-
Notifications
You must be signed in to change notification settings - Fork 62
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
加载预编译模型失败 #56
Comments
生成出来的预编译模型是不是只有几kb,这样的预编译模型是有问题的。执行 |
这个模型不是只有几KB的情况,大小是正常的。原始rknn模型大小为6.9MB,生成的预编译模型大小为5.7MB。 执行export RKNN_LOG_LEVEL=5后,预编译生成的日志如下: D [set_chip_platform_env:229]evaluate model on RV1126 D [setup_norm_tensor:5042]tensor id=1, name=Sigmoid_Sigmoid_203/out0_1, opt = 0, attr.opt =0 D [setup_norm_tensor:5042]tensor id=2, name=Sigmoid_Sigmoid_205/out0_2, opt = 0, attr.opt =0 D [setup_norm_tensor:5042]tensor id=3, name=images_205, opt = 0, attr.opt =0 I [vnn_CreateRKNN:6283]Const tensor initialize D [set_chip_platform_env:229]evaluate model on RV1126 D [setup_norm_tensor:5042]tensor id=1, name=Sigmoid_Sigmoid_203/out0_1, opt = 0, attr.opt =0 D [setup_norm_tensor:5042]tensor id=2, name=Sigmoid_Sigmoid_205/out0_2, opt = 0, attr.opt =0 D [setup_norm_tensor:5042]tensor id=3, name=images_205, opt = 0, attr.opt =0 I [vnn_CreateRKNN:6283]Const tensor initialize D [rknn_init:1985] Input Tensors: 使用该模型进行推理,显示信息如下: |
从预编译日志看不到什么问题。麻烦提供rknn模型以便复现。 |
请问解决了嘛?我也遇到了同样的问题。 |
在rv1126板子上(rknn runtime版本为1.7.5),通过C接口调用生成预编译模型并保存到文件中。
然后在该环境中加载上述预编译模型,返回错误码-6,请问可以如何解决?
捞取的信息如下:
Loading rknn_pre_compile.rknn...
librknn_runtime version 1.7.5 (ccb93f1 build: 2023-07-18 10:56:29 base: 1131)
E [parseModel:145]Invalid RKNN model:1313689686 vs 65550
rknn_init error ret=-6
ExtractorRknn fail
The text was updated successfully, but these errors were encountered: