完成了所有代码

src/globals.py

@@ -30,7 +30,8 @@ OCR_IMG_MAX_WIDTH  = 768
 OCR_IMG_CHANNELS = 1  # 灰度图
 
 # ocr模型训练数据集的最长token数
-MAX_TOKEN_SIZE = 600
+MAX_TOKEN_SIZE = 512     # 模型最长的embedding长度被设置成了512，所以这里必须是512
+# MAX_TOKEN_SIZE = 600
 
 # ocr模型训练时随机缩放的比例
 MAX_RESIZE_RATIO = 1.15

src/models/ocr_model/train/train.py

@@ -32,6 +32,7 @@ def train(model, tokenizer, train_dataset, eval_dataset, collate_fn_with_tokeniz
 
 def evaluate(model, tokenizer, eval_dataset, collate_fn):
     eval_config = CONFIG.copy()
+    eval_config['predict_with_generate'] = True
     generate_config = GenerationConfig(
         max_new_tokens=MAX_TOKEN_SIZE,
         num_beams=1,
@@ -40,31 +41,22 @@ def evaluate(model, tokenizer, eval_dataset, collate_fn):
         eos_token_id=tokenizer.eos_token_id,
         bos_token_id=tokenizer.bos_token_id,
     )
-    # eval_config['use_cpu'] = True
-    eval_config['output_dir'] = 'debug_dir'
-    eval_config['predict_with_generate'] = True
-    eval_config['predict_with_generate'] = True
-    eval_config['dataloader_num_workers'] = 1
-    eval_config['jit_mode_eval'] = False
-    eval_config['torch_compile'] = False
-    eval_config['auto_find_batch_size'] = False
     eval_config['generation_config'] = generate_config
+    eval_config['auto_find_batch_size'] = False
     seq2seq_config = Seq2SeqTrainingArguments(**eval_config)
 
     trainer = Seq2SeqTrainer(
         model,
         seq2seq_config,
 
-        eval_dataset=eval_dataset.select(range(100)),
+        eval_dataset=eval_dataset,
         tokenizer=tokenizer, 
         data_collator=collate_fn,
         compute_metrics=partial(bleu_metric, tokenizer=tokenizer)
     )
 
     res = trainer.evaluate()
-    pause = 1
+    print(res)
-    ...
-    
     
 
 if __name__ == '__main__':
@@ -72,31 +64,27 @@ if __name__ == '__main__':
     script_dirpath = Path(__file__).resolve().parent
     os.chdir(script_dirpath)
 
-
     dataset = load_dataset(
         '/home/lhy/code/TeXify/src/models/ocr_model/train/dataset/latex-formulas/latex-formulas.py',
         'cleaned_formulas'
     )['train']
-    # dataset = load_dataset(
+    dataset = dataset.shuffle(seed=42)
-    #     '/home/lhy/code/TeXify/src/models/ocr_model/train/dataset/latex-formulas/latex-formulas.py',
+    dataset = dataset.flatten_indices()
-    #     'cleaned_formulas'
-    # )['train'].select(range(1000))
 
     tokenizer = TexTeller.get_tokenizer('/home/lhy/code/TeXify/src/models/tokenizer/roberta-tokenizer-550Kformulas')
 
     map_fn = partial(tokenize_fn, tokenizer=tokenizer)
     tokenized_dataset = dataset.map(map_fn, batched=True, remove_columns=dataset.column_names, num_proc=8, load_from_cache_file=True)
-    # tokenized_dataset = dataset.map(map_fn, batched=True, remove_columns=dataset.column_names, num_proc=1)
     tokenized_dataset = tokenized_dataset.with_transform(img_transform_fn)
 
     split_dataset = tokenized_dataset.train_test_split(test_size=0.05, seed=42)    
     train_dataset, eval_dataset = split_dataset['train'], split_dataset['test']
     collate_fn_with_tokenizer = partial(collate_fn, tokenizer=tokenizer)
     # model = TexTeller()
-    model = TexTeller.from_pretrained('/home/lhy/code/TeXify/src/models/ocr_model/train/train_result/checkpoint-80500')
+    model = TexTeller.from_pretrained('/home/lhy/code/TeXify/src/models/ocr_model/train/train_result/bugy_train_without_random_resize/checkpoint-82000')
 
-    enable_train = False
+    enable_train = True
-    enable_evaluate = True
+    enable_evaluate = False
     if enable_train:
         train(model, tokenizer, train_dataset, eval_dataset, collate_fn_with_tokenizer)  
     if enable_evaluate:
@@ -104,42 +92,3 @@ if __name__ == '__main__':
 
 
     os.chdir(cur_path)
-
-
-'''
-if __name__ == '__main__':
-    cur_path = os.getcwd()
-    script_dirpath = Path(__file__).resolve().parent
-    os.chdir(script_dirpath)
-
-
-    dataset = load_dataset(
-        '/home/lhy/code/TeXify/src/models/ocr_model/train/dataset/latex-formulas/latex-formulas.py',
-        'cleaned_formulas'
-    )['train']
-
-    pause = dataset[0]['image']
-    tokenizer = TexTeller.get_tokenizer('/home/lhy/code/TeXify/src/models/tokenizer/roberta-tokenizer-550Kformulas')
-
-    map_fn = partial(tokenize_fn, tokenizer=tokenizer)
-    tokenized_dataset = dataset.map(map_fn, batched=True, remove_columns=dataset.column_names, num_proc=8)
-    tokenized_dataset = tokenized_dataset.with_transform(img_preprocess)
-
-    split_dataset = tokenized_dataset.train_test_split(test_size=0.05, seed=42)    
-    train_dataset, eval_dataset = split_dataset['train'], split_dataset['test']
-    collate_fn_with_tokenizer = partial(collate_fn, tokenizer=tokenizer)
-    # model = TexTeller()
-    model = TexTeller.from_pretrained('/home/lhy/code/TeXify/src/models/ocr_model/train/train_result/checkpoint-81000')
-
-    enable_train = False
-    enable_evaluate = True
-    if enable_train:
-        train(model, tokenizer, train_dataset, eval_dataset, collate_fn_with_tokenizer)  
-    if enable_evaluate:
-        evaluate(model, tokenizer, eval_dataset, collate_fn_with_tokenizer)
-
-
-    os.chdir(cur_path)
-
-
-'''

src/models/ocr_model/train/training_args.py

@@ -4,13 +4,13 @@ CONFIG = {
     # "data_seed": 42,                     # data sampler的采样也固定
     # "full_determinism": True,            # 使整个训练完全固定（这个设置会有害于模型训练，只用于debug）
 
-    "output_dir": "train_result",          # 输出目录
+    "output_dir": "train_result/train_with_random_resize",          # 输出目录
     "overwrite_output_dir": False,         # 如果输出目录存在，不删除原先的内容
     "report_to": ["tensorboard"],          # 输出日志到TensorBoard，
                                            #+通过在命令行：tensorboard --logdir ./logs 来查看日志
 
     "logging_dir": None,                   # TensorBoard日志文件的存储目录(使用默认值)
-    "log_level": "info",                   # 其他可选:‘debug’, ‘info’, ‘warning’, ‘error’ and ‘critical’（由低级别到高级别）
+    "log_level": "warning",                   # 其他可选:‘debug’, ‘info’, ‘warning’, ‘error’ and ‘critical’（由低级别到高级别）
     "logging_strategy": "steps",           # 每隔一定步数记录一次日志
     "logging_steps": 500,                  # 记录日志的步数间隔，可以是int也可以是(0~1)的float，当是float时表示总的训练步数的ratio(比方说可以设置成1.0 / 2000)
                                            #+通常与eval_steps一致
@@ -22,7 +22,7 @@ CONFIG = {
 
     # "label_names": ['your_label_name'],  # 指定data_loader中的标签名，如果不指定则默认为'labels'
 
-    "per_device_train_batch_size": 128,    # 每个GPU的batch size
+    "per_device_train_batch_size": 64,    # 每个GPU的batch size
     "per_device_eval_batch_size": 16,      # 每个GPU的evaluation batch size
     "auto_find_batch_size": True,          # 自动搜索合适的batch size（指数decay）
 

src/models/ocr_model/utils/functional.py

@@ -38,7 +38,6 @@ def collate_fn(samples: List[Dict[str, Any]], tokenizer=None) -> Dict[str, List[
 
     # 左移labels和decoder_attention_mask
     batch['labels'] = left_move(batch['labels'], -100)
-    # batch['decoder_attention_mask'] = left_move(batch['decoder_attention_mask'], 0)
 
     # 把list of Image转成一个tensor with (B, C, H, W)
     batch['pixel_values'] = torch.stack(batch['pixel_values'], dim=0)
@@ -76,48 +75,3 @@ if __name__ == '__main__':
     out = model(**batch)
 
     pause = 1
-
-
-'''
-def left_move(x: torch.Tensor, pad_val):
-    assert len(x.shape) == 2, 'x should be 2-dimensional'
-    lefted_x = torch.ones_like(x)
-    lefted_x[:, :-1] = x[:, 1:]
-    lefted_x[:, -1] = pad_val
-    return lefted_x
-
-
-def tokenize_fn(samples: Dict[str, List[Any]], tokenizer=None) -> Dict[str, List[Any]]:
-    assert tokenizer is not None, 'tokenizer should not be None'
-    tokenized_formula = tokenizer(samples['latex_formula'], return_special_tokens_mask=True)
-    tokenized_formula['pixel_values'] = samples['image']
-    return tokenized_formula
-
-
-def collate_fn(samples: List[Dict[str, Any]], tokenizer=None) -> Dict[str, List[Any]]:
-    assert tokenizer is not None, 'tokenizer should not be None'
-    pixel_values = [dic.pop('pixel_values') for dic in samples]
-
-    clm_collator = DataCollatorForLanguageModeling(tokenizer=tokenizer, mlm=False)
-    
-    batch = clm_collator(samples)
-    batch['pixel_values'] = pixel_values
-    batch['decoder_input_ids'] = batch.pop('input_ids')
-    batch['decoder_attention_mask'] = batch.pop('attention_mask')
-
-    # 左移labels和decoder_attention_mask
-    batch['labels'] = left_move(batch['labels'], -100)
-    batch['decoder_attention_mask'] = left_move(batch['decoder_attention_mask'], 0)
-
-    # 把list of Image转成一个tensor with (B, C, H, W)
-    batch['pixel_values'] = torch.stack(batch['pixel_values'], dim=0)
-    return batch
-
-
-def img_preprocess(samples: Dict[str, List[Any]]) -> Dict[str, List[Any]]:
-    processed_img = train_transform(samples['pixel_values'])
-    samples['pixel_values'] = processed_img
-    return samples
-
-'''
-

src/models/ocr_model/utils/transforms.py

@@ -4,7 +4,6 @@ import numpy as np
 import cv2
 
 from torchvision.transforms import v2
-from PIL import ImageChops, Image
 from typing import List, Union
 
 from ....globals import (
@@ -107,7 +106,7 @@ def random_resize(
     ]
 
 
-def general_transform(images: List[Image.Image]) -> List[torch.Tensor]:
+def general_transform(images: List[np.ndarray]) -> List[torch.Tensor]:
     # 裁剪掉白边
     images = [trim_white_border(image) for image in images]
     # general transform pipeline
@@ -117,16 +116,16 @@ def general_transform(images: List[Image.Image]) -> List[torch.Tensor]:
     return images
 
 
-def train_transform(images: List[np.ndarray]) -> List[torch.Tensor]:
+def train_transform(images: List[List[List[List]]]) -> List[torch.Tensor]:
     assert OCR_IMG_CHANNELS == 1 , "Only support grayscale images for now"
     assert OCR_FIX_SIZE == True, "Only support fixed size images for now"
 
     # random resize first
-    # images = random_resize(images, MIN_RESIZE_RATIO, MAX_RESIZE_RATIO)
+    images = random_resize(images, MIN_RESIZE_RATIO, MAX_RESIZE_RATIO)
     return general_transform(images)
 
 
-def inference_transform(images: List[Image.Image]) -> List[torch.Tensor]:
+def inference_transform(images: List[np.ndarray]) -> List[torch.Tensor]:
     assert OCR_IMG_CHANNELS == 1 , "Only support grayscale images for now"
     assert OCR_FIX_SIZE == True, "Only support fixed size images for now"