Input to DecodeRaw is not a multiple of 8, the size of double

最新推荐文章于 2023-11-13 00:00:52 发布

原创最新推荐文章于 2023-11-13 00:00:52 发布 · 454 阅读

1 ·

CC 4.0 BY-SA版权

文章标签：

#python #tensorflow #深度学习

tf 专栏收录该内容

40 篇文章

订阅专栏

博客内容涉及在使用TensorFlow时遇到的`InvalidArgumentError`，具体错误为输入到`DecodeRaw`的长度不是4的倍数。解决方案是检查`decode_raw`的类型转换，建议将数据类型改为`tf.uint8`，如果需要转换为`tf.float64`，可以先用`tf.cast`进行转换。这个错误通常出现在处理图像或音频数据时，数据预处理不正确导致。

部署运行你感兴趣的模型镜像

tensorflow.python.framework.errors_impl.InvalidArgumentError: Input to DecodeRaw has length 1 that is not a multiple of 4, the size of float
[[node DecodeRaw_1 (defined at /GraphRelated/tf_repos/keras_estimator/aaaa.py:63) ]]

解决方案一：

label = tf.decode_raw(features['label'],tf.uint8)

Because you use tf.decode_raw to convert your image to double(tf.float64) type, which’s size is 8 bytes. so the parsed[‘mel_spec_raw’] should be a multiple of 8. You can print the type of parsed[‘mel_spec_raw’], it should be tf.string, which explain why the size of the parsed[‘mel_spec_raw’] is 165750. Your can change your code to:

# mel_spec1d = tf.decode_raw(parsed['mel_spec_raw'], tf.float64)

mel_spec1d = tf.decode_raw(parsed[‘mel_spec_raw’], tf.uint8)
it might work, because of the size of tf.uint8 is only 1. If you want to convert the type into tf.float64, you can convert the type into tf.float using tf.cast

mel_spec1d = tf.cast(mel_spec1d, tf.float64)

Hope this will help you.

Input to DecodeRaw is not a multiple of 8, the size of double

您可能感兴趣的与本文相关的镜像

Seed-Coder-8B-Base

文本生成

Seed-Coder

Seed-Coder是一个功能强大、透明、参数高效的 8B 级开源代码模型系列，包括基础变体、指导变体和推理变体，由字节团队开源