W0910 21:18:56.955332 3317990 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 6.1, Driver API Version: 12.0, Runtime API Version: 10.2 W0910 21:18:56.955994 3317990 gpu_resources.cc:91] device: 0, cuDNN Version: 8.3. #12674
Unanswered
juneana123
asked this question in
Q&A
Replies: 3 comments
-
|
可以正常训练检测模型，但是在训练识别模型时GPU显示调用，但是只有1%,一直无法进行下一步的训练step |
Beta Was this translation helpful? Give feedback.
0 replies
-
是不是你数据量太大了，这一步是要等很久，训练到第一步就快很多了 |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
我重新更换gpu版本后，显示调用了GPU。对于A数据集可以正常训练，但是对于B数据训练缓慢，而且A数据集数据量是大于B数据集的，这是怎么回事呢。 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
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
[2023/09/10 21:22:08] ppocr INFO: Architecture :
[2023/09/10 21:22:08] ppocr INFO: Backbone :
[2023/09/10 21:22:08] ppocr INFO: depth : [3, 6, 3]
[2023/09/10 21:22:08] ppocr INFO: embed_dim : [64, 128, 256]
[2023/09/10 21:22:08] ppocr INFO: img_size : [32, 320]
[2023/09/10 21:22:08] ppocr INFO: last_stage : True
[2023/09/10 21:22:08] ppocr INFO: local_mixer : [[7, 11], [7, 11], [7, 11]]
[2023/09/10 21:22:08] ppocr INFO: mixer : ['Local', 'Local', 'Local', 'Local', 'Local', 'Local', 'Global', 'Global', 'Global', 'Global', 'Global', 'Global']
[2023/09/10 21:22:08] ppocr INFO: name : SVTRNet
[2023/09/10 21:22:08] ppocr INFO: num_heads : [2, 4, 8]
[2023/09/10 21:22:08] ppocr INFO: out_channels : 96
[2023/09/10 21:22:08] ppocr INFO: out_char_num : 40
[2023/09/10 21:22:08] ppocr INFO: patch_merging : Conv
[2023/09/10 21:22:08] ppocr INFO: prenorm : False
[2023/09/10 21:22:08] ppocr INFO: Head :
[2023/09/10 21:22:08] ppocr INFO: name : CTCHead
[2023/09/10 21:22:08] ppocr INFO: Neck :
[2023/09/10 21:22:08] ppocr INFO: encoder_type : reshape
[2023/09/10 21:22:08] ppocr INFO: name : SequenceEncoder
[2023/09/10 21:22:08] ppocr INFO: Transform : None
[2023/09/10 21:22:08] ppocr INFO: algorithm : SVTR
[2023/09/10 21:22:08] ppocr INFO: model_type : rec
[2023/09/10 21:22:08] ppocr INFO: Eval :
[2023/09/10 21:22:08] ppocr INFO: dataset :
[2023/09/10 21:22:08] ppocr INFO: data_dir : /media/wj/newvolume/REC_DATASET/RCTW_rec/RCTW_recval_imgs/
[2023/09/10 21:22:08] ppocr INFO: label_file_list : ['/media/wj/newvolume/REC_DATASET/RCTW_rec/rec_val.txt']
[2023/09/10 21:22:08] ppocr INFO: name : SimpleDataSet
[2023/09/10 21:22:08] ppocr INFO: transforms :
[2023/09/10 21:22:08] ppocr INFO: DecodeImage :
[2023/09/10 21:22:08] ppocr INFO: channel_first : False
[2023/09/10 21:22:08] ppocr INFO: img_mode : BGR
[2023/09/10 21:22:08] ppocr INFO: CTCLabelEncode : None
[2023/09/10 21:22:08] ppocr INFO: SVTRRecResizeImg :
[2023/09/10 21:22:08] ppocr INFO: image_shape : [3, 32, 320]
[2023/09/10 21:22:08] ppocr INFO: padding : True
[2023/09/10 21:22:08] ppocr INFO: KeepKeys :
[2023/09/10 21:22:08] ppocr INFO: keep_keys : ['image', 'label', 'length']
[2023/09/10 21:22:08] ppocr INFO: loader :
[2023/09/10 21:22:08] ppocr INFO: batch_size_per_card : 256
[2023/09/10 21:22:08] ppocr INFO: drop_last : False
[2023/09/10 21:22:08] ppocr INFO: num_workers : 2
[2023/09/10 21:22:08] ppocr INFO: shuffle : False
[2023/09/10 21:22:08] ppocr INFO: Global :
[2023/09/10 21:22:08] ppocr INFO: cal_metric_during_train : True
[2023/09/10 21:22:08] ppocr INFO: character_dict_path : ppocr/utils/baidu_rec_dict.txt
[2023/09/10 21:22:08] ppocr INFO: character_type : ch
[2023/09/10 21:22:08] ppocr INFO: checkpoints : None
[2023/09/10 21:22:08] ppocr INFO: distributed : False
[2023/09/10 21:22:08] ppocr INFO: epoch_num : 50
[2023/09/10 21:22:08] ppocr INFO: eval_batch_step : [0, 2000]
[2023/09/10 21:22:08] ppocr INFO: infer_img : doc/imgs_words/ch/word_1.jpg
[2023/09/10 21:22:08] ppocr INFO: infer_mode : False
[2023/09/10 21:22:08] ppocr INFO: log_smooth_window : 20
[2023/09/10 21:22:08] ppocr INFO: max_text_length : 100
[2023/09/10 21:22:08] ppocr INFO: pretrained_model : None
[2023/09/10 21:22:08] ppocr INFO: print_batch_step : 1
[2023/09/10 21:22:08] ppocr INFO: save_epoch_step : 10
[2023/09/10 21:22:08] ppocr INFO: save_inference_dir : None
[2023/09/10 21:22:08] ppocr INFO: save_model_dir : ./output/myrec/rec_svtr_RC
[2023/09/10 21:22:08] ppocr INFO: save_res_path : ./output/myrec/rec_svtr_RC/predicts_svtr_tiny_ch_all.txt
[2023/09/10 21:22:08] ppocr INFO: use_gpu : True
[2023/09/10 21:22:08] ppocr INFO: use_space_char : True
[2023/09/10 21:22:08] ppocr INFO: use_visualdl : True
[2023/09/10 21:22:08] ppocr INFO: Loss :
[2023/09/10 21:22:08] ppocr INFO: name : CTCLoss
[2023/09/10 21:22:08] ppocr INFO: Metric :
[2023/09/10 21:22:08] ppocr INFO: main_indicator : acc
[2023/09/10 21:22:08] ppocr INFO: name : RecMetric
[2023/09/10 21:22:08] ppocr INFO: Optimizer :
[2023/09/10 21:22:08] ppocr INFO: beta1 : 0.9
[2023/09/10 21:22:08] ppocr INFO: beta2 : 0.99
[2023/09/10 21:22:08] ppocr INFO: epsilon : 8e-08
[2023/09/10 21:22:08] ppocr INFO: lr :
[2023/09/10 21:22:08] ppocr INFO: learning_rate : 0.0005
[2023/09/10 21:22:08] ppocr INFO: name : Cosine
[2023/09/10 21:22:08] ppocr INFO: warmup_epoch : 2
[2023/09/10 21:22:08] ppocr INFO: name : AdamW
[2023/09/10 21:22:08] ppocr INFO: no_weight_decay_name : norm pos_embed
[2023/09/10 21:22:08] ppocr INFO: one_dim_param_no_weight_decay : True
[2023/09/10 21:22:08] ppocr INFO: weight_decay : 0.05
[2023/09/10 21:22:08] ppocr INFO: PostProcess :
[2023/09/10 21:22:08] ppocr INFO: name : CTCLabelDecode
[2023/09/10 21:22:08] ppocr INFO: Train :
[2023/09/10 21:22:08] ppocr INFO: dataset :
[2023/09/10 21:22:08] ppocr INFO: data_dir : /media/wj/newvolume/REC_DATASET/RCTW_rec/RCTW_rectrain_imgs/
[2023/09/10 21:22:08] ppocr INFO: ext_op_transform_idx : 1
[2023/09/10 21:22:08] ppocr INFO: label_file_list : ['/media/wj/newvolume/REC_DATASET/RCTW_rec/rec_train.txt']
[2023/09/10 21:22:08] ppocr INFO: name : SimpleDataSet
[2023/09/10 21:22:08] ppocr INFO: transforms :
[2023/09/10 21:22:08] ppocr INFO: DecodeImage :
[2023/09/10 21:22:08] ppocr INFO: channel_first : False
[2023/09/10 21:22:08] ppocr INFO: img_mode : BGR
[2023/09/10 21:22:08] ppocr INFO: RecConAug :
[2023/09/10 21:22:08] ppocr INFO: ext_data_num : 2
[2023/09/10 21:22:08] ppocr INFO: image_shape : [32, 320, 3]
[2023/09/10 21:22:08] ppocr INFO: prob : 0.5
[2023/09/10 21:22:08] ppocr INFO: RecAug : None
[2023/09/10 21:22:08] ppocr INFO: CTCLabelEncode : None
[2023/09/10 21:22:08] ppocr INFO: SVTRRecResizeImg :
[2023/09/10 21:22:08] ppocr INFO: image_shape : [3, 32, 320]
[2023/09/10 21:22:08] ppocr INFO: padding : True
[2023/09/10 21:22:08] ppocr INFO: KeepKeys :
[2023/09/10 21:22:08] ppocr INFO: keep_keys : ['image', 'label', 'length']
[2023/09/10 21:22:08] ppocr INFO: loader :
[2023/09/10 21:22:08] ppocr INFO: batch_size_per_card : 128
[2023/09/10 21:22:08] ppocr INFO: drop_last : True
[2023/09/10 21:22:08] ppocr INFO: num_workers : 8
[2023/09/10 21:22:08] ppocr INFO: shuffle : True
[2023/09/10 21:22:08] ppocr INFO: profiler_options : None
[2023/09/10 21:22:08] ppocr INFO: train with paddle 2.3.2 and device Place(gpu:0)
[2023/09/10 21:22:08] ppocr INFO: Initialize indexs of datasets:['/media/wj/newvolume/REC_DATASET/RCTW_rec/rec_train.txt']
[2023/09/10 21:22:08] ppocr INFO: Initialize indexs of datasets:['/media/wj/newvolume/REC_DATASET/RCTW_rec/rec_val.txt']
W0910 21:22:08.206351 3318983 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 6.1, Driver API Version: 12.0, Runtime API Version: 10.2
W0910 21:22:08.207228 3318983 gpu_resources.cc:91] device: 0, cuDNN Version: 8.3.
[2023/09/10 21:22:08] ppocr INFO: train dataloader has 342 iters
[2023/09/10 21:22:08] ppocr INFO: valid dataloader has 49 iters
[2023/09/10 21:22:08] ppocr INFO: train from scratch
[2023/09/10 21:22:08] ppocr INFO: During the training process, after the 0th iteration, an evaluation is run every 2000 iterations
我们提供了AceIssueSolver来帮助你解答问题，你是否想要它来解答(请填写yes/no)?/We provide AceIssueSolver to solve issues, do you want it? (Please write yes/no):yes
请尽量不要包含图片在问题中/Please try to not include the image in the issue.
Beta Was this translation helpful? Give feedback.
All reactions