rockchip-linux / rknn-toolkit

BSD 3-Clause "New" or "Revised" License
830 stars 175 forks source link

Unsuccessful Conversion of Yolo-V10S FP32-TFlite to RKNN #466

Open aayushb-95 opened 3 months ago

aayushb-95 commented 3 months ago

Hello Everybody, I am trying to convert the Yolo-V10S FP32-TFlite model to RKNN. However, I am encountering the following error:

/home/aayush/miniconda3/envs/rknn_toolkit_1.7.5/bin/python 
/home/aayush/yolo_create_rknn.py 
W rknn-toolkit version: 1.7.5
--> Config model
--> Loading model
TFLite Model Loaded /home/aayush/v10s.tflite
--> Building model
E Catch exception when loading tflite model: /home/aayush/v10s.tflite!
E Traceback (most recent call last):
E   File "rknn/base/RKNNlib/converter/tflite_loader.py", line 573, in rknn.base.RKNNlib.converter.tflite_loader.ModelParser.parse
E   File "rknn/base/RKNNlib/converter/tflite_loader.py", line 503, in rknn.base.RKNNlib.converter.tflite_loader.ModelParser._build_layers
E   File "rknn/base/RKNNlib/converter/tflite_loader.py", line 338, in rknn.base.RKNNlib.converter.tflite_loader.ModelParser._get_layer_RKNN_op
E   File "rknn/base/RKNNlib/converter/lite/RKNN_map_op.py", line 844, in rknn.base.RKNNlib.converter.lite.RKNN_map_op.RKNNMapOp.map_RKNN_ops
E   File "rknn/base/RKNNlib/converter/lite/RKNN_map_op.py", line 811, in rknn.base.RKNNlib.converter.lite.RKNN_map_op._cast
E AttributeError: 'NoneType' object has no attribute 'InDataType'
E Please feedback the detailed log file <log_feedback_to_the_rknn_toolkit_dev_team.log> to the RKNN Toolkit development team.
E You can also check github issues: https://github.com/rockchip-linux/rknn-toolkit/issues
Traceback (most recent call last):
  File "/home/aayush/yolo_create_rknn.py", line 49, in <module>
    ret = rknn.build(do_quantization=True, dataset='/home/aayush/img_list.txt', pre_compile=False, rknn_batch_size=1)
  File "/home/aayush/miniconda3/envs/rknn_toolkit_1.7.5/lib/python3.8/site-packages/rknn/api/rknn.py", line 272, in build
    inputs = self.rknn_base.net.get_input_layers()
AttributeError: 'NoneType' object has no attribute 'get_input_layers'

This error generally occurs when the RKNN cannot read the input layer of the model. However, the FP32 TFLite model is working absolutely fine.