From 3d0283e59ca8254b60cd475fcf912a82979f6913 Mon Sep 17 00:00:00 2001 From: lujun Date: Mon, 3 Dec 2018 16:53:39 +0800 Subject: [PATCH] update print out,test=develop --- 01.fit_a_line/README.cn.md | 18 ++++++++---------- 01.fit_a_line/index.cn.html | 18 ++++++++---------- 01.fit_a_line/train.py | 20 +++++++++----------- 3 files changed, 25 insertions(+), 31 deletions(-) diff --git a/01.fit_a_line/README.cn.md b/01.fit_a_line/README.cn.md index 9324580..9067bda 100644 --- a/01.fit_a_line/README.cn.md +++ b/01.fit_a_line/README.cn.md @@ -155,7 +155,7 @@ use_cuda = False place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace() ``` -除此之外，还可以定义一个事件响应器来处理类似`打印训练进程`的事件： +除此之外，还可以通过画图，来展现`训练进程`： ```python # Plot data @@ -165,10 +165,6 @@ train_title = "Train cost" test_title = "Test cost" plot_cost = Ploter(train_title, test_title) - -def event_handler(title, loop_step, handler_val): - plot_cost.append(title, loop_step, handler_val) - plot_cost.plot() ``` ### 创建训练过程 @@ -216,11 +212,13 @@ def train_loop(main_program): feed=feeder.feed(data_train), fetch_list=[avg_loss]) if step % 10 == 0: # record a train cost every 10 batches - event_handler(train_title, step, avg_loss_value[0]) + plot_cost.append(train_title, step, avg_loss_value[0]) + plot_cost.plot() if step % 100 == 0: # record a test cost every 100 batches test_metics = train_test(train_program=main_program, feeder=feeder_test) - event_handler(test_title, step, test_metics[0]) + plot_cost.append(test_title, step, test_metics[0]) + plot_cost.plot() # If the accuracy is good enough, we can stop the training. if test_metics[0] < 10.0: return @@ -229,9 +227,9 @@ def train_loop(main_program): if math.isnan(float(avg_loss_value)): sys.exit("got NaN loss, training failed.") - if params_dirname is not None: - # We can save the trained parameters for the inferences later - fluid.io.save_inference_model(params_dirname, ['x'], + if params_dirname is not None: + # We can save the trained parameters for the inferences later + fluid.io.save_inference_model(params_dirname, ['x'], [y_predict], exe) ``` diff --git a/01.fit_a_line/index.cn.html b/01.fit_a_line/index.cn.html index 13585cb..99bc1d6 100644 --- a/01.fit_a_line/index.cn.html +++ b/01.fit_a_line/index.cn.html @@ -197,7 +197,7 @@ use_cuda = False place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace() ``` -除此之外，还可以定义一个事件响应器来处理类似`打印训练进程`的事件： +除此之外，还可以通过画图，来展现`训练进程`： ```python # Plot data @@ -207,10 +207,6 @@ train_title = "Train cost" test_title = "Test cost" plot_cost = Ploter(train_title, test_title) - -def event_handler(title, loop_step, handler_val): - plot_cost.append(title, loop_step, handler_val) - plot_cost.plot() ``` ### 创建训练过程 @@ -258,11 +254,13 @@ def train_loop(main_program): feed=feeder.feed(data_train), fetch_list=[avg_loss]) if step % 10 == 0: # record a train cost every 10 batches - event_handler(train_title, step, avg_loss_value[0]) + plot_cost.append(train_title, step, avg_loss_value[0]) + plot_cost.plot() if step % 100 == 0: # record a test cost every 100 batches test_metics = train_test(train_program=main_program, feeder=feeder_test) - event_handler(test_title, step, test_metics[0]) + plot_cost.append(test_title, step, test_metics[0]) + plot_cost.plot() # If the accuracy is good enough, we can stop the training. if test_metics[0] < 10.0: return @@ -271,9 +269,9 @@ def train_loop(main_program): if math.isnan(float(avg_loss_value)): sys.exit("got NaN loss, training failed.") - if params_dirname is not None: - # We can save the trained parameters for the inferences later - fluid.io.save_inference_model(params_dirname, ['x'], + if params_dirname is not None: + # We can save the trained parameters for the inferences later + fluid.io.save_inference_model(params_dirname, ['x'], [y_predict], exe) ``` diff --git a/01.fit_a_line/train.py b/01.fit_a_line/train.py index 9902b89..b487074 100644 --- a/01.fit_a_line/train.py +++ b/01.fit_a_line/train.py @@ -21,11 +21,6 @@ import math import sys -# event_handler prints training and testing info -def event_handler(title, loop_step, handler_val): - print("%s, Step %d, Cost %f" % (title, loop_step, handler_val)) - - def main(): batch_size = 20 @@ -93,10 +88,13 @@ def main(): feed=feeder.feed(data_train), fetch_list=[avg_loss]) if step % 10 == 0: # record a train cost every 10 batches - event_handler(train_title, step, avg_loss_value[0]) + print("%s, Step %d, Cost %f" % + (train_title, step, avg_loss_value[0])) + if step % 100 == 0: # record a test cost every 100 batches test_metics = train_test( program=test_program, feeder=feeder_test) - event_handler(test_title, step, test_metics[0]) + print("%s, Step %d, Cost %f" % + (test_title, step, test_metics[0])) # If the accuracy is good enough, we can stop the training. if test_metics[0] < 10.0: return @@ -105,10 +103,10 @@ def main(): if math.isnan(float(avg_loss_value)): sys.exit("got NaN loss, training failed.") - if params_dirname is not None: - # We can save the trained parameters for the inferences later - fluid.io.save_inference_model(params_dirname, ['x'], [y_predict], - exe) + if params_dirname is not None: + # We can save the trained parameters for the inferences later + fluid.io.save_inference_model(params_dirname, ['x'], + [y_predict], exe) train_loop() -- GitLab