rknn-toolkit
rknn-toolkit copied to clipboard
rockchip-linux
W:tensorflow:From D:\Anaconda3\envs\rk36\lib\site-packages\rknn\api\rknn.py:278: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead. W:tensorflow:From D:\Anaconda3\envs\rk36\lib\site-packages\rknn\api\rknn.py:278: The name tf.image.resize_nearest_neighbor is deprecated. Please use tf.compat.v1.image.resize_nearest_neighbor instead. 2024-07-04 14:26:53.726182: E tensorflow/core/lib/jpeg/jpeg_mem.cc:323] Premature end of JPEG...
I am trying to convert the TFLite model to RKNN model and I am getting the following error. ``` W rknn-toolkit version: 1.7.5 W Channels(512) of input node: input_0 >...
Hi, I am trying to convert YoloV8/V9 ONNX model to RKNN. However, I am getting the following error : ``` W rknn-toolkit version: 1.7.5 W Catch exception when optimize onnx...
开发板：rv1109 环境：1.7.0、1.7.3、1.75都试过， 1. 先模型分段，然后在线转预编译模型，报IOCTL_CMD_RKNN_PRECOMPILE fail, ack=0, data_len=0!，生成的预编译模型只有4 KB。  3. 不进行模型分段可以成功转预编译模型，人脸识别模型adaface转换失败但是其他模型可以转成功。 4. 连扳推理，分段前后模型输出均正常 想问下支持模型分段后再转预编译吗，这个IOCTL_CMD_RKNN_PRECOMPILE fail, ack=0, data_len=0有什么解决方法吗。
Cannot reshape a tensor with 96000 elements to shape [20,120] (2400 elements) for '{{node MatMul_p2o.MatMul.0_131_reshape_for_in1_312/MatMul_p2o.MatMul.0_131_reshape_for_in1_312}} = Reshape[T=DT_FLOAT, Tshape=DT_INT32](Add_p2o.Add.234_134_2/cond/Identity, MatMul_p2o.MatMul.0_131_reshape_for_in1_312/MatMul_p2o.MatMul.0_131_reshape_for_in1_312/shape)' with input shapes: [20,40,120], [2] and with input tensors computed as...
linaro@138V001000212263:~/Desktop/rknn-toolkit/rknn-toolkit-lite/examples/inference_with_lite$ python test.py --> Load RKNN model done --> Init runtime environment I RKNN: [13:32:31.065] RKNN Runtime Information: librknnrt version: 1.5.0 (e6fe0c678@2023-05-25T08:09:20) I RKNN: [13:32:31.067] RKNN Driver Information: version: 0.9.0...
Hi I made a single layer network: def export_pytorch_model(): net = Conv3d(3, 64, kernel_size=[2,3,3], stride=[1,2,2]) net.eval() trace_model = torch.jit.trace(net, torch.Tensor(1,3,2,180,360)) trace_model.save('./sqnet.pt') Then follow the rknn api as below: input_size_list =...
从onnx导出为rknn时提示Conv算子不支持，目前定位的问题是rknn无法加载weight为空的卷积算子。这个该如何解决。   [log_feedback_to_the_rknn_toolkit_dev_team.log](https://github.com/rockchip-linux/rknn-toolkit/files/15144557/log_feedback_to_the_rknn_toolkit_dev_team.log) [测试onnx精度.zip](https://github.com/rockchip-linux/rknn-toolkit/files/15144561/onnx.zip)
done I NPUTransfer: Starting NPU Transfer Client, Transfer version 2.1.0 (b5861e7@2020-11-23T11:50:36) E RKNNAPI: rknn_init, driver open fail! ret = -9(ERROR_PIPE)! E Catch exception when init runtime! E Traceback (most recent...
