`auto_convert_mixed_precision` Error: two nodes with same node name error occurred during

[KunMengcode]

I want to convert a model into an AMP model. This is my code:

def convert_float32_to_mixed_precision(fp32_model_path, mixed_precision_model_path):
    from onnxconverter_common import auto_mixed_precision
    import onnx

    model = onnx.load(fp32_model_path)

    import numpy as np
    np.random.seed(123)
    test_data = {"image": 2*np.random.rand(1, 3, 640, 640).astype(np.float32)-1.0,"scale_factor":[[1,1]]}

    def validate(res1, res2):
        for r1, r2 in zip(res1, res2):
            if not np.allclose(r1, r2, rtol=0.01, atol=0.001):
                return False
        return True

    model_fp16 = auto_mixed_precision.auto_convert_mixed_precision(model, test_data, validate, keep_io_types=True)
    onnx.save(model_fp16, mixed_precision_model_path)

fp32_model_path = 'F32.onnx'
mixed_precision_model_path = 'AMP.onnx'

print("Convert to mixed precision starts...")
convert_float32_to_mixed_precision(fp32_model_path, mixed_precision_model_path)
print("Conversion finished.")

OR

from onnxconverter_common import auto_mixed_precision
import onnx
import numpy as np

test_data = {"image": 2*np.random.rand(1, 3, 640, 640).astype(np.float32)-1.0,"scale_factor":[[1,1]]}
model = onnx.load("float32.onnx")
model_fp16 = auto_mixed_precision.auto_convert_mixed_precision(model, test_data, rtol=0.01, atol=0.001, keep_io_types=True)
onnx.save(model_fp16, "AMP-float32.onnx")

Using my model to run this code resulted in an error:1 : FAIL : This is an invalid model. Error: two nodes with same node name (_output_cast0). But I don't have duplicate node names. This is my model: Google Drive

[xiaowuhu]

@KunMengcode Can I have you model to repro?

[KunMengcode]

@xiaowuhu I am using the ONNX F32 model exported from the PPYOLOE model in the PaddleDetection library. I think it should be equivalent to this？PPYOLOE_pytorch (structurally) And [Baidu Yun Link]：https://pan.baidu.com/s/1BYw-tC3wqA5zpV90XeYvFQ?pwd=el24 [code]：el24

[itapty-ily]

Did you fix this issue. I am fixing this problem

[KunMengcode]

Did you fix this issue. I am fixing this problem

I haven't solved this problem yet

[KunMengcode]

Did you fix this issue. I am fixing this problem

What is your situation @itapty-ily