During train Text Direction Classification, training is stop at loading eval model. #12966

diagram298 · 2024-02-21T04:44:32Z

diagram298
Feb 21, 2024

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

系统环境/System Environment：linux, CUDA12.1, paddlepaddle-gpu 2.6.0
版本号/Version：Paddle： PaddleOCR：问题相关组件/Related components：
What can I do for this problem? My data size -> trainimg: 270000, evalimg: 32 <- Is it too unbalanced?
cls.yml

Global:
  use_gpu: true
  epoch_num: 100
  log_smooth_window: 20
  print_batch_step: 100
  save_model_dir: ./output/cls/mv3/
  save_epoch_step: 1
  # evaluation is run every 5000 iterations after the 4000th iteration
  eval_batch_step:  [0, 100]
  cal_metric_during_train: True
  pretrained_model: best_accuracy
  checkpoints:
  save_inference_dir:
  use_visualdl: False
  infer_img: doc/imgs_words_en/word_10.png
  label_list: ['0','180']

Architecture:
  model_type: cls
  algorithm: CLS
  Transform:
  Backbone:
    name: MobileNetV3
    scale: 0.35
    model_name: small
  Neck:
  Head:
    name: ClsHead
    class_dim: 2

Loss:
  name: ClsLoss

Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
    name: Cosine
    learning_rate: 0.001
  regularizer:
    name: 'L2'
    factor: 0

PostProcess:
  name: ClsPostProcess

Metric:
  name: ClsMetric
  main_indicator: acc

Train:
  dataset:
    name: SimpleDataSet
    data_dir: ./train_data/cls
    label_file_list:
      - ./train_data/cls/ttrain.txt
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - ClsLabelEncode: # Class handling label
      - BaseDataAugmentation:
      - RandAugment:
      - ClsResizeImg:
          image_shape: [3, 48, 192]
      - KeepKeys:
          keep_keys: ['image', 'label'] # dataloader will return list in this order
  loader:
    shuffle: True
    batch_size_per_card: 32 # [*] 512
    drop_last: True
    num_workers: 0

Eval:
  dataset:
    name: SimpleDataSet
    data_dir: ./train_data/cls
    label_file_list:
      - ./train_data/cls/ttest.txt
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - ClsLabelEncode: # Class handling label
      - ClsResizeImg:
          image_shape: [3, 48, 192]
      - KeepKeys:
          keep_keys: ['image', 'label'] # dataloader will return list in this order
  loader:
    shuffle: False
    drop_last: False
    batch_size_per_card: 32 # [*] 512
    num_workers: 4

command log : python tools/train.py -c configs/cls/cls_mv3.yml

[2024/02/21 13:23:24] ppocr INFO: Architecture : 
[2024/02/21 13:23:24] ppocr INFO:     Backbone : 
[2024/02/21 13:23:24] ppocr INFO:         model_name : small
[2024/02/21 13:23:24] ppocr INFO:         name : MobileNetV3
[2024/02/21 13:23:24] ppocr INFO:         scale : 0.35
[2024/02/21 13:23:24] ppocr INFO:     Head : 
[2024/02/21 13:23:24] ppocr INFO:         class_dim : 2
[2024/02/21 13:23:24] ppocr INFO:         name : ClsHead
[2024/02/21 13:23:24] ppocr INFO:     Neck : None
[2024/02/21 13:23:24] ppocr INFO:     Transform : None
[2024/02/21 13:23:24] ppocr INFO:     algorithm : CLS
[2024/02/21 13:23:24] ppocr INFO:     model_type : cls
[2024/02/21 13:23:24] ppocr INFO: Eval : 
[2024/02/21 13:23:24] ppocr INFO:     dataset : 
[2024/02/21 13:23:24] ppocr INFO:         data_dir : ./train_data/cls
[2024/02/21 13:23:24] ppocr INFO:         label_file_list : ['./train_data/cls/ttest.txt']
[2024/02/21 13:23:24] ppocr INFO:         name : SimpleDataSet
[2024/02/21 13:23:24] ppocr INFO:         transforms : 
[2024/02/21 13:23:24] ppocr INFO:             DecodeImage : 
[2024/02/21 13:23:24] ppocr INFO:                 channel_first : False
[2024/02/21 13:23:24] ppocr INFO:                 img_mode : BGR
[2024/02/21 13:23:24] ppocr INFO:             ClsLabelEncode : None
[2024/02/21 13:23:24] ppocr INFO:             ClsResizeImg : 
[2024/02/21 13:23:24] ppocr INFO:                 image_shape : [3, 48, 192]
[2024/02/21 13:23:24] ppocr INFO:             KeepKeys : 
[2024/02/21 13:23:24] ppocr INFO:                 keep_keys : ['image', 'label']
[2024/02/21 13:23:24] ppocr INFO:     loader : 
[2024/02/21 13:23:24] ppocr INFO:         batch_size_per_card : 32
[2024/02/21 13:23:24] ppocr INFO:         drop_last : False
[2024/02/21 13:23:24] ppocr INFO:         num_workers : 4
[2024/02/21 13:23:24] ppocr INFO:         shuffle : False
[2024/02/21 13:23:24] ppocr INFO: Global : 
[2024/02/21 13:23:24] ppocr INFO:     cal_metric_during_train : True
[2024/02/21 13:23:24] ppocr INFO:     checkpoints : None
[2024/02/21 13:23:24] ppocr INFO:     distributed : False
[2024/02/21 13:23:24] ppocr INFO:     epoch_num : 100
[2024/02/21 13:23:24] ppocr INFO:     eval_batch_step : [0, 100]
[2024/02/21 13:23:24] ppocr INFO:     infer_img : doc/imgs_words_en/word_10.png
[2024/02/21 13:23:24] ppocr INFO:     label_list : ['0', '180']
[2024/02/21 13:23:24] ppocr INFO:     log_smooth_window : 20
[2024/02/21 13:23:24] ppocr INFO:     pretrained_model : best_accuracy
[2024/02/21 13:23:24] ppocr INFO:     print_batch_step : 100
[2024/02/21 13:23:24] ppocr INFO:     save_epoch_step : 1
[2024/02/21 13:23:24] ppocr INFO:     save_inference_dir : None
[2024/02/21 13:23:24] ppocr INFO:     save_model_dir : ./output/cls/mv3/
[2024/02/21 13:23:24] ppocr INFO:     use_gpu : True
[2024/02/21 13:23:24] ppocr INFO:     use_visualdl : False
[2024/02/21 13:23:24] ppocr INFO: Loss : 
[2024/02/21 13:23:24] ppocr INFO:     name : ClsLoss
[2024/02/21 13:23:24] ppocr INFO: Metric : 
[2024/02/21 13:23:24] ppocr INFO:     main_indicator : acc
[2024/02/21 13:23:24] ppocr INFO:     name : ClsMetric
[2024/02/21 13:23:24] ppocr INFO: Optimizer : 
[2024/02/21 13:23:24] ppocr INFO:     beta1 : 0.9
[2024/02/21 13:23:24] ppocr INFO:     beta2 : 0.999
[2024/02/21 13:23:24] ppocr INFO:     lr : 
[2024/02/21 13:23:24] ppocr INFO:         learning_rate : 0.001
[2024/02/21 13:23:24] ppocr INFO:         name : Cosine
[2024/02/21 13:23:24] ppocr INFO:     name : Adam
[2024/02/21 13:23:24] ppocr INFO:     regularizer : 
[2024/02/21 13:23:24] ppocr INFO:         factor : 0
[2024/02/21 13:23:24] ppocr INFO:         name : L2
[2024/02/21 13:23:24] ppocr INFO: PostProcess : 
[2024/02/21 13:23:24] ppocr INFO:     name : ClsPostProcess
[2024/02/21 13:23:24] ppocr INFO: Train : 
[2024/02/21 13:23:24] ppocr INFO:     dataset : 
[2024/02/21 13:23:24] ppocr INFO:         data_dir : ./train_data/cls
[2024/02/21 13:23:24] ppocr INFO:         label_file_list : ['./train_data/cls/ttrain.txt']
[2024/02/21 13:23:24] ppocr INFO:         name : SimpleDataSet
[2024/02/21 13:23:24] ppocr INFO:         transforms : 
[2024/02/21 13:23:24] ppocr INFO:             DecodeImage : 
[2024/02/21 13:23:24] ppocr INFO:                 channel_first : False
[2024/02/21 13:23:24] ppocr INFO:                 img_mode : BGR
[2024/02/21 13:23:24] ppocr INFO:             ClsLabelEncode : None
[2024/02/21 13:23:24] ppocr INFO:             BaseDataAugmentation : None
[2024/02/21 13:23:24] ppocr INFO:             RandAugment : None
[2024/02/21 13:23:24] ppocr INFO:             ClsResizeImg : 
[2024/02/21 13:23:24] ppocr INFO:                 image_shape : [3, 48, 192]
[2024/02/21 13:23:24] ppocr INFO:             KeepKeys : 
[2024/02/21 13:23:24] ppocr INFO:                 keep_keys : ['image', 'label']
[2024/02/21 13:23:24] ppocr INFO:     loader : 
[2024/02/21 13:23:24] ppocr INFO:         batch_size_per_card : 32
[2024/02/21 13:23:24] ppocr INFO:         drop_last : True
[2024/02/21 13:23:24] ppocr INFO:         num_workers : 0
[2024/02/21 13:23:24] ppocr INFO:         shuffle : True
[2024/02/21 13:23:24] ppocr INFO: profiler_options : None
[2024/02/21 13:23:24] ppocr INFO: train with paddle 2.6.0 and device Place(gpu:0)
[2024/02/21 13:23:24] ppocr INFO: Initialize indexs of datasets:['./train_data/cls/ttrain.txt']
list index out of range
[2024/02/21 13:23:24] ppocr INFO: Initialize indexs of datasets:['./train_data/cls/ttest.txt']
W0221 13:23:24.456318 2191625 gpu_resources.cc:119] Please NOTE: device: 0, GPU Compute Capability: 8.6, Driver API Version: 12.2, Runtime API Version: 11.8
W0221 13:23:24.457041 2191625 gpu_resources.cc:164] device: 0, cuDNN Version: 8.9.
[2024/02/21 13:23:24] ppocr INFO: train dataloader has 8465 iters
[2024/02/21 13:23:24] ppocr INFO: valid dataloader has 1 iters
[2024/02/21 13:23:24] ppocr INFO: load pretrain successful from best_accuracy
[2024/02/21 13:23:24] ppocr INFO: During the training process, after the 0th iteration, an evaluation is run every 100 iterations
[2024/02/21 13:23:29] ppocr INFO: epoch: [1/100], global_step: 100, lr: 0.001000, acc: 1.000000, loss: 0.037324, avg_reader_cost: 0.02331 s, avg_batch_cost: 0.04661 s, avg_samples: 32.0, ips: 686.59691 samples/s, eta: 10:57:27
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.67it/s]
[2024/02/21 13:23:29] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2420.168920624617
[2024/02/21 13:23:29] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:29] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2420.168920624617, best_epoch: 1
[2024/02/21 13:23:33] ppocr INFO: epoch: [1/100], global_step: 200, lr: 0.001000, acc: 1.000000, loss: 0.022498, avg_reader_cost: 0.02331 s, avg_batch_cost: 0.04282 s, avg_samples: 32.0, ips: 747.25229 samples/s, eta: 10:30:42
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 14.33it/s]
[2024/02/21 13:23:33] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2232.1258606352903
[2024/02/21 13:23:33] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:33] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2232.1258606352903, best_epoch: 1
[2024/02/21 13:23:38] ppocr INFO: epoch: [1/100], global_step: 300, lr: 0.001000, acc: 0.984375, loss: 0.035846, avg_reader_cost: 0.02390 s, avg_batch_cost: 0.04418 s, avg_samples: 32.0, ips: 724.36124 samples/s, eta: 10:28:06
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 14.49it/s]
[2024/02/21 13:23:38] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3174.6470504754247
[2024/02/21 13:23:38] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:38] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3174.6470504754247, best_epoch: 1
[2024/02/21 13:23:42] ppocr INFO: epoch: [1/100], global_step: 400, lr: 0.001000, acc: 0.968750, loss: 0.044509, avg_reader_cost: 0.02405 s, avg_batch_cost: 0.04328 s, avg_samples: 32.0, ips: 739.34395 samples/s, eta: 10:23:36
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 15.34it/s]
[2024/02/21 13:23:42] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3474.0831392038103
[2024/02/21 13:23:42] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:42] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3474.0831392038103, best_epoch: 1
[2024/02/21 13:23:47] ppocr INFO: epoch: [1/100], global_step: 500, lr: 0.001000, acc: 1.000000, loss: 0.026678, avg_reader_cost: 0.02545 s, avg_batch_cost: 0.04400 s, avg_samples: 32.0, ips: 727.19710 samples/s, eta: 10:22:55
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 15.66it/s]
[2024/02/21 13:23:47] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3407.3196415424845
[2024/02/21 13:23:47] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:47] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3407.3196415424845, best_epoch: 1
[2024/02/21 13:23:51] ppocr INFO: epoch: [1/100], global_step: 600, lr: 0.001000, acc: 1.000000, loss: 0.017840, avg_reader_cost: 0.02247 s, avg_batch_cost: 0.04140 s, avg_samples: 32.0, ips: 773.00422 samples/s, eta: 10:16:18
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 12.58it/s]
[2024/02/21 13:23:51] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2273.0659982725624
[2024/02/21 13:23:51] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:51] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2273.0659982725624, best_epoch: 1
[2024/02/21 13:23:55] ppocr INFO: epoch: [1/100], global_step: 700, lr: 0.001000, acc: 1.000000, loss: 0.020293, avg_reader_cost: 0.02373 s, avg_batch_cost: 0.04195 s, avg_samples: 32.0, ips: 762.83578 samples/s, eta: 10:12:40
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 14.62it/s]
[2024/02/21 13:23:55] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2602.0264433329457
[2024/02/21 13:23:55] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:23:55] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2602.0264433329457, best_epoch: 1
[2024/02/21 13:24:00] ppocr INFO: epoch: [1/100], global_step: 800, lr: 0.001000, acc: 1.000000, loss: 0.012668, avg_reader_cost: 0.02599 s, avg_batch_cost: 0.04480 s, avg_samples: 32.0, ips: 714.32114 samples/s, eta: 10:14:57
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 12.81it/s]
[2024/02/21 13:24:00] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2715.803565286012
[2024/02/21 13:24:00] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:00] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2715.803565286012, best_epoch: 1
[2024/02/21 13:24:04] ppocr INFO: epoch: [1/100], global_step: 900, lr: 0.001000, acc: 1.000000, loss: 0.020033, avg_reader_cost: 0.02294 s, avg_batch_cost: 0.04211 s, avg_samples: 32.0, ips: 759.93685 samples/s, eta: 10:12:30
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 11.35it/s]
[2024/02/21 13:24:04] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 990.6025344856854
[2024/02/21 13:24:04] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:04] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 990.6025344856854, best_epoch: 1
[2024/02/21 13:24:09] ppocr INFO: epoch: [1/100], global_step: 1000, lr: 0.001000, acc: 1.000000, loss: 0.018790, avg_reader_cost: 0.02161 s, avg_batch_cost: 0.04077 s, avg_samples: 32.0, ips: 784.90606 samples/s, eta: 10:08:38
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 14.47it/s]
[2024/02/21 13:24:09] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3210.0288912273986
[2024/02/21 13:24:09] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:09] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3210.0288912273986, best_epoch: 1
[2024/02/21 13:24:13] ppocr INFO: epoch: [1/100], global_step: 1100, lr: 0.001000, acc: 1.000000, loss: 0.012634, avg_reader_cost: 0.02281 s, avg_batch_cost: 0.04186 s, avg_samples: 32.0, ips: 764.51392 samples/s, eta: 10:06:51
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.67it/s]
[2024/02/21 13:24:13] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2846.6114103923646
[2024/02/21 13:24:13] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:13] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2846.6114103923646, best_epoch: 1
[2024/02/21 13:24:17] ppocr INFO: epoch: [1/100], global_step: 1200, lr: 0.001000, acc: 1.000000, loss: 0.023840, avg_reader_cost: 0.02255 s, avg_batch_cost: 0.03922 s, avg_samples: 32.0, ips: 815.87545 samples/s, eta: 10:02:16
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 15.07it/s]
[2024/02/21 13:24:17] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2858.79844086136
[2024/02/21 13:24:17] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:17] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2858.79844086136, best_epoch: 1
[2024/02/21 13:24:21] ppocr INFO: epoch: [1/100], global_step: 1300, lr: 0.001000, acc: 0.984375, loss: 0.030372, avg_reader_cost: 0.02430 s, avg_batch_cost: 0.04247 s, avg_samples: 32.0, ips: 753.41692 samples/s, eta: 10:01:53
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 15.38it/s]
[2024/02/21 13:24:21] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3225.7673524322245
[2024/02/21 13:24:21] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:21] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3225.7673524322245, best_epoch: 1
[2024/02/21 13:24:26] ppocr INFO: epoch: [1/100], global_step: 1400, lr: 0.001000, acc: 1.000000, loss: 0.016639, avg_reader_cost: 0.02485 s, avg_batch_cost: 0.04235 s, avg_samples: 32.0, ips: 755.55466 samples/s, eta: 10:01:26
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.83it/s]
[2024/02/21 13:24:26] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2593.9301548035482
[2024/02/21 13:24:26] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:26] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2593.9301548035482, best_epoch: 1
[2024/02/21 13:24:30] ppocr INFO: epoch: [1/100], global_step: 1500, lr: 0.001000, acc: 1.000000, loss: 0.021692, avg_reader_cost: 0.01923 s, avg_batch_cost: 0.03987 s, avg_samples: 32.0, ips: 802.65312 samples/s, eta: 9:58:43
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.99it/s]
[2024/02/21 13:24:30] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2682.2087929656277
[2024/02/21 13:24:30] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:30] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2682.2087929656277, best_epoch: 1
[2024/02/21 13:24:34] ppocr INFO: epoch: [1/100], global_step: 1600, lr: 0.001000, acc: 1.000000, loss: 0.016204, avg_reader_cost: 0.02375 s, avg_batch_cost: 0.04227 s, avg_samples: 32.0, ips: 757.10513 samples/s, eta: 9:58:25
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 12.09it/s]
[2024/02/21 13:24:34] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3305.3668915923754
[2024/02/21 13:24:34] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:34] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3305.3668915923754, best_epoch: 1
[2024/02/21 13:24:39] ppocr INFO: epoch: [1/100], global_step: 1700, lr: 0.001000, acc: 1.000000, loss: 0.011013, avg_reader_cost: 0.02315 s, avg_batch_cost: 0.04209 s, avg_samples: 32.0, ips: 760.20461 samples/s, eta: 9:58:01
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.99it/s]
[2024/02/21 13:24:39] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2619.086914100613
[2024/02/21 13:24:39] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:39] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2619.086914100613, best_epoch: 1
[2024/02/21 13:24:43] ppocr INFO: epoch: [1/100], global_step: 1800, lr: 0.001000, acc: 1.000000, loss: 0.008535, avg_reader_cost: 0.02464 s, avg_batch_cost: 0.04277 s, avg_samples: 32.0, ips: 748.13473 samples/s, eta: 9:58:11
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.28it/s]
[2024/02/21 13:24:43] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2492.807250845065
[2024/02/21 13:24:43] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:43] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2492.807250845065, best_epoch: 1
[2024/02/21 13:24:47] ppocr INFO: epoch: [1/100], global_step: 1900, lr: 0.001000, acc: 1.000000, loss: 0.005821, avg_reader_cost: 0.02216 s, avg_batch_cost: 0.04139 s, avg_samples: 32.0, ips: 773.20849 samples/s, eta: 9:57:17
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 12.28it/s]
[2024/02/21 13:24:47] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 2284.284902224416
[2024/02/21 13:24:47] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:47] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 2284.284902224416, best_epoch: 1
[2024/02/21 13:24:52] ppocr INFO: epoch: [1/100], global_step: 2000, lr: 0.001000, acc: 1.000000, loss: 0.013334, avg_reader_cost: 0.02400 s, avg_batch_cost: 0.04278 s, avg_samples: 32.0, ips: 747.96259 samples/s, eta: 9:57:28
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 13.77it/s]
[2024/02/21 13:24:52] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3465.8298817332025
[2024/02/21 13:24:52] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:52] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3465.8298817332025, best_epoch: 1
[2024/02/21 13:24:56] ppocr INFO: epoch: [1/100], global_step: 2100, lr: 0.001000, acc: 0.984375, loss: 0.036552, avg_reader_cost: 0.02460 s, avg_batch_cost: 0.04339 s, avg_samples: 32.0, ips: 737.49092 samples/s, eta: 9:58:02
eval model:: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 14.66it/s]
[2024/02/21 13:24:56] ppocr INFO: cur metric, acc: 0.9999996875000976, fps: 3144.6715868887795
[2024/02/21 13:24:56] ppocr INFO: save best model is to ./output/cls/mv3/best_accuracy
[2024/02/21 13:24:56] ppocr INFO: best metric, acc: 0.9999996875000976, is_float16: False, fps: 3144.6715868887795, best_epoch: 1
[2024/02/21 13:25:01] ppocr INFO: epoch: [1/100], global_step: 2200, lr: 0.001000, acc: 1.000000, loss: 0.012487, avg_reader_cost: 0.02222 s, avg_batch_cost: 0.04177 s, avg_samples: 32.0, ips: 766.11362 samples/s, eta: 9:57:30
eval model::   0%|                                                                                                                               | 0/1 [00:00<?, ?it/s]

Model is stop after that log " eval model:: 0%| | 0/1 [00:00<?, ?it/s]"

Sunting78 · 2024-02-23T03:00:33Z

Sunting78
Feb 23, 2024
Collaborator

您好，请问这个问题是能够稳定复现的吗？看起来之前的eval model都能正常进行，检测一下是否有程序中断

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

During train Text Direction Classification, training is stop at loading eval model. #12966

Uh oh!

{{title}}

Uh oh!

Replies: 1 comment

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

During train Text Direction Classification, training is stop at loading eval model. #12966

Uh oh!

diagram298 Feb 21, 2024

Replies: 1 comment

Uh oh!

Sunting78 Feb 23, 2024 Collaborator

diagram298
Feb 21, 2024

Sunting78
Feb 23, 2024
Collaborator