TensorFlow Keras图执行错误求助：遵循教程训练仍报错

我知道已经有类似问题被提出，但我严格按照教程操作还是出现错误，查遍了所有可能的解决方案仍不知道怎么修正。我初步判断问题和形状或标签有关，但找不到解决办法，求帮忙。

我的代码 #

import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
from tensorflow.keras.models import Sequential 
from tensorflow.keras.layers import Dense, Input, Activation
from tensorflow.keras.datasets import boston_housing
from tensorflow.keras import layers

SEED_VALUE = 65

# 固定随机种子保证结果可复现
np.random.seed(SEED_VALUE)
tf.random.set_seed(SEED_VALUE)

# 加载波士顿房价数据集
(X_train, y_train), (X_test, y_test ) = boston_housing.load_data()
print(X_train.shape)
print("\n")
print("输入特征样例: ", X_train[0])
print("\n")
print("标签样例: ", y_train[0])

boston_features = {
'Average Number of Rooms': 5,
}

# 提取单特征（平均房间数）
X_train_1d = X_train[:, boston_features['Average Number of Rooms']]
print(X_train_1d.shape)

X_test_1d = X_test[:, boston_features['Average Number of Rooms']]

# 绘制特征与标签的散点图
plt.figure(figsize=(15,5))
plt.xlabel('平均房间数')
plt.ylabel('房价中位数 [$K]')
plt.grid("on")
plt.scatter(X_train_1d[:], y_train, color='green', alpha=0.5);

# 构建单神经元模型
model = Sequential()
model.add(Dense(units=1, input_shape=(1,)))

# 打印模型结构
model.summary()

# 编译模型
model.compile(optimizer=tf.keras.optimizers.RMSprop(learning_rate=.005), loss='mse')

# 训练模型
history = model.fit(X_train_1d, y_train, batch_size=16, epochs=101, validation_split=0.3)

运行错误信息 #

Epoch 1/101
2023-01-21 12:03:45.701983: I        tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:114] Plugin optimizer for device_type GPU is enabled.
2023-01-21 12:03:45.757923: W tensorflow/core/framework/op_kernel.cc:1830] OP_REQUIRES failed     at xla_ops.cc:418 : NOT_FOUND: could not find registered platform with id: 0x281ae11b0
2023-01-21 12:03:45.757952: W tensorflow/core/framework/op_kernel.cc:1830] OP_REQUIRES failed     at xla_ops.cc:418 : NOT_FOUND: could not find registered platform with id: 0x281ae11b0
---------------------------------------------------------------------------
NotFoundError                             Traceback (most recent call last)
Cell In[28], line 1
----> 1 history = model.fit(X_train_1d, y_train, batch_size=16, epochs=101,      validation_split=0.3)

File ~/miniconda3/envs/tensorflow/lib/python3.10/site-packages/keras/utils/traceback_utils.py:70, in filter_traceback.<locals>.error_handler(*args, **kwargs)
     67     filtered_tb = _process_traceback_frames(e.__traceback__)
     68     # To get the full stack trace, call:
     69     # `tf.debugging.disable_traceback_filtering()`
---> 70     raise e.with_traceback(filtered_tb) from None
     71 finally:
     72     del filtered_tb

File ~/miniconda3/envs/tensorflow/lib/python3.10/site-packages/tensorflow/python/eager/execute.py:52, in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
     50 try:
     51   ctx.ensure_initialized()
---> 52   tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
     53                                       inputs, attrs, num_outputs)
     54 except core._NotOkStatusException as e:
     55   if name is not None:

NotFoundError: Graph execution error:

这个错误和数据形状、标签无关，是TensorFlow的GPU适配问题，可通过以下方式解决：

• 强制用CPU运行：在代码开头添加以下代码，禁用GPU，让TensorFlow使用CPU计算：

import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"

• 检查版本兼容性：当前TensorFlow版本和系统的CUDA、cuDNN版本不匹配，导致XLA加速模块找不到GPU平台。对照TensorFlow官方版本兼容表，安装对应版本的CUDA和cuDNN。

• 关闭XLA优化：添加配置禁用XLA，避免相关错误：

tf.config.optimizer.set_jit(False)

• 升级TensorFlow：部分旧版本存在GPU平台注册bug，升级到最新稳定版可修复：

pip install --upgrade tensorflow

内容的提问来源于stack exchange，提问作者Andres Beregovich