search

PaddlePaddle / PaddleSlim

PaddleSlim is an open-source library for deep model compression and architecture search.
https://paddleslim.readthedocs.io/zh_CN/latest/
Apache License 2.0
1.56k stars 345 forks source link

OCR模型自动压缩PPOCRV4_rec报错 #1883

Open huangguifeng opened 3 months ago

huangguifeng commented 3 months ago

2024-07-08 18:00:06,589-INFO：选定的策略：['qat_dis'] 回溯（最近一次调用最后一次）： 文件“run.py”，第 171 行，在 main()中 文件“run.py”，第 164 行，在 main ac.compress() 中 文件“/root/miniconda3/lib/python3.8/site-packages/paddleslim/auto_compression/compressor.py”，第 586 行，在 compress self.single_strategy_compress(strategy, config, strategies_idx, 文件“/root/miniconda3/lib/python3.8/site-packages/paddleslim/auto_compression/compressor.py”，第 769 行，在 single_strategy_compress train_program_info, test_program_info = self._prepare_program( 文件“/root/miniconda3/lib/python3.8/site-packages/paddleslim/auto_compression/compressor.py”, 第 506 行, 在 _prepare_program train_program_info 中, test_program_info = build_distill_program( 文件“/root/miniconda3/lib/python3.8/site-packages/paddleslim/auto_compression/create_compressed_program.py”, 第 327 行, 在 build_distill_program train_program 中, data_name_map = _load_program_and_merge( 文件“/root/miniconda3/lib/python3.8/site-packages/paddleslim/auto_compression/create_compressed_program.py”, 第 219 行, 在 _load_program_and_merge 中 合并( 文件“/root/miniconda3/lib/python3.8/site-packages/paddleslim/dist/single_distiller.py”, 行197，在合并 student_program.global_block().append_op( 文件“/root/miniconda3/lib/python3.8/site-packages/paddle/fluid/framework.py”中，第 4013 行，在 append_op op = Operator( 文件“/root/miniconda3/lib/python3.8/site-packages/paddle/fluid/framework.py”中，第 2889 行，在init 中 引发 ValueError( ValueError: 运算符“dropout”的输出设置不正确，应设置：[Mask]。

huangguifeng commented 3 months ago

另外 det模型 自动压缩召回率一直为0 2024-07-08 18:17:02,228-INFO: metric eval *** 2024-07-08 18:17:02,228-INFO: precision:0.0 2024-07-08 18:17:02,228-INFO: recall:0.0 2024-07-08 18:17:02,228-INFO: hmean:0

huangguifeng commented 3 months ago

直接用这个案例 ，然后模型也是文档里提供的模型。但是运行报这个错误 https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.8/deploy/slim/auto_compression/run.py

Global: model_type: det model_dir: C:\Users\Administrator\Downloads\det_server_qat_v100 model_filename: inference.pdmodel params_filename: inference.pdiparams algorithm: DB

C++ Traceback (most recent call last):

0 paddle::framework::StandaloneExecutor::Run(paddle::framework::Scope, std::vector<std::string, std::allocator > const&, std::vector<std::string, std::allocator > const&) 1 paddle::framework::InterpreterCore::Run(std::vector<std::string, std::allocator > const&, bool) 2 paddle::framework::interpreter::BuildOpFuncList(phi::Place const&, paddle::framework::BlockDesc const&, std::set<std::string, std::less, std::allocator > const&, std::vector<paddle::framework::OpFuncNode, std::allocator >, paddle::framework::VariableScope, paddle::framework::interpreter::ExecutionConfig const&, bool, bool) 3 paddle::framework::StructKernelImpl<paddle::operators::FakeQuantizeMovingAverageAbsMaxKernel<float, phi::GPUContext>, void>::Compute(phi::KernelContext) 4 paddle::operators::FakeMovingAverageAbsMaxKernelBase<float, phi::GPUContext>::Compute(paddle::framework::ExecutionContext const&) const 5 phi::DenseTensor::mutable_data(phi::Place const&, phi::DataType, unsigned long) 6 phi::DenseTensor::set_type(phi::DataType)

Error Message Summary:

FatalError: Segmentation fault is detected by the operating system. [TimeInfo: Aborted at 1720491686 (unix time) try "date -d @1720491686" if you are using GNU date ] [SignalInfo: SIGSEGV (@0x68) received by PID 40490 (TID 0x7f739046a340) from PID 104 ]

这是需要修改那里吗？为什么运行会报这个错误。。假如用 自己训练的模型召回率会一直为0