标签：单机 Keras 多卡 checkpoint train test import model save

注意：此模式下不能用fit_generator() 方式训练

""" GPU test
"""
import os
import sys
os.system('pip install -i https://pypi.tuna.tsinghua.edu.cn/simple keras==2.3.1')
from tensorflow.keras import Sequential
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Input,Dense
from tensorflow.keras import layers
from tensorflow.keras.callbacks import ModelCheckpoint,EarlyStopping
import tensorflow as tf
from tensorflow import keras
import numpy as np
import pickle
import time

checkpoint_save_dir = "/models/embedding_recall/ckpt"
best_model_name = "best_model.hdf5"
if not os.path.exists(checkpoint_save_dir):
    os.makedirs(checkpoint_save_dir)

with open(r"/models/embedding_recall/resources/minist.pkl","rb") as fr:
    data = pickle.load(fr)
    
def make_or_restore_model():
    # Either restore the latest model, or create a fresh one
    # if there is no checkpoint available.
    checkpoints = [checkpoint_dir + "/" + name for name in os.listdir(checkpoint_dir)]
    if checkpoints:
        latest_checkpoint = max(checkpoints, key=os.path.getctime)
        print("Restoring from", latest_checkpoint)
        return keras.models.load_model(latest_checkpoint)
    print("Creating a new model")
    return get_compiled_model()

def get_compiled_model():
    inputs = Input(shape=(784,))
    inputs.shape
    inputs.dtype
    dense = Dense(64, activation="relu")
    x = dense(inputs)
    x = Dense(64, activation="relu")(x)
    outputs = Dense(10)(x)
    model = Model(inputs=inputs, outputs=outputs, name="my_model")
    model.compile(
    loss = keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    optimizer = keras.optimizers.RMSprop(),
    metrics = ["accuracy"],)
    return model

def make_or_restore_model(checkpoint_save_dir, model_name):
    # Either restore the latest model, or create a fresh one
    # if there is no checkpoint available.
    if checkpoint_save_dir:
        latest_checkpoint = os.path.join(checkpoint_save_dir, model_name)
        print("Restoring from", latest_checkpoint)
        return keras.models.load_model(latest_checkpoint)
    else:
        return None

def run_training(epochs):
    strategy = tf.distribute.MirroredStrategy()
    print("Number of devices:{}".format(strategy.num_replicas_in_sync))
    with strategy.scope():
        model = get_compiled_model()
    (x_train, y_train),(x_test, y_test) = data[0],data[1]
    x_train = x_train.reshape(60000, 784).astype("float32")/255
    x_test = x_test.reshape(10000, 784).astype("float32")/255
    
    early_stop = EarlyStopping(monitor='loss', patience=3, verbose=1)
    checkpoint = ModelCheckpoint(os.path.join(checkpoint_save_dir, best_model_name),
                             monitor='loss', verbose=1, save_best_only=True, mode='min')
    
    callbacks_list = [checkpoint, early_stop]
    t1 = time.time()
    history = model.fit(x_train, y_train, batch_size=100, epochs=epochs, callbacks=callbacks_list)
    t2 = time.time()
#     test_scores = model.evaluate(x_test, y_test, batch_size=100,verbose=2)
#     print("test loss:{}".format(test_scores[0]))
#     print("test acc:{}".format(test_scores[1]))
#     print("total spent:{}".format(t2-t1))

def continue_training(epochs):
    strategy = tf.distribute.MirroredStrategy()
    print("Number of devices:{}".format(strategy.num_replicas_in_sync))
#     with strategy.scope():
    model = make_or_restore_model(checkpoint_save_dir, best_model_name)
    (x_train, y_train),(x_test, y_test) = data[0],data[1]
    x_train = x_train.reshape(60000, 784).astype("float32")/255
    x_test = x_test.reshape(10000, 784).astype("float32")/255
    
    early_stop = EarlyStopping(monitor='loss', patience=3, verbose=1)
    checkpoint = ModelCheckpoint(os.path.join(checkpoint_save_dir, best_model_name),
                             monitor='loss', verbose=1, save_best_only=True, mode='min')
    
    callbacks_list = [checkpoint, early_stop]
    t1 = time.time()
    history = model.fit(x_train, y_train, batch_size=100, epochs=epochs, callbacks=callbacks_list)
    t2 = time.time()
    
run_training(epochs=5)
# continue_training(epochs=100)

标签：单机,Keras,多卡,checkpoint,train,test,import,model,save
来源： https://www.cnblogs.com/demo-deng/p/15856442.html

本站声明： 1. iCode9 技术分享网（下文简称本站）提供的所有内容，仅供技术学习、探讨和分享；
2. 关于本站的所有留言、评论、转载及引用，纯属内容发起人的个人观点，与本站观点和立场无关；
3. 关于本站的所有言论和文字，纯属内容发起人的个人观点，与本站观点和立场无关；
4. 本站文章均是网友提供，不完全保证技术分享内容的完整性、准确性、时效性、风险性和版权归属；如您发现该文章侵犯了您的权益，可联系我们第一时间进行删除；
5. 本站为非盈利性的个人网站，所有内容不会用来进行牟利，也不会利用任何形式的广告来间接获益，纯粹是为了广大技术爱好者提供技术内容和技术思想的分享性交流网站。