How to use GPU in paddleocr

[joel1895]

I have set the flag use_gpu = True in paddle ocr, and still it's using cpu. How can we make this right?

[shiyutang]

To use GPU in PaddleOCR, there are a few steps you need to follow:

1. Install the GPU version of PaddlePaddle: Make sure you have installed the GPU version of PaddlePaddle, which includes CUDA and cuDNN support. You can install it by following the instructions provided in the PaddlePaddle documentation.

2. Check GPU availability: Ensure that your GPU is properly recognized by PaddlePaddle. You can run the following code snippet to check if PaddlePaddle can detect your GPU:

   import paddle
   gpu_available  = paddle.device.is_compiled_with_cuda()
   print("GPU available:", gpu_available)

   If the output is True, it means PaddlePaddle has detected the GPU.

3. Set the use_gpu flag in PaddleOCR: In your code or configuration file, make sure you set the use_gpu flag to True to enable GPU usage. For example, if you are using the PaddleOCR API, you can set it like this:

   from paddleocr import PaddleOCR
   
   ocr = PaddleOCR(use_gpu=True)

   If you are using a configuration file, set the use_gpu flag to true in the configuration file.

4. Check GPU usage during inference: When performing inference with PaddleOCR, you can monitor GPU usage using tools such as nvidia-smi (for NVIDIA GPUs) or other GPU monitoring tools. This will help you verify if the GPU is being utilized during inference.

If you have followed these steps and still encounter issues where PaddleOCR is using the CPU instead of the GPU, there might be other factors causing the problem. Ensure that your system environment and dependencies are correctly set up for GPU usage, and double-check the installation of PaddlePaddle with GPU support. Additionally, check for any error messages or warnings during the PaddleOCR initialization or inference process, as they might provide more insights into the issue.

[AGRocky]

Hey guys I am finding errors in importing PaddleOcr, draw_ocr. I am trying to run it on colab could guys plz help me out with this
ImportError: libssl.so.1.1: cannot open shared object file: No such file or directory

[lijwchn]

Find the version in this site (http://security.ubuntu.com/ubuntu/pool/main/o/openssl/) that suits your system to install , my computer is ubuntu，such as:

wget http://security.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.1-1ubuntu2.1~18.04.23_amd64.deb
sudo dpkg -i libssl1.1_1.1.1-1ubuntu2.1~18.04.23_amd64.deb

[codenoid]

@shiyutang

cuda-samples/Samples/1_Utilities/deviceQuery$ ./deviceQuery
./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "NVIDIA GeForce RTX 3060"
  CUDA Driver Version / Runtime Version          12.2 / 12.2
  CUDA Capability Major/Minor version number:    8.6
  Total amount of global memory:                 12288 MBytes (12884377600 bytes)
  (028) Multiprocessors, (128) CUDA Cores/MP:    3584 CUDA Cores
  GPU Max Clock rate:                            1777 MHz (1.78 GHz)
  Memory Clock rate:                             7501 Mhz
  Memory Bus Width:                              192-bit
  L2 Cache Size:                                 2359296 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total shared memory per multiprocessor:        102400 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Managed Memory:                Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      No
  Device PCI Domain ID / Bus ID / location ID:   0 / 1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 12.2, CUDA Runtime Version = 12.2, NumDevs = 1
Result = PASS

but when I try to use paddle to detect gpu, the result are:

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.

[youcanyouupsb]

Python 3.10.13 (main, Sep 11 2023, 13:44:35) [GCC 11.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.

import paddle
paddle.utils.run_check()
Running verify PaddlePaddle program ...
I0425 16:05:19.611985 1475101 interpretercore.cc:237] New Executor is Running.
I0425 16:05:19.667097 1475101 interpreter_util.cc:518] Standalone Executor is Used.
PaddlePaddle works well on 1 CPU.
PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now.
gpu_available=paddle.device.is_compiled_with_cuda()
print("GPU available:",gpu_available)
GPU available: False

[youcanyouupsb]

env

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Thu_Mar_28_02:18:24_PDT_2024
Cuda compilation tools, release 12.4, V12.4.131
Build cuda_12.4.r12.4/compiler.34097967_0

---------------------------
Name: paddleocr
Version: 2.7.0.1
--------------------------
Name: paddlepaddle
Version: 2.5.2
--------------------------
Name: paddlepaddle-gpu
Version: 2.5.1
-------------------------

Test Command:

$ python
>>> import paddle
>>> paddle.utils.run_check()
Running verify PaddlePaddle program ... 
PaddlePaddle works well on 1 GPU.
PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now.
>>> 
or
>>> gpu_available=paddle.device.is_compiled_with_cuda()
>>> print("GPU available:",gpu_available)
GPU available: True
>>>

---

Test Case:

from paddleocr import PaddleOCR, draw_ocr
import time
# Initialize PaddleOCR, attempt to use GPU
ocr = PaddleOCR(use_angle_cls=True, lang='ch', use_gpu=0, show_log=False)

# Read an image
image_path = 'cs.png'
# High precision timing
start_time = time.perf_counter()
# Use PaddleOCR for detection and recognition
result = ocr.ocr(image_path, cls=True)
# High precision timing ends
end_time = time.perf_counter()
# Calculate execution time
execution_time = end_time - start_time
print(f"High precision interface execution time: {execution_time} seconds")
# Print results
print(result)
# If needed, draw the recognition results on the image
# boxes, txts, scores = result
# for box, txt, score in zip(boxes, txts, scores):
#     print(f'Text: {txt}, Confidence: {score}')
#     draw_ocr(image_path, box, txt, score, 0, ocr.font_path)

---

Runtime Dependency Library Files:

libcudnn_ops_infer.so.8
libcudnn_cnn_train.so.8
libcudnn_cnn_infer.so.8
libcudnn.so.8

---

CPU Processing Result:

$ python test.py 
High precision interface execution time: 2.5933946020086296 seconds
[[[[84.0, 15.0], [101.0, 15.0], [101.0, 31.0], [84.0, 31.0]], ('C', 0.8629724383354187)]....

---

GPU Processing Result:

$ python test.py 
High precision interface execution time: 0.4999611990060657 seconds
[[[[84.0, 15.0], [101.0, 15.0], [101.0, 31.0], [84.0, 31.0]], ('C', 0.8637353181838989)].....

---

The result is perfect！