bs = x1.shape[0]
if self.dropout_input > 0:
x0 = F.dropout(x0, p=self.dropout_input, training=self.training)
x1 = F.dropout(x1, p=self.dropout_input, training=self.training)
x0_chunks = paddle.split(x0, self.chunks, -1)
x1_chunks = paddle.split(x1, self.chunks, -1)
zs = []
for x0_c, x1_c, m0, m1 in zip(x0_chunks, x1_chunks, self.merge_linears0,
self.merge_linears1):
m = m0(x0_c) * m1(x1_c) # bs x split_size*rank
m = m.reshape([bs, self.rank, -1])
z = paddle.sum(m, 1)
if self.pos_norm == 'before_cat':
z = paddle.sqrt(F.relu(z)) - paddle.sqrt(F.relu(-z))
z = F.normalize(z)
zs.append(z)
z = paddle.concat(zs, 1)
if self.pos_norm == 'after_cat':
z = paddle.sqrt(F.relu(z)) - paddle.sqrt(F.relu(-z))
z = F.normalize(z)
if self.dropout_pre_lin > 0:
z = F.dropout(z, p=self.dropout_pre_lin, training=self.training)
z = self.linear_out(z)
if self.dropout_output > 0:
z = F.dropout(z, p=self.dropout_output, training=self.training)
return z
def chunk_sizes(self, dim, chunks):
split_size = (dim + chunks - 1) // chunks
sizes_list = [split_size] * chunks
sizes_list[-1] = sizes_list[-1] - (sum(sizes_list) - dim)
return sizes_list
......@@ -45,6 +45,8 @@ def build_post_process(config, global_config=None):
config = copy.deepcopy(config)
module_name = config.pop('name')
if module_name == "None":
return
if global_config is not None:
config.update(global_config)
assert module_name in support_dict, Exception(
......
......@@ -54,7 +54,8 @@ def main():
config['Architecture']["Head"]['out_channels'] = char_num
model = build_model(config['Architecture'])
extra_input = config['Architecture']['algorithm'] in ["SRN", "SAR"]
extra_input = config['Architecture'][
'algorithm'] in ["SRN", "NRTR", "SAR", "SEED"]
if "model_type" in config['Architecture'].keys():
model_type = config['Architecture']['model_type']
else:
......@@ -68,7 +69,6 @@ def main():
# build metric
eval_class = build_metric(config['Metric'])
# start eval
metric = program.eval(model, valid_dataloader, post_process_class,
eval_class, model_type, extra_input)
......
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import numpy as np
import paddle.nn.functional as F
import os
import sys
__dir__ = os.path.dirname(os.path.abspath(__file__))
sys.path.append(__dir__)
sys.path.append(os.path.abspath(os.path.join(__dir__, '..')))
os.environ["FLAGS_allocator_strategy"] = 'auto_growth'
import cv2
import paddle
from ppocr.data import create_operators, transform
from ppocr.modeling.architectures import build_model
from ppocr.utils.save_load import init_model
import tools.program as program
def read_class_list(filepath):
dict = {}
with open(filepath, "r") as f:
lines = f.readlines()
for line in lines:
key, value = line.split(" ")
dict[key] = value.rstrip()
return dict
def draw_kie_result(batch, node, idx_to_cls, count):
img = batch[6].copy()
boxes = batch[7]
h, w = img.shape[:2]
pred_img = np.ones((h, w * 2, 3), dtype=np.uint8) * 255
max_value, max_idx = paddle.max(node, -1), paddle.argmax(node, -1)
node_pred_label = max_idx.numpy().tolist()
node_pred_score = max_value.numpy().tolist()
for i, box in enumerate(boxes):
if i >= len(node_pred_label):
break
new_box = [[box[0], box[1]], [box[2], box[1]], [box[2], box[3]],
[box[0], box[3]]]
Pts = np.array([new_box], np.int32)
cv2.polylines(
img, [Pts.reshape((-1, 1, 2))],
True,
color=(255, 255, 0),
thickness=1)
x_min = int(min([point[0] for point in new_box]))
y_min = int(min([point[1] for point in new_box]))
pred_label = str(node_pred_label[i])
if pred_label in idx_to_cls:
pred_label = idx_to_cls[pred_label]
pred_score = '{:.2f}'.format(node_pred_score[i])
text = pred_label + '(' + pred_score + ')'
cv2.putText(pred_img, text, (x_min * 2, y_min),
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 0, 0), 1)
vis_img = np.ones((h, w * 3, 3), dtype=np.uint8) * 255
vis_img[:, :w] = img
vis_img[:, w:] = pred_img
save_kie_path = os.path.dirname(config['Global']['save_res_path']) + "/kie_results/"
if not os.path.exists(save_kie_path):
os.makedirs(save_kie_path)
save_path = os.path.join(save_kie_path, str(count) + ".png")
cv2.imwrite(save_path, vis_img)
logger.info("The Kie Image saved in {}".format(save_path))
def main():
global_config = config['Global']
# build model
model = build_model(config['Architecture'])
init_model(config, model, logger)
# create data ops
transforms = []
for op in config['Eval']['dataset']['transforms']:
transforms.append(op)
data_dir = config['Eval']['dataset']['data_dir']
ops = create_operators(transforms, global_config)
save_res_path = config['Global']['save_res_path']
class_path = config['Global']['class_path']
idx_to_cls = read_class_list(class_path)
if not os.path.exists(os.path.dirname(save_res_path)):
os.makedirs(os.path.dirname(save_res_path))
model.eval()
with open(save_res_path, "wb") as fout:
with open(config['Global']['infer_img'], "rb") as f:
lines = f.readlines()
for index, data_line in enumerate(lines):
data_line = data_line.decode('utf-8')
substr = data_line.strip("\n").split("\t")
img_path, label = data_dir + "/" + substr[0], substr[1]
data = {'img_path': img_path, 'label': label}
with open(data['img_path'], 'rb') as f:
img = f.read()
data['image'] = img
batch = transform(data, ops)
batch_pred = [0] * len(batch)
for i in range(len(batch)):
batch_pred[i] = paddle.to_tensor(
np.expand_dims(
batch[i], axis=0))
node, edge = model(batch_pred)
node = F.softmax(node, -1)
draw_kie_result(batch, node, idx_to_cls, index)
logger.info("success!")
if __name__ == '__main__':
config, device, logger, vdl_writer = program.preprocess()
main()
......@@ -227,6 +227,10 @@ def train(config,
images = batch[0]
if use_srn:
model_average = True
if model_type == 'table' or extra_input:
preds = model(images, data=batch[1:])
if model_type == "kie":
preds = model(batch)
train_start = time.time()
# use amp
......@@ -266,7 +270,7 @@ def train(config,
if cal_metric_during_train: # only rec and cls need
batch = [item.numpy() for item in batch]
if model_type == 'table':
if model_type in ['table', 'kie']:
eval_class(preds, batch)
else:
post_result = post_process_class(preds, batch[1])
......@@ -399,17 +403,20 @@ def eval(model,
start = time.time()
if model_type == 'table' or extra_input:
preds = model(images, data=batch[1:])
if model_type == "kie":
preds = model(batch)
else:
preds = model(images)
batch = [item.numpy() for item in batch]
# Obtain usable results from post-processing methods
total_time += time.time() - start
# Evaluate the results of the current batch
if model_type == 'table':
if model_type in ['table', 'kie']:
eval_class(preds, batch)
else:
post_result = post_process_class(preds, batch[1])
eval_class(post_result, batch)
pbar.update(1)
total_frame += len(images)
# Get final metric,eg. acc or hmean
......@@ -498,8 +505,13 @@ def preprocess(is_train=False):
assert alg in [
'EAST', 'DB', 'SAST', 'Rosetta', 'CRNN', 'STARNet', 'RARE', 'SRN',
'CLS', 'PGNet', 'Distillation', 'NRTR', 'TableAttn', 'SAR', 'PSE',
'SEED'
'SEED', 'SDMGR'
]
windows_not_support_list = ['PSE']
if platform.system() == "Windows" and alg in windows_not_support_list:
logger.warning('{} is not support in Windows now'.format(
windows_not_support_list))
sys.exit()
device = 'gpu:{}'.format(dist.ParallelEnv().dev_id) if use_gpu else 'cpu'
device = paddle.set_device(device)
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册