文字识别训练在初始化dataset的索引时时报错list index out of range,然后会卡在During the training process, after the 0th iteration, an evaluation is run every 2000 iterations #13250
-
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem
(PaddleOcr) ~\Desktop\PaddleOCR git:[main]
python tools/train.py -c configs/rec/PP-OCRv3/ch_PP-OCRv3_rec_distillation.yml
[2024/05/11 09:03:45] ppocr INFO: Architecture :
[2024/05/11 09:03:45] ppocr INFO: Models :
[2024/05/11 09:03:45] ppocr INFO: Student :
[2024/05/11 09:03:45] ppocr INFO: Backbone :
[2024/05/11 09:03:45] ppocr INFO: last_conv_stride : [1, 2]
[2024/05/11 09:03:45] ppocr INFO: last_pool_kernel_size : [2, 2]
[2024/05/11 09:03:45] ppocr INFO: last_pool_type : avg
[2024/05/11 09:03:45] ppocr INFO: name : MobileNetV1Enhance
[2024/05/11 09:03:45] ppocr INFO: scale : 0.5
[2024/05/11 09:03:45] ppocr INFO: Head :
[2024/05/11 09:03:45] ppocr INFO: head_list :
[2024/05/11 09:03:45] ppocr INFO: CTCHead :
[2024/05/11 09:03:45] ppocr INFO: Head :
[2024/05/11 09:03:45] ppocr INFO: fc_decay : 1e-05
[2024/05/11 09:03:45] ppocr INFO: Neck :
[2024/05/11 09:03:45] ppocr INFO: depth : 2
[2024/05/11 09:03:45] ppocr INFO: dims : 64
[2024/05/11 09:03:45] ppocr INFO: hidden_dims : 120
[2024/05/11 09:03:45] ppocr INFO: name : svtr
[2024/05/11 09:03:45] ppocr INFO: use_guide : True
[2024/05/11 09:03:45] ppocr INFO: SARHead :
[2024/05/11 09:03:45] ppocr INFO: enc_dim : 512
[2024/05/11 09:03:45] ppocr INFO: max_text_length : 25
[2024/05/11 09:03:45] ppocr INFO: name : MultiHead
[2024/05/11 09:03:45] ppocr INFO: Transform : None
[2024/05/11 09:03:45] ppocr INFO: algorithm : SVTR_LCNet
[2024/05/11 09:03:45] ppocr INFO: freeze_params : False
[2024/05/11 09:03:45] ppocr INFO: model_type : rec
[2024/05/11 09:03:45] ppocr INFO: pretrained : None
[2024/05/11 09:03:45] ppocr INFO: return_all_feats : True
[2024/05/11 09:03:45] ppocr INFO: Teacher :
[2024/05/11 09:03:45] ppocr INFO: Backbone :
[2024/05/11 09:03:45] ppocr INFO: last_conv_stride : [1, 2]
[2024/05/11 09:03:45] ppocr INFO: last_pool_kernel_size : [2, 2]
[2024/05/11 09:03:45] ppocr INFO: last_pool_type : avg
[2024/05/11 09:03:45] ppocr INFO: name : MobileNetV1Enhance
[2024/05/11 09:03:45] ppocr INFO: scale : 0.5
[2024/05/11 09:03:45] ppocr INFO: Head :
[2024/05/11 09:03:45] ppocr INFO: head_list :
[2024/05/11 09:03:45] ppocr INFO: CTCHead :
[2024/05/11 09:03:45] ppocr INFO: Head :
[2024/05/11 09:03:45] ppocr INFO: fc_decay : 1e-05
[2024/05/11 09:03:45] ppocr INFO: Neck :
[2024/05/11 09:03:45] ppocr INFO: depth : 2
[2024/05/11 09:03:45] ppocr INFO: dims : 64
[2024/05/11 09:03:45] ppocr INFO: hidden_dims : 120
[2024/05/11 09:03:45] ppocr INFO: name : svtr
[2024/05/11 09:03:45] ppocr INFO: use_guide : True
[2024/05/11 09:03:45] ppocr INFO: SARHead :
[2024/05/11 09:03:45] ppocr INFO: enc_dim : 512
[2024/05/11 09:03:45] ppocr INFO: max_text_length : 25
[2024/05/11 09:03:45] ppocr INFO: name : MultiHead
[2024/05/11 09:03:45] ppocr INFO: Transform : None
[2024/05/11 09:03:45] ppocr INFO: algorithm : SVTR_LCNet
[2024/05/11 09:03:45] ppocr INFO: freeze_params : False
[2024/05/11 09:03:45] ppocr INFO: model_type : rec
[2024/05/11 09:03:45] ppocr INFO: pretrained : None
[2024/05/11 09:03:45] ppocr INFO: return_all_feats : True
[2024/05/11 09:03:45] ppocr INFO: algorithm : Distillation
[2024/05/11 09:03:45] ppocr INFO: model_type : rec
[2024/05/11 09:03:45] ppocr INFO: name : DistillationModel
[2024/05/11 09:03:45] ppocr INFO: Eval :
[2024/05/11 09:03:45] ppocr INFO: dataset :
[2024/05/11 09:03:45] ppocr INFO: data_dir : ./train_data/rec/
[2024/05/11 09:03:45] ppocr INFO: label_file_list : ['./train_data/rec/label.txt']
[2024/05/11 09:03:45] ppocr INFO: name : SimpleDataSet
[2024/05/11 09:03:45] ppocr INFO: transforms :
[2024/05/11 09:03:45] ppocr INFO: DecodeImage :
[2024/05/11 09:03:45] ppocr INFO: channel_first : False
[2024/05/11 09:03:45] ppocr INFO: img_mode : BGR
[2024/05/11 09:03:45] ppocr INFO: MultiLabelEncode : None
[2024/05/11 09:03:45] ppocr INFO: RecResizeImg :
[2024/05/11 09:03:45] ppocr INFO: image_shape : [3, 48, 320]
[2024/05/11 09:03:45] ppocr INFO: KeepKeys :
[2024/05/11 09:03:45] ppocr INFO: keep_keys : ['image', 'label_ctc', 'label_sar', 'length', 'valid_ratio']
[2024/05/11 09:03:45] ppocr INFO: loader :
[2024/05/11 09:03:45] ppocr INFO: batch_size_per_card : 128
[2024/05/11 09:03:45] ppocr INFO: drop_last : False
[2024/05/11 09:03:45] ppocr INFO: num_workers : 4
[2024/05/11 09:03:45] ppocr INFO: shuffle : False
[2024/05/11 09:03:45] ppocr INFO: Global :
[2024/05/11 09:03:45] ppocr INFO: cal_metric_during_train : True
[2024/05/11 09:03:45] ppocr INFO: character_dict_path : ppocr/utils/ppocr_keys_v1.txt
[2024/05/11 09:03:45] ppocr INFO: checkpoints : None
[2024/05/11 09:03:45] ppocr INFO: d2s_train_image_shape : [3, 48, -1]
[2024/05/11 09:03:45] ppocr INFO: debug : False
[2024/05/11 09:03:45] ppocr INFO: distributed : False
[2024/05/11 09:03:45] ppocr INFO: epoch_num : 800
[2024/05/11 09:03:45] ppocr INFO: eval_batch_step : [0, 2000]
[2024/05/11 09:03:45] ppocr INFO: infer_img : doc/imgs_words/ch/word_1.jpg
[2024/05/11 09:03:45] ppocr INFO: infer_mode : False
[2024/05/11 09:03:45] ppocr INFO: log_smooth_window : 20
[2024/05/11 09:03:45] ppocr INFO: max_text_length : 25
[2024/05/11 09:03:45] ppocr INFO: pretrained_model : ./pretrain_models/ch_PP-OCRv3_rec_train/best_accuracy
[2024/05/11 09:03:45] ppocr INFO: print_batch_step : 10
[2024/05/11 09:03:45] ppocr INFO: save_epoch_step : 3
[2024/05/11 09:03:45] ppocr INFO: save_inference_dir : None
[2024/05/11 09:03:45] ppocr INFO: save_model_dir : ./output/rec_ppocr_v3_distillation
[2024/05/11 09:03:45] ppocr INFO: save_res_path : ./output/rec/predicts_ppocrv3_distillation.txt
[2024/05/11 09:03:45] ppocr INFO: use_gpu : False
[2024/05/11 09:03:45] ppocr INFO: use_space_char : True
[2024/05/11 09:03:45] ppocr INFO: use_visualdl : False
[2024/05/11 09:03:45] ppocr INFO: Loss :
[2024/05/11 09:03:45] ppocr INFO: loss_config_list :
[2024/05/11 09:03:45] ppocr INFO: DistillationDMLLoss :
[2024/05/11 09:03:45] ppocr INFO: act : softmax
[2024/05/11 09:03:45] ppocr INFO: dis_head : ctc
[2024/05/11 09:03:45] ppocr INFO: key : head_out
[2024/05/11 09:03:45] ppocr INFO: model_name_pairs : [['Student', 'Teacher']]
[2024/05/11 09:03:45] ppocr INFO: multi_head : True
[2024/05/11 09:03:45] ppocr INFO: name : dml_ctc
[2024/05/11 09:03:45] ppocr INFO: use_log : True
[2024/05/11 09:03:45] ppocr INFO: weight : 1.0
[2024/05/11 09:03:45] ppocr INFO: DistillationDMLLoss :
[2024/05/11 09:03:45] ppocr INFO: act : softmax
[2024/05/11 09:03:45] ppocr INFO: dis_head : sar
[2024/05/11 09:03:45] ppocr INFO: key : head_out
[2024/05/11 09:03:45] ppocr INFO: model_name_pairs : [['Student', 'Teacher']]
[2024/05/11 09:03:45] ppocr INFO: multi_head : True
[2024/05/11 09:03:45] ppocr INFO: name : dml_sar
[2024/05/11 09:03:45] ppocr INFO: use_log : True
[2024/05/11 09:03:45] ppocr INFO: weight : 0.5
[2024/05/11 09:03:45] ppocr INFO: DistillationDistanceLoss :
[2024/05/11 09:03:45] ppocr INFO: key : backbone_out
[2024/05/11 09:03:45] ppocr INFO: mode : l2
[2024/05/11 09:03:45] ppocr INFO: model_name_pairs : [['Student', 'Teacher']]
[2024/05/11 09:03:45] ppocr INFO: weight : 1.0
[2024/05/11 09:03:45] ppocr INFO: DistillationCTCLoss :
[2024/05/11 09:03:45] ppocr INFO: key : head_out
[2024/05/11 09:03:45] ppocr INFO: model_name_list : ['Student', 'Teacher']
[2024/05/11 09:03:45] ppocr INFO: multi_head : True
[2024/05/11 09:03:45] ppocr INFO: weight : 1.0
[2024/05/11 09:03:45] ppocr INFO: DistillationSARLoss :
[2024/05/11 09:03:45] ppocr INFO: key : head_out
[2024/05/11 09:03:45] ppocr INFO: model_name_list : ['Student', 'Teacher']
[2024/05/11 09:03:45] ppocr INFO: multi_head : True
[2024/05/11 09:03:45] ppocr INFO: weight : 1.0
[2024/05/11 09:03:45] ppocr INFO: name : CombinedLoss
[2024/05/11 09:03:45] ppocr INFO: Metric :
[2024/05/11 09:03:45] ppocr INFO: base_metric_name : RecMetric
[2024/05/11 09:03:45] ppocr INFO: ignore_space : False
[2024/05/11 09:03:45] ppocr INFO: key : Student
[2024/05/11 09:03:45] ppocr INFO: main_indicator : acc
[2024/05/11 09:03:45] ppocr INFO: name : DistillationMetric
[2024/05/11 09:03:45] ppocr INFO: Optimizer :
[2024/05/11 09:03:45] ppocr INFO: beta1 : 0.9
[2024/05/11 09:03:45] ppocr INFO: beta2 : 0.999
[2024/05/11 09:03:45] ppocr INFO: lr :
[2024/05/11 09:03:45] ppocr INFO: decay_epochs : [700]
[2024/05/11 09:03:45] ppocr INFO: name : Piecewise
[2024/05/11 09:03:45] ppocr INFO: values : [0.0005, 5e-05]
[2024/05/11 09:03:45] ppocr INFO: warmup_epoch : 5
[2024/05/11 09:03:45] ppocr INFO: name : Adam
[2024/05/11 09:03:45] ppocr INFO: regularizer :
[2024/05/11 09:03:45] ppocr INFO: factor : 3e-05
[2024/05/11 09:03:45] ppocr INFO: name : L2
[2024/05/11 09:03:45] ppocr INFO: PostProcess :
[2024/05/11 09:03:45] ppocr INFO: key : head_out
[2024/05/11 09:03:45] ppocr INFO: model_name : ['Student', 'Teacher']
[2024/05/11 09:03:45] ppocr INFO: multi_head : True
[2024/05/11 09:03:45] ppocr INFO: name : DistillationCTCLabelDecode
[2024/05/11 09:03:45] ppocr INFO: Train :
[2024/05/11 09:03:45] ppocr INFO: dataset :
[2024/05/11 09:03:45] ppocr INFO: data_dir : ./train_data/rec/
[2024/05/11 09:03:45] ppocr INFO: ext_op_transform_idx : 1
[2024/05/11 09:03:45] ppocr INFO: label_file_list : ['./train_data/rec/label.txt']
[2024/05/11 09:03:45] ppocr INFO: name : SimpleDataSet
[2024/05/11 09:03:45] ppocr INFO: transforms :
[2024/05/11 09:03:45] ppocr INFO: DecodeImage :
[2024/05/11 09:03:45] ppocr INFO: channel_first : False
[2024/05/11 09:03:45] ppocr INFO: img_mode : BGR
[2024/05/11 09:03:45] ppocr INFO: RecConAug :
[2024/05/11 09:03:45] ppocr INFO: ext_data_num : 2
[2024/05/11 09:03:45] ppocr INFO: image_shape : [48, 320, 3]
[2024/05/11 09:03:45] ppocr INFO: max_text_length : 25
[2024/05/11 09:03:45] ppocr INFO: prob : 0.5
[2024/05/11 09:03:45] ppocr INFO: RecAug : None
[2024/05/11 09:03:45] ppocr INFO: MultiLabelEncode : None
[2024/05/11 09:03:45] ppocr INFO: RecResizeImg :
[2024/05/11 09:03:45] ppocr INFO: image_shape : [3, 48, 320]
[2024/05/11 09:03:45] ppocr INFO: KeepKeys :
[2024/05/11 09:03:45] ppocr INFO: keep_keys : ['image', 'label_ctc', 'label_sar', 'length', 'valid_ratio']
[2024/05/11 09:03:45] ppocr INFO: loader :
[2024/05/11 09:03:45] ppocr INFO: batch_size_per_card : 128
[2024/05/11 09:03:45] ppocr INFO: drop_last : True
[2024/05/11 09:03:45] ppocr INFO: num_workers : 4
[2024/05/11 09:03:45] ppocr INFO: shuffle : True
[2024/05/11 09:03:45] ppocr INFO: profiler_options : None
[2024/05/11 09:03:45] ppocr INFO: train with paddle 2.6.1 and device Place(cpu)
[2024/05/11 09:03:45] ppocr INFO: Initialize indexs of datasets:['./train_data/rec/label.txt']
list index out of range
[2024/05/11 09:03:45] ppocr INFO: Initialize indexs of datasets:['./train_data/rec/label.txt']
[2024/05/11 09:03:47] ppocr INFO: train dataloader has 782 iters
[2024/05/11 09:03:47] ppocr INFO: valid dataloader has 783 iters
[2024/05/11 09:03:47] ppocr INFO: load pretrain successful from ./pretrain_models/ch_PP-OCRv3_rec_train/best_accuracy
[2024/05/11 09:03:47] ppocr INFO: During the training process, after the 0th iteration, an evaluation is run every 2000 iterations |
Beta Was this translation helpful? Give feedback.
Replies: 7 comments 5 replies
-
将use_gpu设置为True |
Beta Was this translation helpful? Give feedback.
-
谢谢 |
Beta Was this translation helpful? Give feedback.
-
@zhangyubo0722 您好,我的use_gpu=True,但是还是卡在这里就结束了。 |
Beta Was this translation helpful? Give feedback.
-
@555myhair 更新一下paddle版本试试,用paddle 3.0 beta 版本 |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
数据里是不是有不在char_dict.txt中的字符?@555myhair |
Beta Was this translation helpful? Give feedback.
-
我的数据里面都是数字和字母,里面都有的 |
Beta Was this translation helpful? Give feedback.
检查一下数据标注的格式对不对