我使用此网站上的预训练模型：https://github.com/tensorflow/models/tree/master/research/slim/nets/mobilenet 我下载了v3-large_224_1.0_float.pb模型链接是（Large dm = 1（float））。运行它之后，我经常会得到这个结果， [

===== TENSORFLOW RESULTS =======
cello,violoncello (score = 1.00000)
hip,rose hip,rosehip (score = 0.00000)
banjo (score = 0.00000)
steel drum (score = 0.00000)
marimba,xylophone (score = 0.00000)

] 1 然后，我更改输入图像，但结果是相同的。我确实知道如何解决这个问题。谢谢。 我的代码如下：

# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements.  See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership.  The ASF licenses this file
# to you under the Apache License,Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License.  You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,# software distributed under the License is distributed on an
# "AS IS" BASIS,WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND,either express or implied.  See the License for the
# specific language governing permissions and limitations
# under the License.
"""
Compile Tensorflow Models
=========================
This article is an introductory tutorial to deploy tensorflow models with tvm.

For us to begin with,tensorflow python module is required to be installed.

Please refer to https://www.tensorflow.org/install
"""

# tvm,relay
import tvm
from tvm import te
from tvm import relay

# os and numpy
import numpy as np
import os.path

# Tensorflow imports
import tensorflow as tf
try:
    tf_compat_v1 = tf.compat.v1
except ImportError:
    tf_compat_v1 = tf

# Tensorflow utility functions
import tvm.relay.testing.tf as tf_testing

# Target settings
# Use these commented settings to build for cuda.
#target = 'cuda'
#target_host = 'llvm'
#layout = "NCHW"
#ctx = tvm.gpu(0)
target = 'llvm'
target_host = 'llvm'
layout = None
ctx = tvm.cpu(0)

pwd = os.path.dirname(os.path.realpath(__file__))
img_url=os.path.join(pwd,"../data")
model_path = os.path.join(pwd,'model/v3-large_224_1.0_float.pb')
map_proto_path = os.path.join(pwd,'../data/imagenet_2012_challenge_label_map_proto.pbtxt')
label_path = os.path.join(pwd,'../data/imagenet_synset_to_human_label_map.txt')

# preprocess one picture
def preprocess(img_path):
    jpg = open(img_path,'rb')
    image = tf.image.decode_jpeg(jpg.read(),channels = 3)

    image = tf.image.central_crop(image,0.875)
    image = tf.expand_dims(image,0)
    image = tf.image.resize_bilinear(image,[224,224],align_corners=False)
    image = tf.subtract(image,0.5)
    image = tf.multiply(image,2.0)
    return image

# prepare input data
def prepare_input(num):
    x_all = []
    for i in range(num):
        #load jpg
        index = i % 67
        img_path = os.path.join(
            img_url,'imagenet_images/val/ILSVRC2012_val_000000%02d.JPEG' % (index+1))
        x = preprocess(img_path)
        x_all.append(x)

    tx = tf.concat(x_all,0)
    with tf.Session() as session:
        x = tx.eval()
    return x

# set number pictures to predict
BATCH = 1
print("Test mobilenet-v3 model with batch " + str(BATCH))
# prepare input data
x = prepare_input(BATCH)

######################################################################
# Inference on tensorflow
# -----------------------
# Run the corresponding model on tensorflow

def create_graph():
    """Creates a graph from saved GraphDef file and returns a saver."""
    # Creates graph from saved graph_def.pb.
    with tf_compat_v1.gfile.GFile(model_path,'rb') as f:
        graph_def = tf_compat_v1.GraphDef()
        graph_def.ParseFromString(f.read())
        graph = tf.import_graph_def(graph_def,name='')
        # Call the utility to import the graph definition into default graph.
        graph_def = tf_testing.ProcessGraphDefParam(graph_def)

def run_inference_on_image(image):
    """Runs inference on an image.

    Parameters
    ----------
    image: String
        Image file name.

    Returns
    -------
        Nothing
    """
    # Creates graph from saved GraphDef.
    create_graph()

    with tf_compat_v1.Session() as sess:
        softmax_tensor = sess.graph.get_tensor_by_name('MobilenetV3/Predictions/Softmax:0')
        predictions = sess.run(softmax_tensor,{'input:0': image})

        predictions = np.squeeze(predictions)

        # Creates node ID --> English string lookup.
        node_lookup = tf_testing.NodeLookup(label_lookup_path=map_proto_path,uid_lookup_path=label_path)

        # Print top 5 predictions from tensorflow.
        top_k = predictions.argsort()[-5:][::-1]
        print ("===== TENSORFLOW RESULTS =======")
        for node_id in top_k:
            human_string = node_lookup.id_to_string(node_id)
            score = predictions[node_id]
            print('%s (score = %.5f)' % (human_string,score))

run_inference_on_image(x)