Question

我正在创建一个CNN鉴别器模型来验证gan的音频。它取自gan的发生器部分的输出张量。但由于某些原因，当使用带有来自生成器的音频张量的编译模型时，我得到了这个甚至没有指定行号的奇怪错误。

代码：

<header class="primary-header">

  <h5 class="logo-name">Roger Anderson</h5>

  <nav class="nav">
    <ul>
      <li><a href="about.html">about</a></li>
      <!--
              -->
      <li><a href="portfolio.html">portfolio</a></li>
      <!--
              -->
      <li><a href="contact.html">contact</a></li>
    </ul>
  </nav>

</header>

错误：

def build_audio_discriminator(audio_shape, num_classes):

    model = Sequential()

    model.add(Conv1D(32, kernel_size=(2), padding="same", input_shape=audio_shape))
    model.add(MaxPooling1D(pool_size=(2)))
    model.add(Dropout(0.25))
    model.add(Dense(128, activation='relu'))
    model.add(Dropout(0.25))
    model.add(Dense(128))

    model.summary()

    audio_shape_ = (None, audio_shape[1])
    audio = Input(shape=audio_shape_)

    # Extract feature representation
    features = model(audio)

    # Determine validity and label of the image
    validity = Dense(1, activation="sigmoid")(features)
    label = Dense(num_classes+1, activation="softmax")(features)

    return Model(audio, [validity, label])


# Build and compile the discriminator
#audio_shape: (31, 214161), num_classes: 1
audio_discriminator = build_audio_discriminator(audio_shape, num_classes)
audio_discriminator.compile(loss=losses, optimizer=optimizer, metrics=['accuracy'])

# audio: Tensor("model_4/sequential_4/activation_4/Softmax:0", shape=(?, 214161), dtype=float32)
audio_valid, audio_target_label = audio_discriminator(audio)

Answer 1

Solution was to reshape input Tensor to proper Conv1D input shape: (-1, 214161, 1)

def post_list(request, tag_slug=None):
    object_list = Post.published.all()
    tag = None
    if tag_slug:
        tag = get_object_or_404(Tag, slug=tag_slug)
        object_list = object_list.filter(tags__in=[tag])
    paginator = Paginator(object_list, 3) # 3 posts in each page
    page = request.GET.get('page')
    try:
         posts = paginator.page(page)
    except PageNotAnInteger:
    # If page is not an integer deliver the first page
        posts = paginator.page(1)
    except EmptyPage:
    # If page is out of range deliver last page of results
        posts = paginator.page(paginator.num_pages)
    return render(request, 'blog.html', {'page': page, 'posts': posts, 'tag': tag})

The solution shown above lets you run into a problem though, because the object returned by tf.reshape is not a Keras tensor but a standard tensor without _keras_history flag. My recommendation is to add:

key << { myDriver.myMap.keySet() }()

instead the solution above to your model.

Keras模型溢出未知错误| （IndexError：列表索引超出范围）

1 个答案: