Question

我正在尝试从https://github.com/raghakot/keras-vis/blob/master/examples/mnist/attention.ipynb复制结果以产生显着性为零的图。我输入了完全相同的代码，并按建议使用了：

from vis.visualization import visualize_saliency
from vis.utils import utils
from keras import activations

# Utility to search for layer index by name. 
# Alternatively we can specify this as -1 since it corresponds to the last layer.
layer_idx = utils.find_layer_idx(model, 'preds')

# Swap softmax with linear
model.layers[layer_idx].activation = activations.linear
model = utils.apply_modifications(model)

grads = visualize_saliency(model, layer_idx, filter_indices=class_idx,seed_input=x_test[idx])
# Plot with 'jet' colormap to visualize as a heatmap.
plt.imshow(grads, cmap='jet')

但是，我不断收到以下错误：

InvalidArgumentError: conv2d_1_input_3:0 is both fed and fetched.

我在其他地方看过，并且看到了升级keras-vis的建议，虽然这样做了，但还是出现了相同的错误。错误似乎在

中

grads = visualize_saliency(model, layer_idx, filter_indices=class_idx,seed_input=x_test[idx])

当我注释掉这一行时，没有错误显示。

我该如何解决？

Answer 1

已解决！如果有人遇到此问题，请使用： pip安装git + git：//github.com/raghakot/keras-vis.git --upgrade --no-deps

keras-vis可视化显着性问题

1 个答案: