tensorflow代码分析
- tensorflow/python/
- client
- data
- kernel_tests
- ops
- dataset_ops.py
- iterator_ops.py
- readers.py
- util
- debug
tensorflow的调试工具TFBDG。 - eager
- estimator
高阶API，可以大幅度的减少我们构建模型的代码量。- canned
- init.py
- baseline.py
- baseline_test.py
- dnn.py
- dnn_linear_combined.py
- dnn_linear_combined_test.py
- dnn_test.py
- dnn_testing_utils.py
- head.py
- head_test.py
- linear.py
- linear_test.py
- linear_testing_utils.py
- metric_keys.py
- optimizers.py
- optimizers_test.py
- parsing_utils.py
- parsing_utils_test.py
- prediction_keys.py
- export
- inputs
- canned
- feature_column
- framework
- grappler
- keras
高阶API，可以大幅度的减少我们构建模型的代码量。 Keras API的纯tensorflow实现都封装在这个包中了。- _impl/keras
- activations
- applications
- backend
- callbacks
- constraints
- datasets
- estimator
- initializers
- layers
- losses
- metrics
- models
- optimizers
- preprocessing
- regularizers
- utils
- wrappers
- kernel_tests
- layers
- lib
- ops
- linalg
- losses
- accumulate_n_benchmark.py
- array_grad.py
- array_ops.py
- batch_norm_benchmark.py
- bitwise_ops.py
- bitwise_ops_test.py
- candidate_sampling_ops.py
- check_ops.py
- clip_ops.py
- clip_ops_test.py
- concat_benchmark.py
- confusion_matrix.py
- control_flow_grad.py
- control_flow_ops.py
- control_flow_ops_test.py
- control_flow_util.py
- conv2d_benchmark.py
- ctc_ops.py
- data_flow_grad.py
- data_flow_ops.py
- dequantize_op_test.py
- embedding_ops.py
- functional_ops.py
- gradient_checker.py
- gradient_checker_test.py
- gradients.py
- gradients_impl.py
- gradients_test.py
- hidden_ops.txt
- histogram_ops.py
- histogram_ops_test.py
- image_grad.py
- image_grad_test.py
- image_ops.py
- image_ops_impl.py
- image_ops_test.py
- init_ops.py
- initializers_ns.py
- io_ops.py
- linalg_grad.py
- linalg_ops.py
- logging_ops.py
- lookup_ops.py
- math_grad.py
- math_grad_test.py
- math_ops.py
低等数学
tf.scalar_mul(scalar,x) 求x的scalar倍
tf.abs(x,name=None) 求x的绝对值
tf.negative(x,name=None) 求x的负数
tf.sign(x,name=None) 求x的符号
tf.reciprocal(x,name=None) 求x的倒数
tf.square(x,name=None) 求x的平方
tf.round(x,name=None) 求离x最近的整数，若有两值，取偶数。
tf.sqrt(x,name=None) 求x的平方根
tf.rsqrt(x,name=None) 求（x的平方根）的倒数
tf.exp(x,name=None) 求e的x次幂
tf.expm1(x,name=None) 求（e的x次幂）减1
tf.log(x,name=None) 求x的自然对数
tf.log1p(x,name=None) 求x加1的自然对数
tf.ceil(x,name=None) 求比x大的最小整数
tf.floor(x,name=None)求比x小的最大整数
tf.cos(x,name=None)求cos(x)
tf.sin(x,name=None)求sin(x)
tf.lbeta(x,name=None)求ln(|Beta(x)|)
tf.tan(x,name=None) 求tan
tf.acos(x,name=None) 求acos
tf.asin(x,name=None) 求asin
tf.atan(x,name=None) 求atan
tf.lgamma(x,name=None)求ln(gamma(x))
tf.digamma(x,name=None)求lgamma的导数
tf.erf(x,name=None) 计算高斯误差
tf.erfc(x,name=None) 计算1-高斯误差
tf.rint(x,name=None) 计算离x最近的整数，若为中间值，取偶数值。 - math_ops_test.py
- matmul_benchmark.py
- matmul_benchmark_test.py
- metrics.py
- metrics_impl.py
- nn.py
- nn_batchnorm_test.py
- nn_fused_batchnorm_test.py
- nn_grad.py
- nn_grad_test.py
- nn_impl.py
- nn_ops.py
- nn_test.py
- nn_xent_test.py
- numerics.py
- parsing_ops.py
- partitioned_variables.py
- quantized_conv_ops_test.py
- random_ops.py
- resource_variable_ops.py
- resources.py
- rnn.py
- rnn_cell.py
- rnn_cell_impl.py
- script_ops.py
- sdca_ops.py
- session_ops.py
- sets.py
- sets_impl.py
- sparse_grad.py
- sparse_ops.py
- special_math_ops.py
- special_math_ops_test.py
- spectral_grad.py
- spectral_ops.py
- spectral_ops_test_util.py
- split_benchmark.py
- standard_ops.py
- state_grad.py
- state_ops.py
- string_ops.py
- summary_op_util.py
- summary_ops.py
- template.py
- tensor_array_grad.py
- tensor_array_ops.py
- transpose_benchmark.py
- variable_scope.py
- variables.py
- weights_broadcast_ops.py
- platform
- profiler
- saved_model
- summary
- tools
- training
- user_ops
sennchi
