Created by: davidlinhl

在执行infer.py时遇到一个很模糊的报错，不清楚是什么情况

(med) [lin@lin-mjr infer]$ python nii_inf.py --conf=model/deploy.yaml --input_dir /home/lin/Desktop/github/med/med_lib/infer/imgs/ --ext png 
I0902 15:17:39.815834 17959 analysis_predictor.cc:138] Profiler is deactivated, and no profiling report will be generated.
I0902 15:17:39.830946 17959 analysis_predictor.cc:875] MODEL VERSION: 1.7.2
I0902 15:17:39.830972 17959 analysis_predictor.cc:877] PREDICTOR VERSION: 1.8.4
--- Running analysis [ir_graph_build_pass]
--- Running analysis [ir_graph_clean_pass]
--- Running analysis [ir_analysis_pass]
--- Running IR pass [is_test_pass]
--- Running IR pass [simplify_with_basic_ops_pass]
--- Running IR pass [conv_affine_channel_fuse_pass]
--- Running IR pass [conv_eltwiseadd_affine_channel_fuse_pass]
--- Running IR pass [conv_bn_fuse_pass]
I0902 15:17:39.870817 17959 graph_pattern_detector.cc:101] ---  detected 18 subgraphs
--- Running IR pass [conv_eltwiseadd_bn_fuse_pass]
--- Running IR pass [embedding_eltwise_layernorm_fuse_pass]
--- Running IR pass [multihead_matmul_fuse_pass_v2]
--- Running IR pass [fc_fuse_pass]
--- Running IR pass [fc_elementwise_layernorm_fuse_pass]
--- Running IR pass [conv_elementwise_add_act_fuse_pass]
I0902 15:17:39.883591 17959 graph_pattern_detector.cc:101] ---  detected 18 subgraphs
--- Running IR pass [conv_elementwise_add2_act_fuse_pass]
--- Running IR pass [conv_elementwise_add_fuse_pass]
--- Running IR pass [transpose_flatten_concat_fuse_pass]
--- Running IR pass [runtime_context_cache_pass]
--- Running analysis [ir_params_sync_among_devices_pass]
I0902 15:17:39.888110 17959 ir_params_sync_among_devices_pass.cc:41] Sync params from CPU to GPU
--- Running analysis [adjust_cudnn_workspace_size_pass]
--- Running analysis [inference_op_replace_pass]
--- Running analysis [memory_optimize_pass]
I0902 15:17:39.920619 17959 memory_optimize_pass.cc:223] Cluster name : image  size: 3145728
I0902 15:17:39.920639 17959 memory_optimize_pass.cc:223] Cluster name : relu_7.tmp_0  size: 8388608
I0902 15:17:39.920644 17959 memory_optimize_pass.cc:223] Cluster name : relu_9.tmp_0  size: 2097152
I0902 15:17:39.920650 17959 memory_optimize_pass.cc:223] Cluster name : bilinear_interp_1.tmp_0  size: 16777216
I0902 15:17:39.920655 17959 memory_optimize_pass.cc:223] Cluster name : relu_3.tmp_0  size: 33554432
I0902 15:17:39.920658 17959 memory_optimize_pass.cc:223] Cluster name : bilinear_interp_3.tmp_0  size: 67108864
I0902 15:17:39.920663 17959 memory_optimize_pass.cc:223] Cluster name : relu_1.tmp_0  size: 67108864
I0902 15:17:39.920668 17959 memory_optimize_pass.cc:223] Cluster name : concat_3.tmp_0  size: 134217728
--- Running analysis [ir_graph_to_program_pass]
I0902 15:17:39.927340 17959 analysis_predictor.cc:496] ======= optimize end =======
W0902 15:17:39.949319 17959 device_context.cc:252] Please NOTE: device: 0, CUDA Capability: 61, Driver API Version: 10.2, Runtime API Version: 10.0
W0902 15:17:39.950911 17959 device_context.cc:260] device: 0, cuDNN Version: 8.0.
W0902 15:17:41.156368 17959 init.cc:226] Warning: PaddlePaddle catches a failure signal, it may not work properly
W0902 15:17:41.156400 17959 init.cc:228] You could check whether you killed PaddlePaddle thread/process accidentally or report the case to PaddlePaddle
W0902 15:17:41.156406 17959 init.cc:231] The detail failure signal is:

W0902 15:17:41.156414 17959 init.cc:234] *** Aborted at 1599031061 (unix time) try "date -d @1599031061" if you are using GNU date ***
W0902 15:17:41.157629 17959 init.cc:234] PC: @                0x0 (unknown)
W0902 15:17:41.157713 17959 init.cc:234] *** SIGSEGV (@0x0) received by PID 17959 (TID 0x7f08f935d740) from PID 0; stack trace: ***
W0902 15:17:41.158828 17959 init.cc:234]     @     0x7f08f969b0f0 (unknown)
W0902 15:17:41.159871 17959 init.cc:234]     @                0x0 (unknown)
Segmentation fault (core dumped)

deploy.yaml

DEPLOY:
        USE_GPU : 1
        USE_PR : 0
        MODEL_PATH : "./model"
        MODEL_FILENAME : "__model__"
        PARAMS_FILENAME : "__params__"
        EVAL_CROP_SIZE : (512, 512)
        MEAN : [0.5, 0.5, 0.5]
        STD : [0.5, 0.5, 0.5]
        IMAGE_TYPE : "rgb"
        NUM_CLASSES : 2
        CHANNELS : 3
        PRE_PROCESSOR : "SegPreProcessor"
        PREDICTOR_MODE : "ANALYSIS"
        BATCH_SIZE : 1

可能这里添加点额外信息会好点