Search forum and issue tracker for opencl topic. Zdenko
st 19. 2. 2020 o 8:27 Tim Finnegan <[email protected]> napísal(a): > I'm attempting to run GPU Acceleration during training using the OpenCL > libraries. > > I have built tesseract to use openCL, and installed the NVidia Compute > driver 440 on my Ubuntu 19.10 installation > > Whenever I run tesstrain.sh, however, I run into the issue that the > program refuses to select the proper GPU. Rather than use my > NVidia GeForce GTX 1060 6GB device, it will select the CPU as my default > OpenCL device even though it detects my GPU, > and scores it better in the built-in benchmark. > > Setting TESSERACT_OPENCL_DEVICE=1 seems to do nothing as nvidia-smi shows > that the process is not utilizing my GPU. > > Here is my tesstrain.sh output: > > === Starting training for language 'eng' >> [Tue 18 Feb 2020 04:55:13 PM PST] /usr/local/bin/text2image >> --fonts_dir=/usr/share/fonts --ptsize 12 --font=Chit >> --outputbase=/tmp/font_tmp.Hk8xAdjwI8/sample_text.txt >> --text=/tmp/font_tmp.Hk8xAdjwI8/sample_text.txt >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 >> Rendered page 0 to file /tmp/font_tmp.Hk8xAdjwI8/sample_text.txt.tif >> === Phase I: Generating training images === >> Rendering using Chit >> [Tue 18 Feb 2020 04:55:15 PM PST] /usr/local/bin/text2image >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 --fonts_dir=/usr/share/fonts >> --strip_unrenderable_words --leading=32 --xsize=3600 --char_spacing=0.0 >> --exposure=1 --outputbase=/tmp/eng-2020-02-18.Gfj/eng.Chit.exp1 >> --max_pages=0 --font=Chit --ptsize 12 >> --text=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.training_text >> Stripped 35 unrenderable words >> Rendered page 0 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.tif >> Stripped 6 unrenderable words >> Rendered page 1 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.tif >> Rendering using Chit >> [Tue 18 Feb 2020 04:55:17 PM PST] /usr/local/bin/text2image >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 --fonts_dir=/usr/share/fonts >> --strip_unrenderable_words --leading=32 --xsize=3600 --char_spacing=0.0 >> --exposure=2 --outputbase=/tmp/eng-2020-02-18.Gfj/eng.Chit.exp2 >> --max_pages=0 --font=Chit --ptsize 12 >> --text=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.training_text >> Stripped 35 unrenderable words >> Rendered page 0 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.tif >> Stripped 6 unrenderable words >> Rendered page 1 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.tif >> Rendering using Chit >> [Tue 18 Feb 2020 04:55:19 PM PST] /usr/local/bin/text2image >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 --fonts_dir=/usr/share/fonts >> --strip_unrenderable_words --leading=32 --xsize=3600 --char_spacing=0.0 >> --exposure=3 --outputbase=/tmp/eng-2020-02-18.Gfj/eng.Chit.exp3 >> --max_pages=0 --font=Chit --ptsize 12 >> --text=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.training_text >> Stripped 35 unrenderable words >> Rendered page 0 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.tif >> Stripped 6 unrenderable words >> Rendered page 1 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.tif >> Rendering using Chit >> [Tue 18 Feb 2020 04:55:22 PM PST] /usr/local/bin/text2image >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 --fonts_dir=/usr/share/fonts >> --strip_unrenderable_words --leading=32 --xsize=3600 --char_spacing=0.0 >> --exposure=4 --outputbase=/tmp/eng-2020-02-18.Gfj/eng.Chit.exp4 >> --max_pages=0 --font=Chit --ptsize 12 >> --text=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.training_text >> Stripped 35 unrenderable words >> Rendered page 0 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.tif >> Stripped 6 unrenderable words >> Rendered page 1 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.tif >> Rendering using Chit >> [Tue 18 Feb 2020 04:55:24 PM PST] /usr/local/bin/text2image >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 --fonts_dir=/usr/share/fonts >> --strip_unrenderable_words --leading=32 --xsize=3600 --char_spacing=0.0 >> --exposure=5 --outputbase=/tmp/eng-2020-02-18.Gfj/eng.Chit.exp5 >> --max_pages=0 --font=Chit --ptsize 12 >> --text=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.training_text >> Stripped 35 unrenderable words >> Rendered page 0 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.tif >> Stripped 6 unrenderable words >> Rendered page 1 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.tif >> Rendering using Chit >> [Tue 18 Feb 2020 04:55:27 PM PST] /usr/local/bin/text2image >> --fontconfig_tmpdir=/tmp/font_tmp.Hk8xAdjwI8 --fonts_dir=/usr/share/fonts >> --strip_unrenderable_words --leading=32 --xsize=3600 --char_spacing=0.0 >> --exposure=6 --outputbase=/tmp/eng-2020-02-18.Gfj/eng.Chit.exp6 >> --max_pages=0 --font=Chit --ptsize 12 >> --text=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.training_text >> Stripped 35 unrenderable words >> Rendered page 0 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.tif >> Stripped 6 unrenderable words >> Rendered page 1 to file /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.tif >> === Phase UP: Generating unicharset and unichar properties files === >> [Tue 18 Feb 2020 04:55:28 PM PST] /usr/local/bin/unicharset_extractor >> --output_unicharset /tmp/eng-2020-02-18.Gfj/eng.unicharset --norm_mode 1 >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.box >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.box >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.box >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.box >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.box >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.box >> Extracting unicharset from box file >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.box >> Extracting unicharset from box file >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.box >> Extracting unicharset from box file >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.box >> Extracting unicharset from box file >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.box >> Extracting unicharset from box file >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.box >> Extracting unicharset from box file >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.box >> Other case É of é is not in unicharset >> Wrote unicharset file /tmp/eng-2020-02-18.Gfj/eng.unicharset >> [Tue 18 Feb 2020 04:55:29 PM PST] >> /usr/local/bin/set_unicharset_properties -U >> /tmp/eng-2020-02-18.Gfj/eng.unicharset -O >> /tmp/eng-2020-02-18.Gfj/eng.unicharset -X >> /tmp/eng-2020-02-18.Gfj/eng.xheights >> --script_dir=/home/tim/PycharmProjects/RnD/OCR_Dataset/langdata >> Loaded unicharset of size 102 from file >> /tmp/eng-2020-02-18.Gfj/eng.unicharset >> Setting unichar properties >> Other case É of é is not in unicharset >> Setting script properties >> Warning: properties incomplete for index 25 = ~ >> Writing unicharset to file /tmp/eng-2020-02-18.Gfj/eng.unicharset >> === Phase E: Generating lstmf files === >> Using >> TESSDATA_PREFIX=/home/tim/PycharmProjects/RnD/OCR_Dataset/tessdata_best >> [Tue 18 Feb 2020 04:55:29 PM PST] /usr/local/bin/tesseract >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.tif >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1 --psm 6 lstm.train >> [Tue 18 Feb 2020 04:55:29 PM PST] /usr/local/bin/tesseract >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.tif >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3 --psm 6 lstm.train >> [Tue 18 Feb 2020 04:55:29 PM PST] /usr/local/bin/tesseract >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.tif >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4 --psm 6 lstm.train >> [Tue 18 Feb 2020 04:55:29 PM PST] /usr/local/bin/tesseract >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.tif >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2 --psm 6 lstm.train >> [Tue 18 Feb 2020 04:55:29 PM PST] /usr/local/bin/tesseract >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.tif >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5 --psm 6 lstm.train >> [Tue 18 Feb 2020 04:55:29 PM PST] /usr/local/bin/tesseract >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.tif >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6 --psm 6 lstm.train >> [DS] Profile read from file (tesseract_opencl_profile_devices.dat). >> [DS] Device[1] 1:GeForce GTX 1060 6GB score is 1.846448 >> [DS] Profile read from file (tesseract_opencl_profile_devices.dat). >> [DS] Device[2] 0:(null) score is 0.503215 >> [DS] Selected Device[2]: "(null)" (Native) >> [DS] Overriding Device Selection (TESSERACT_OPENCL_DEVICE=1, 1) >> [DS] Overridden Device[1]: "GeForce GTX 1060 6GB" (OpenCL) >> [DS] Device[1] 1:GeForce GTX 1060 6GB score is 1.846448 >> [DS] Device[2] 0:(null) score is 0.503215 >> [DS] Selected Device[2]: "(null)" (Native) >> [DS] Overriding Device Selection (TESSERACT_OPENCL_DEVICE=1, 1) >> [DS] Overridden Device[1]: "GeForce GTX 1060 6GB" (OpenCL) >> [DS] Profile read from file (tesseract_opencl_profile_devices.dat). >> [DS] Device[1] 1:GeForce GTX 1060 6GB score is 1.846448 >> [DS] Device[2] 0:(null) score is 0.503215 >> [DS] Selected Device[2]: "(null)" (Native) >> [DS] Overriding Device Selection (TESSERACT_OPENCL_DEVICE=1, 1) >> [DS] Overridden Device[1]: "GeForce GTX 1060 6GB" (OpenCL) >> [DS] Profile read from file (tesseract_opencl_profile_devices.dat). >> [DS] Device[1] 1:GeForce GTX 1060 6GB score is 1.846448 >> [DS] Device[2] 0:(null) score is 0.503215 >> [DS] Selected Device[2]: "(null)" (Native) >> [DS] Overriding Device Selection (TESSERACT_OPENCL_DEVICE=1, 1) >> [DS] Overridden Device[1]: "GeForce GTX 1060 6GB" (OpenCL) >> [DS] Profile read from file (tesseract_opencl_profile_devices.dat). >> [DS] Device[1] 1:GeForce GTX 1060 6GB score is 1.846448 >> [DS] Device[2] 0:(null) score is 0.503215 >> [DS] Selected Device[2]: "(null)" (Native) >> [DS] Overriding Device Selection (TESSERACT_OPENCL_DEVICE=1, 1) >> [DS] Overridden Device[1]: "GeForce GTX 1060 6GB" (OpenCL) >> [DS] Profile read from file (tesseract_opencl_profile_devices.dat). >> [DS] Device[1] 1:GeForce GTX 1060 6GB score is 1.846448 >> [DS] Device[2] 0:(null) score is 0.503215 >> [DS] Selected Device[2]: "(null)" (Native) >> [DS] Overriding Device Selection (TESSERACT_OPENCL_DEVICE=1, 1) >> [DS] Overridden Device[1]: "GeForce GTX 1060 6GB" (OpenCL) >> Tesseract Open Source OCR Engine v5.0.0-alpha with Leptonica >> Tesseract Open Source OCR Engine v5.0.0-alpha with Leptonica >> Tesseract Open Source OCR Engine v5.0.0-alpha with Leptonica >> Page 1 >> Page 1 >> Page 1 >> Tesseract Open Source OCR Engine v5.0.0-alpha with Leptonica >> Page 1 >> Tesseract Open Source OCR Engine v5.0.0-alpha with Leptonica >> Tesseract Open Source OCR Engine v5.0.0-alpha with Leptonica >> Page 1 >> Page 1 >> Page 2 >> Loaded 56/56 lines (1-56) of document >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.lstmf >> Page 2 >> Page 2 >> Page 2 >> Loaded 56/56 lines (1-56) of document >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.lstmf >> Loaded 56/56 lines (1-56) of document >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.lstmf >> Loaded 56/56 lines (1-56) of document >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.lstmf >> Page 2 >> Loaded 56/56 lines (1-56) of document >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.lstmf >> Page 2 >> Loaded 56/56 lines (1-56) of document >> /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.lstmf >> === Constructing LSTM training data === >> [Tue 18 Feb 2020 04:55:33 PM PST] /usr/local/bin/combine_lang_model >> --input_unicharset /tmp/eng-2020-02-18.Gfj/eng.unicharset --script_dir >> /home/tim/PycharmProjects/RnD/OCR_Dataset/langdata --words >> /home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.wordlist >> --numbers >> /home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.numbers --puncs >> /home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.punc >> --output_dir /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA --lang eng >> Loaded unicharset of size 102 from file >> /tmp/eng-2020-02-18.Gfj/eng.unicharset >> Setting unichar properties >> Other case É of é is not in unicharset >> Setting script properties >> Config file is optional, continuing... >> Failed to read data from: >> /home/tim/PycharmProjects/RnD/OCR_Dataset/langdata/eng/eng.config >> Null char=2 >> Reducing Trie to SquishedDawg >> Reducing Trie to SquishedDawg >> Reducing Trie to SquishedDawg >> === Saving box/tiff pairs for training data === >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.box to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.box to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.box to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.box to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.box to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.box to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.tif to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.tif to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.tif to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.tif to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.tif to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.tif to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> === Moving lstmf files for training data === >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp1.lstmf to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp2.lstmf to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp3.lstmf to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp4.lstmf to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp5.lstmf to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA >> Moving /tmp/eng-2020-02-18.Gfj/eng.Chit.exp6.lstmf to >> /home/tim/PycharmProjects/RnD/OCR_Dataset/DATA > > > Created starter traineddata for LSTM training of language 'eng' > > > Run 'lstmtraining' command to continue LSTM training for language 'eng' > > > And here is my nvidia-smi output during the training process: > > Tue Feb 18 17:01:50 2020 >> >> +-----------------------------------------------------------------------------+ >> | NVIDIA-SMI 440.33.01 Driver Version: 440.33.01 CUDA Version: >> 10.2 | >> >> |-------------------------------+----------------------+----------------------+ >> | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. >> ECC | >> | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util >> Compute M. | >> >> |===============================+======================+======================| >> | 0 GeForce GTX 106... On | 00000000:01:00.0 On | >> N/A | >> | 22% 56C P0 23W / 120W | 402MiB / 6072MiB | 0% >> Default | >> >> +-------------------------------+----------------------+----------------------+ >> >> >> >> +-----------------------------------------------------------------------------+ >> | Processes: GPU >> Memory | >> | GPU PID Type Process name Usage >> | >> >> |=============================================================================| >> | 0 1254 G /usr/lib/xorg/Xorg >> 32MiB | >> | 0 2003 G /usr/lib/xorg/Xorg >> 155MiB | >> | 0 2223 G /usr/bin/gnome-shell >> 97MiB | >> | 0 2724 G ...p/pycharm-professional/183/jbr/bin/java >> 2MiB | >> | 0 4932 G /usr/bin/nvidia-settings >> 0MiB | >> | 0 5071 G ...AAAAAAAAAAAAAAgAAAAAAAAA --shared-files >> 62MiB | >> >> +-----------------------------------------------------------------------------+ >> > > -- > You received this message because you are subscribed to the Google Groups > "tesseract-ocr" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/tesseract-ocr/984f04a4-c5da-495c-8231-83c15e51fd77%40googlegroups.com > <https://groups.google.com/d/msgid/tesseract-ocr/984f04a4-c5da-495c-8231-83c15e51fd77%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/CAJbzG8zAnvW83XrLLijsTMCGz1nEZEaCZXxhGzs7E5eajc0h2g%40mail.gmail.com.
