使用pnnx转换onnx的模型，调用ncnn::Net::load_param(AAssetManager*, char const*)崩溃

[BandageWorm]

error log | 日志或报错信息 | ログ

10:25:27.067 29257-29257 DEBUG A signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x000000000000001c
10:25:27.067 29257-29257 DEBUG A Cause: null pointer dereference
10:25:27.067 29257-29257 DEBUG A x0 0000007b96459038 x1 b400007b0b1ef800 x2 0000000000000020 x3 0000000000000016
10:25:27.067 29257-29257 DEBUG A x4 0000007b1598e820 x5 0000007b1598e840 x6 72686664711f716e x7 7f7f7f7f7f7f7f7f
10:25:27.067 29257-29257 DEBUG A x8 b400007b0b228700 x9 b400007b1597b700 x10 0000000000000320 x11 0000000000000001
10:25:27.067 29257-29257 DEBUG A x12 0000000000000000 x13 0000000000000001 x14 0000000004000000 x15 0000000000000028
10:25:27.067 29257-29257 DEBUG A x16 0000007c55b82a10 x17 0000007c55b72bc0 x18 0000007b13696000 x19 b400007b157d0300
10:25:27.067 29257-29257 DEBUG A x20 0000007b0815121e x21 0000007b0813de25 x22 0000000000000000 x23 0000007b14727000
10:25:27.067 29257-29257 DEBUG A x24 000000000000000d x25 0000000100000000 x26 0000007c55b85c20 x27 0000000000000001
10:25:27.067 29257-29257 DEBUG A x28 0000007b08117f99 x29 0000007b14722e10
10:25:27.067 29257-29257 DEBUG A lr 0000007b08460644 sp 0000007b14722d80 pc 0000007b08460670 pst 0000000080001000
10:25:27.067 29257-29257 DEBUG A 80 total frames
10:25:27.067 29257-29257 DEBUG A backtrace:
10:25:27.067 29257-29257 DEBUG A #00 pc 000000000042b670 /data/app/~~P7TyRA4ppV0KgDKwUpYJkQ==/com.kurt.imgtester-xg_hIUcId6kv5NjKNCmGQg==/base.apk!librvmncnn.so (offset 0x4890000) (BuildId: d27f3c6c1a2780b7b6bfa3dd3546b1e106027187)
10:25:27.067 29257-29257 DEBUG A #1 pc 000000000042d60c /data/app/~~P7TyRA4ppV0KgDKwUpYJkQ==/com.kurt.imgtester-xg_hIUcId6kv5NjKNCmGQg==/base.apk!librvmncnn.so (offset 0x4890000) (BuildId: d27f3c6c1a2780b7b6bfa3dd3546b1e106027187)
10:25:27.067 29257-29257 DEBUG A #2 pc 000000000042fa7c /data/app/~~P7TyRA4ppV0KgDKwUpYJkQ==/com.kurt.imgtester-xg_hIUcId6kv5NjKNCmGQg==/base.apk!librvmncnn.so (offset 0x4890000) (BuildId: d27f3c6c1a2780b7b6bfa3dd3546b1e106027187)
10:25:27.067 29257-29257 DEBUG A #3 pc 00000000003e04e4 /data/app/~~P7TyRA4ppV0KgDKwUpYJkQ==/com.kurt.imgtester-xg_hIUcId6kv5NjKNCmGQg==/base.apk!librvmncnn.so (offset 0x4890000) (SEG::load(AAssetManager*, char const*, char const*)+56) (BuildId: d27f3c6c1a2780b7b6bfa3dd3546b1e106027187)
10:25:27.067 29257-29257 DEBUG A #4 pc 00000000003d9664 /data/app/~~P7TyRA4ppV0KgDKwUpYJkQ==/com.kurt.imgtester-xg_hIUcId6kv5NjKNCmGQg==/base.apk!librvmncnn.so (offset 0x4890000) (Java_com_tencent_rvmncnn_RVMNcnn_loadModel2+228) (BuildId: d27f3c6c1a2780b7b6bfa3dd3546b1e106027187)

addr2line -Cfe librvmncnn.so 42fa7c
ncnn::Net::load_param(AAssetManager*, char const*)
net.cpp:?
addr2line -Cfe librvmncnn.so 42d60c
ncnn::Net::load_param(ncnn::DataReader const&)
net.cpp:?
addr2line -Cfe librvmncnn.so 42b670
ncnn::Net::clear()
net.cpp:?

model | 模型 | モデル

1. original model

param文件：
deeplabv3p_mv2.ncnn.txt

how to reproduce | 复现步骤 | 再現方法

1. 下载https://github.com/nolanliou/mobile-deeplab-v3-plus 中的MobilenetV2 256x256 pb模型
2. 转换为onnx模型：python -m tf2onnx.convert --input .\deeplab_v3_plus_mnv2_aspp_decoder_256.pb --inputs Input:0 --outputs Output:0 --output deeplabv3p_mv2.onnx --opset 11
3. pnnx转化ncnn模型：pnnx .\deeplabv3p_mv2.onnx inputshape=[1,256,256,3]f32
4. Android端加载ncnn.param出现如上错误