初步修改完成,但仍然有问题

This commit is contained in:
三洋三洋
2024-03-27 04:54:49 +00:00
parent 6373e19132
commit dbf35fe9c4
3 changed files with 28 additions and 8 deletions

View File

@@ -41,6 +41,7 @@ class TexTeller(VisionEncoderDecoderModel):
if __name__ == "__main__": if __name__ == "__main__":
pause = 1
# texteller = TexTeller() # texteller = TexTeller()
# from ..utils.inference import inference # from ..utils.inference import inference
# model = TexTeller.from_pretrained('/home/lhy/code/TexTeller/src/models/ocr_model/model/ckpt') # model = TexTeller.from_pretrained('/home/lhy/code/TexTeller/src/models/ocr_model/model/ckpt')

View File

@@ -15,7 +15,7 @@ from ...globals import MAX_TOKEN_SIZE, MIN_WIDTH, MIN_HEIGHT
def train(model, tokenizer, train_dataset, eval_dataset, collate_fn_with_tokenizer): def train(model, tokenizer, train_dataset, eval_dataset, collate_fn_with_tokenizer):
training_args = TrainingArguments(**CONFIG) training_args = TrainingArguments(**CONFIG)
debug_mode = False debug_mode = True
if debug_mode: if debug_mode:
training_args.auto_find_batch_size = False training_args.auto_find_batch_size = False
training_args.num_train_epochs = 2 training_args.num_train_epochs = 2
@@ -96,6 +96,10 @@ if __name__ == '__main__':
# model = TexTeller() # model = TexTeller()
model = TexTeller.from_pretrained('/home/lhy/code/TexTeller/src/models/ocr_model/model/ckpt') model = TexTeller.from_pretrained('/home/lhy/code/TexTeller/src/models/ocr_model/model/ckpt')
# ================= debug =======================
foo = train_dataset[:3]
# ================= debug =======================
enable_train = True enable_train = True
enable_evaluate = True enable_evaluate = True
if enable_train: if enable_train:

View File

@@ -5,7 +5,6 @@ import cv2
from torchvision.transforms import v2 from torchvision.transforms import v2
from typing import List, Union from typing import List, Union
from augraphy import *
from PIL import Image from PIL import Image
from ...globals import ( from ...globals import (
@@ -15,8 +14,10 @@ from ...globals import (
IMAGE_MEAN, IMAGE_STD, IMAGE_MEAN, IMAGE_STD,
MAX_RESIZE_RATIO, MIN_RESIZE_RATIO MAX_RESIZE_RATIO, MIN_RESIZE_RATIO
) )
from .ocr_aug import ocr_augmentation_pipeline
train_pipeline = default_augraphy_pipeline(scan_only=True) # train_pipeline = default_augraphy_pipeline(scan_only=True)
train_pipeline = ocr_augmentation_pipeline()
general_transform_pipeline = v2.Compose([ general_transform_pipeline = v2.Compose([
v2.ToImage(), # Convert to tensor, only needed if you had a PIL image v2.ToImage(), # Convert to tensor, only needed if you had a PIL image
@@ -76,11 +77,24 @@ def trim_white_border(image: np.ndarray):
return trimmed_image return trimmed_image
# BUGY CODE
def add_white_border(image: np.ndarray, max_size: int) -> np.ndarray: def add_white_border(image: np.ndarray, max_size: int) -> np.ndarray:
randi = [random.randint(0, max_size) for _ in range(4)] randi = [random.randint(0, max_size) for _ in range(4)]
pad_height_size = randi[1] + randi[3]
pad_width_size = randi[0] + randi[2]
if (pad_height_size + image.shape[0] < 30):
compensate_height = int((30 - (pad_height_size + image.shape[0])) * 0.5) + 1
randi[1] += compensate_height
randi[3] += compensate_height
if (pad_width_size + image.shape[1] < 30):
compensate_width = int((30 - (pad_width_size + image.shape[1])) * 0.5) + 1
randi[0] += compensate_width
randi[2] += compensate_width
return v2.functional.pad( return v2.functional.pad(
image, torch.from_numpy(image).permute(2, 0, 1),
padding=randi padding=randi,
padding_mode='constant',
fill=(255, 255, 255)
) )
@@ -127,11 +141,12 @@ def train_transform(images: List[Image.Image]) -> List[torch.Tensor]:
# 裁剪掉白边 # 裁剪掉白边
images = [trim_white_border(image) for image in images] images = [trim_white_border(image) for image in images]
# 增加白边 # 增加白边
images = [add_white_border(image, max_size=35) for image in images] # images = [add_white_border(image, max_size=35) for image in images]
# 数据增强 # 数据增强
images = [train_pipeline(image) for image in images] # images = [train_pipeline(image.permute(1, 2, 0).numpy()) for image in images]
# general transform pipeline # general transform pipeline
images = general_transform_pipeline(images) # imgs: List[PIL.Image.Image] images = general_transform_pipeline(images)
# images = [general_transform_pipeline(image) for image in images]
# padding to fixed size # padding to fixed size
images = padding(images, OCR_IMG_SIZE) images = padding(images, OCR_IMG_SIZE)
return images return images