Meet unsupported input_2_t dtype for matmul

[Tracersboy]

rknn tool kit 1.5.2
rknn_server 1.5.2
我存在两个onnx，名为1steg2_new.onnx和1steg2_old.onnx。两个onnx除了权重都一样，1steg2_new.onnx在rknntoolkit在模拟器上（rknn.init_runtime()）推理正常，但使用adb连接（rknn.init_runtime(target="rk3568", device_id="df8e2b3164dd6e6a")），就会报错toolkit端报错如下。1steg2_old.onnx均正常。

rknn_server端的报错如下

toolkit端的使用的代码如下

# -*- coding:utf-8 -*-
import os
import urllib
import traceback
import time
import sys
import numpy as np
import cv2
from rknn.api import RKNN
from PIL import Image

if __name__ == '__main__':
    rknn = RKNN(verbose=True)
    rknn.list_devices()

    print('--> Config model')
    rknn.config(target_platform="rk3568")


    # Load ONNX model
    print('--> Loading model')
    ret = rknn.load_onnx(model="1steg2_new.onnx")
    if ret != 0:
        print('Load model failed!')
        exit(ret)
    print('done')

    print('--> Building model')
    ret = rknn.build(do_quantization=False)

    if ret != 0:
        print('Build model failed!')
        exit(ret)
    print('done')
    # Init runtime environment
    print('--> Init runtime environment')
    # ret = rknn.init_runtime()
    ret = rknn.init_runtime(target="rk3568", device_id="df8e2b3164dd6e6a")
    if ret != 0:
        print('Init runtime environment failed!')
        exit(ret)
    print('done')

    outputs = rknn.inference(inputs=[np.zeros((1, 16, 512)).astype(np.float32)], data_format="nchw")  #
    print(outputs[0].shape)
    print(outputs[0])

两个onnx的地址为链接：https://pan.baidu.com/s/1uiE_XkMCSox2yzVJYby9zw
提取码：z4d8

望开发者能尽快定位问题，辛苦了