Colab初体验

老师说我闹着玩

于 2022-10-24 16:55:53 发布

阅读量375

点赞数

分类专栏： python杂记文章标签：深度学习 pytorch python 经验分享

本文链接：https://blog.csdn.net/weixin_44738304/article/details/127496125

版权

python杂记专栏收录该内容

2 篇文章 0 订阅

订阅专栏

print('hello world!')#万能测试

hello world!

测试GPU运行速度

import tensorflow as tf
device_name = tf.test.gpu_device_name()
if device_name != '/device:GPU:0':
  raise SystemError('GPU device not found')
else: 
  print(device_name)

/device:GPU:0

!nvidia-smi

Mon Oct 24 08:46:47 2022       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |
| N/A   46C    P0    28W /  70W |    312MiB / 15109MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
+-----------------------------------------------------------------------------+

import torch
print(torch.__version__)

加载TF2

%tensorflow_version 2.9.1 #Colab only includes TensorFlow 2.x; %tensorflow_version has no effect.
import tensorflow
print(tensorflow.__version__)

Colab only includes TensorFlow 2.x; %tensorflow_version has no effect.
2.9.2

添加Google的云硬盘到这个工作环境下

from google.colab import drive
drive.mount('/content/drive/')

Drive already mounted at /content/drive/; to attempt to forcibly remount, call drive.mount("/content/drive/", force_remount=True).

切换工作目录

目录名中最好不要有空格。如果有空格，用到此目录名时需要在空格前加\进行转义


%cd "/content/drive/MyDrive/Learing_COLAB"
%ls #显示当前文件夹下的目录

/content/drive/MyDrive/Learing_COLAB
GitHub源码.zip

import os 
# Thisnotepath=os.path.abspath(__file__)#只在代码文件中好用，这里是命令行一样的东西，会报错
Thisnotepath=os.getcwd()

print(Thisnotepath)

/content/drive/MyDrive/Learing_COLAB

# import zipfile
 
# f = zipfile.ZipFile("/content/drive/MyDrive/Learing_COLAB/GitHub源码.zip",'r') # 压缩文件位置
# for file in f.namelist():
#     f.extract(file,"../")               # 解压位置
# f.close()

# /content/drive/MyDrive/deep-learning-for-image-processing-master
%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master"
%ls #显示当前文件夹下的目录

/content/drive/MyDrive/deep-learning-for-image-processing-master
[0m[01;34marticle_link[0m/       [01;34mothers_project[0m/            README.md
[01;34mcourse_ppt[0m/         [01;34mpytorch_classification[0m/    summary_problem.md
[01;34mdata_set[0m/           [01;34mpytorch_keypoint[0m/          [01;34mtensorflow_classification[0m/
[01;34mdeploying_service[0m/  [01;34mpytorch_object_detection[0m/
LICENSE             [01;34mpytorch_segmentation[0m/

用魔法命令运行.py文件

下载花类的数据集

import os
import zipfile

import matplotlib.pyplot as plt
import numpy as np
import requests
from six.moves import urllib
from tqdm import tqdm

def download_from_url(url, dst):
    """
    @param: url to download file
    @param: dst place to put the file
    """
    file_size = int(urllib.request.urlopen(url).info().get('Content-Length', -1))
    if os.path.exists(dst):
        first_byte = os.path.getsize(dst)
    else:
        first_byte = 0
    print(file_size)
    if first_byte >= file_size:
        return file_size
    header = {"Range": "bytes=%s-%s" % (first_byte, file_size)}
    pbar = tqdm(total=file_size, initial=first_byte, unit='B', unit_scale=True, desc=url.split('/')[-1])
    req = requests.get(url, headers=header, stream=True)
    with (open(dst, 'ab')) as f:
        for chunk in req.iter_content(chunk_size=1024):
            if chunk:
                f.write(chunk)
                pbar.update(1024)
    pbar.close()
    return file_size


def download_m2nist_if_not_exist():
    """

    :return:
    """
    data_rootdir = r'/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set'#在和文件同级别的目录下，建立文件夹
    if not os.path.exists(data_rootdir):  # 保存路径不存在，则创建该路径
        os.mkdir(data_rootdir)
    #由于是爬虫，需要设置一下文件写入的名字和后缀
    m2nist_zip_path = os.path.join(data_rootdir, 'flower_photos.tgz')
    if os.path.exists(m2nist_zip_path):
        return
    os.makedirs(data_rootdir, exist_ok=True)
    # 下载的链接地址
    m2nist_zip_url = 'https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz'
    # 执行下载
    download_from_url(m2nist_zip_url, m2nist_zip_path)
    # 解压
    zipf = zipfile.ZipFile(m2nist_zip_path)
    zipf.extractall(data_rootdir)
    zipf.close()

download_m2nist_if_not_exist()

# import zipfile
 
# f = zipfile.ZipFile("/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_photos.zip",'r') # 压缩文件位置
# for file in f.namelist():
#     f.extract(file,"./flower_photos")               # 解压位置
# f.close()

# tar zxvf filename.tgz
%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_data/flower_photos"
%ls #显示当前文件夹下的目录
!tar zxvf r'/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_photos.tgz'
#切换到当前目录
# %cd "/content/drive/MyDrive/deep-learning-for-image-processing-master"
# %ls #显示当前文件夹下的目录

/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_data/flower_photos
[0m[01;34mdaisy[0m/  [01;34mdandelion[0m/  LICENSE.txt  [01;34mroses[0m/  [01;34msunflowers[0m/  [01;34mtulips[0m/
tar (child): r/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_photos.tgz: Cannot open: No such file or directory
tar (child): Error is not recoverable: exiting now
tar: Child returned status 2
tar: Error is not recoverable: exiting now

%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_data/flower_photos"
%ls #显示当前文件夹下的目录
!tar zxvf r'/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_photos.tgz'
%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set"

# import os
from shutil import copy, rmtree
import random


def mk_file(file_path: str):
    if os.path.exists(file_path):
        # 如果文件夹存在，则先删除原文件夹在重新创建
        rmtree(file_path)
    os.makedirs(file_path)


def main():
    # 保证随机可复现
    random.seed(0)

    # 将数据集中10%的数据划分到验证集中
    split_rate = 0.1

    # 指向你解压后的flower_photos文件夹
    cwd = os.getcwd()
    data_root = os.path.join(cwd, "flower_data")
    origin_flower_path = os.path.join(data_root, "flower_photos")
    assert os.path.exists(origin_flower_path), "path '{}' does not exist.".format(origin_flower_path)

    flower_class = [cla for cla in os.listdir(origin_flower_path)
                    if os.path.isdir(os.path.join(origin_flower_path, cla))]

    # 建立保存训练集的文件夹
    train_root = os.path.join(data_root, "train")
    mk_file(train_root)
    for cla in flower_class:
        # 建立每个类别对应的文件夹
        mk_file(os.path.join(train_root, cla))

    # 建立保存验证集的文件夹
    val_root = os.path.join(data_root, "val")
    mk_file(val_root)
    for cla in flower_class:
        # 建立每个类别对应的文件夹
        mk_file(os.path.join(val_root, cla))

    for cla in flower_class:
        cla_path = os.path.join(origin_flower_path, cla)
        images = os.listdir(cla_path)
        num = len(images)
        # 随机采样验证集的索引
        eval_index = random.sample(images, k=int(num*split_rate))
        for index, image in enumerate(images):
            if image in eval_index:
                # 将分配至验证集中的文件复制到相应目录
                image_path = os.path.join(cla_path, image)
                new_path = os.path.join(val_root, cla)
                copy(image_path, new_path)
            else:
                # 将分配至训练集中的文件复制到相应目录
                image_path = os.path.join(cla_path, image)
                new_path = os.path.join(train_root, cla)
                copy(image_path, new_path)
            print("\r[{}] processing [{}/{}]".format(cla, index+1, num), end="")  # processing bar
        print()

    print("processing done!")


if __name__ == '__main__':
    main()

/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_data/flower_photos
[0m[01;34mdaisy[0m/  [01;34mdandelion[0m/  LICENSE.txt  [01;34mroses[0m/  [01;34msunflowers[0m/  [01;34mtulips[0m/
tar (child): r/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_photos.tgz: Cannot open: No such file or directory
tar (child): Error is not recoverable: exiting now
tar: Child returned status 2
tar: Error is not recoverable: exiting now
/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set
[daisy] processing [633/633]
[dandelion] processing [898/898]
[sunflowers] processing [806/806]
[tulips] processing [1317/1317]
[roses] processing [641/641]
processing done!

%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_data"
%ls #显示当前文件夹下的目录

/content/drive/MyDrive/deep-learning-for-image-processing-master/data_set/flower_data
[0m[01;34mflower_photos[0m/  [01;34mtrain[0m/  [01;34mval[0m/


%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master/pytorch_classification/Test3_vggnet"
%ls #显示当前文件夹下的目录


!python '/content/drive/MyDrive/deep-learning-for-image-processing-master/pytorch_classification/Test3_vggnet/train.py'

using cuda:0 device.
Using 2 dataloader workers every process
using 3868 images for training, 427 images for validation.
train epoch[1/30] loss:1.281: 100% 121/121 [01:03<00:00,  1.91it/s]
100% 14/14 [00:02<00:00,  5.00it/s]
[epoch 1] train_loss: 1.407  val_accuracy: 0.457
train epoch[2/30] loss:1.571: 100% 121/121 [00:59<00:00,  2.05it/s]
100% 14/14 [00:03<00:00,  3.50it/s]
[epoch 2] train_loss: 1.285  val_accuracy: 0.461
train epoch[3/30] loss:1.376: 100% 121/121 [00:59<00:00,  2.02it/s]
100% 14/14 [00:03<00:00,  3.67it/s]
[epoch 3] train_loss: 1.266  val_accuracy: 0.466
train epoch[4/30] loss:1.305: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:03<00:00,  3.72it/s]
[epoch 4] train_loss: 1.256  val_accuracy: 0.475
train epoch[5/30] loss:1.566: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:03<00:00,  3.54it/s]
[epoch 5] train_loss: 1.210  val_accuracy: 0.504
train epoch[6/30] loss:0.911: 100% 121/121 [01:00<00:00,  2.00it/s]
100% 14/14 [00:04<00:00,  3.43it/s]
[epoch 6] train_loss: 1.121  val_accuracy: 0.625
train epoch[7/30] loss:0.992: 100% 121/121 [00:59<00:00,  2.02it/s]
100% 14/14 [00:03<00:00,  3.53it/s]
[epoch 7] train_loss: 1.061  val_accuracy: 0.644
train epoch[8/30] loss:1.062: 100% 121/121 [00:59<00:00,  2.02it/s]
100% 14/14 [00:03<00:00,  3.71it/s]
[epoch 8] train_loss: 0.997  val_accuracy: 0.621
train epoch[9/30] loss:1.116: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:03<00:00,  3.57it/s]
[epoch 9] train_loss: 0.969  val_accuracy: 0.681
train epoch[10/30] loss:0.746: 100% 121/121 [00:59<00:00,  2.02it/s]
100% 14/14 [00:03<00:00,  3.68it/s]
[epoch 10] train_loss: 0.929  val_accuracy: 0.707
train epoch[11/30] loss:0.697: 100% 121/121 [00:59<00:00,  2.02it/s]
100% 14/14 [00:03<00:00,  3.60it/s]
[epoch 11] train_loss: 0.890  val_accuracy: 0.712
train epoch[12/30] loss:0.785: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.69it/s]
[epoch 12] train_loss: 0.876  val_accuracy: 0.707
train epoch[13/30] loss:0.782: 100% 121/121 [00:59<00:00,  2.05it/s]
100% 14/14 [00:03<00:00,  3.65it/s]
[epoch 13] train_loss: 0.861  val_accuracy: 0.710
train epoch[14/30] loss:0.643: 100% 121/121 [00:59<00:00,  2.02it/s]
100% 14/14 [00:03<00:00,  3.92it/s]
[epoch 14] train_loss: 0.816  val_accuracy: 0.724
train epoch[15/30] loss:0.556: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:03<00:00,  3.79it/s]
[epoch 15] train_loss: 0.766  val_accuracy: 0.738
train epoch[16/30] loss:0.801: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.59it/s]
[epoch 16] train_loss: 0.802  val_accuracy: 0.663
train epoch[17/30] loss:0.636: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:04<00:00,  3.35it/s]
[epoch 17] train_loss: 0.754  val_accuracy: 0.754
train epoch[18/30] loss:0.507: 100% 121/121 [01:00<00:00,  2.01it/s]
100% 14/14 [00:03<00:00,  3.64it/s]
[epoch 18] train_loss: 0.729  val_accuracy: 0.740
train epoch[19/30] loss:0.621: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.58it/s]
[epoch 19] train_loss: 0.735  val_accuracy: 0.759
train epoch[20/30] loss:0.824: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:04<00:00,  3.44it/s]
[epoch 20] train_loss: 0.697  val_accuracy: 0.747
train epoch[21/30] loss:0.848: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:05<00:00,  2.58it/s]
[epoch 21] train_loss: 0.706  val_accuracy: 0.770
train epoch[22/30] loss:0.747: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:03<00:00,  3.55it/s]
[epoch 22] train_loss: 0.678  val_accuracy: 0.759
train epoch[23/30] loss:0.420: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:04<00:00,  3.45it/s]
[epoch 23] train_loss: 0.652  val_accuracy: 0.761
train epoch[24/30] loss:0.380: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.63it/s]
[epoch 24] train_loss: 0.660  val_accuracy: 0.759
train epoch[25/30] loss:1.107: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:04<00:00,  3.13it/s]
[epoch 25] train_loss: 0.653  val_accuracy: 0.778
train epoch[26/30] loss:0.548: 100% 121/121 [00:59<00:00,  2.03it/s]
100% 14/14 [00:03<00:00,  3.60it/s]
[epoch 26] train_loss: 0.627  val_accuracy: 0.770
train epoch[27/30] loss:0.530: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.59it/s]
[epoch 27] train_loss: 0.644  val_accuracy: 0.770
train epoch[28/30] loss:0.488: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.72it/s]
[epoch 28] train_loss: 0.610  val_accuracy: 0.787
train epoch[29/30] loss:0.487: 100% 121/121 [01:00<00:00,  2.01it/s]
100% 14/14 [00:04<00:00,  3.50it/s]
[epoch 29] train_loss: 0.590  val_accuracy: 0.763
train epoch[30/30] loss:0.505: 100% 121/121 [00:59<00:00,  2.04it/s]
100% 14/14 [00:03<00:00,  3.75it/s]
[epoch 30] train_loss: 0.609  val_accuracy: 0.808
Finished Training

# /content/drive/MyDrive/deep-learning-for-image-processing-master/pytorch_classification/Test4_googlenet/train.py
%cd "/content/drive/MyDrive/deep-learning-for-image-processing-master/pytorch_classification/Test4_googlenet"
%ls

/content/drive/MyDrive/deep-learning-for-image-processing-master/pytorch_classification/Test4_googlenet
class_indices.json  model.py  predict.py  train.py

!python '/content/drive/MyDrive/deep-learning-for-image-processing-master/pytorch_classification/Test4_googlenet/train.py'

using cuda:0 device.
Using 2 dataloader workers every process
using 3868 images for training, 427 images for validation.
train epoch[1/64] loss:2.411: 100% 16/16 [00:23<00:00,  1.44s/it]
100% 2/2 [00:02<00:00,  1.31s/it]
[epoch 1] train_loss: 5.018  val_accuracy: 0.389
train epoch[2/64] loss:1.711: 100% 16/16 [00:23<00:00,  1.45s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 2] train_loss: 2.123  val_accuracy: 0.539
train epoch[3/64] loss:1.510: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.40s/it]
[epoch 3] train_loss: 1.817  val_accuracy: 0.520
train epoch[4/64] loss:2.249: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.40s/it]
[epoch 4] train_loss: 1.821  val_accuracy: 0.597
train epoch[5/64] loss:1.562: 100% 16/16 [00:22<00:00,  1.44s/it]
100% 2/2 [00:04<00:00,  2.36s/it]
[epoch 5] train_loss: 1.710  val_accuracy: 0.593
train epoch[6/64] loss:1.582: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 6] train_loss: 1.578  val_accuracy: 0.651
train epoch[7/64] loss:1.163: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 7] train_loss: 1.457  val_accuracy: 0.674
train epoch[8/64] loss:1.430: 100% 16/16 [00:22<00:00,  1.44s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 8] train_loss: 1.466  val_accuracy: 0.642
train epoch[9/64] loss:1.180: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 9] train_loss: 1.440  val_accuracy: 0.660
train epoch[10/64] loss:1.068: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 10] train_loss: 1.327  val_accuracy: 0.618
train epoch[11/64] loss:1.482: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.38s/it]
[epoch 11] train_loss: 1.357  val_accuracy: 0.604
train epoch[12/64] loss:1.660: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 12] train_loss: 1.465  val_accuracy: 0.717
train epoch[13/64] loss:1.518: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.32s/it]
[epoch 13] train_loss: 1.296  val_accuracy: 0.714
train epoch[14/64] loss:1.100: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.33s/it]
[epoch 14] train_loss: 1.215  val_accuracy: 0.766
train epoch[15/64] loss:1.881: 100% 16/16 [00:25<00:00,  1.60s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 15] train_loss: 1.183  val_accuracy: 0.672
train epoch[16/64] loss:1.119: 100% 16/16 [00:22<00:00,  1.44s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 16] train_loss: 1.310  val_accuracy: 0.700
train epoch[17/64] loss:1.028: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 17] train_loss: 1.193  val_accuracy: 0.721
train epoch[18/64] loss:1.090: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 18] train_loss: 1.141  val_accuracy: 0.698
train epoch[19/64] loss:0.692: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 19] train_loss: 1.177  val_accuracy: 0.742
train epoch[20/64] loss:1.115: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 20] train_loss: 1.061  val_accuracy: 0.585
train epoch[21/64] loss:1.457: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 21] train_loss: 1.262  val_accuracy: 0.735
train epoch[22/64] loss:0.905: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 22] train_loss: 1.106  val_accuracy: 0.747
train epoch[23/64] loss:0.857: 100% 16/16 [00:22<00:00,  1.43s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 23] train_loss: 1.036  val_accuracy: 0.724
train epoch[24/64] loss:0.910: 100% 16/16 [00:25<00:00,  1.57s/it]
100% 2/2 [00:02<00:00,  1.33s/it]
[epoch 24] train_loss: 1.020  val_accuracy: 0.740
train epoch[25/64] loss:1.008: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 25] train_loss: 1.030  val_accuracy: 0.674
train epoch[26/64] loss:1.175: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 26] train_loss: 1.054  val_accuracy: 0.686
train epoch[27/64] loss:0.620: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 27] train_loss: 1.016  val_accuracy: 0.733
train epoch[28/64] loss:0.860: 100% 16/16 [00:22<00:00,  1.43s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 28] train_loss: 0.951  val_accuracy: 0.766
train epoch[29/64] loss:0.746: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 29] train_loss: 0.929  val_accuracy: 0.747
train epoch[30/64] loss:1.387: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 30] train_loss: 0.976  val_accuracy: 0.738
train epoch[31/64] loss:0.824: 100% 16/16 [00:25<00:00,  1.57s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 31] train_loss: 0.921  val_accuracy: 0.766
train epoch[32/64] loss:0.996: 100% 16/16 [00:23<00:00,  1.46s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 32] train_loss: 0.973  val_accuracy: 0.672
train epoch[33/64] loss:1.363: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 33] train_loss: 1.106  val_accuracy: 0.740
train epoch[34/64] loss:1.219: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 34] train_loss: 0.967  val_accuracy: 0.726
train epoch[35/64] loss:0.863: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 35] train_loss: 1.026  val_accuracy: 0.752
train epoch[36/64] loss:0.723: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 36] train_loss: 0.864  val_accuracy: 0.756
train epoch[37/64] loss:0.785: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.43s/it]
[epoch 37] train_loss: 0.817  val_accuracy: 0.740
train epoch[38/64] loss:0.574: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:04<00:00,  2.15s/it]
[epoch 38] train_loss: 0.946  val_accuracy: 0.787
train epoch[39/64] loss:1.012: 100% 16/16 [00:23<00:00,  1.48s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 39] train_loss: 0.880  val_accuracy: 0.761
train epoch[40/64] loss:0.936: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 40] train_loss: 0.843  val_accuracy: 0.817
train epoch[41/64] loss:0.640: 100% 16/16 [00:22<00:00,  1.44s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 41] train_loss: 0.763  val_accuracy: 0.796
train epoch[42/64] loss:0.833: 100% 16/16 [00:22<00:00,  1.43s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 42] train_loss: 0.770  val_accuracy: 0.787
train epoch[43/64] loss:0.655: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 43] train_loss: 0.816  val_accuracy: 0.766
train epoch[44/64] loss:0.646: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.40s/it]
[epoch 44] train_loss: 0.830  val_accuracy: 0.815
train epoch[45/64] loss:0.748: 100% 16/16 [00:22<00:00,  1.43s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 45] train_loss: 0.769  val_accuracy: 0.817
train epoch[46/64] loss:0.792: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 46] train_loss: 0.742  val_accuracy: 0.794
train epoch[47/64] loss:0.554: 100% 16/16 [00:23<00:00,  1.45s/it]
100% 2/2 [00:04<00:00,  2.36s/it]
[epoch 47] train_loss: 0.739  val_accuracy: 0.789
train epoch[48/64] loss:0.321: 100% 16/16 [00:23<00:00,  1.46s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 48] train_loss: 0.701  val_accuracy: 0.778
train epoch[49/64] loss:0.836: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 49] train_loss: 0.720  val_accuracy: 0.756
train epoch[50/64] loss:0.345: 100% 16/16 [00:24<00:00,  1.52s/it]
100% 2/2 [00:02<00:00,  1.36s/it]
[epoch 50] train_loss: 0.798  val_accuracy: 0.799
train epoch[51/64] loss:0.420: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.39s/it]
[epoch 51] train_loss: 0.767  val_accuracy: 0.815
train epoch[52/64] loss:0.714: 100% 16/16 [00:22<00:00,  1.43s/it]
100% 2/2 [00:02<00:00,  1.33s/it]
[epoch 52] train_loss: 0.733  val_accuracy: 0.789
train epoch[53/64] loss:0.989: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 53] train_loss: 0.724  val_accuracy: 0.810
train epoch[54/64] loss:0.623: 100% 16/16 [00:23<00:00,  1.44s/it]
100% 2/2 [00:02<00:00,  1.40s/it]
[epoch 54] train_loss: 0.684  val_accuracy: 0.824
train epoch[55/64] loss:0.752: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 55] train_loss: 0.665  val_accuracy: 0.778
train epoch[56/64] loss:0.630: 100% 16/16 [00:22<00:00,  1.41s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 56] train_loss: 0.692  val_accuracy: 0.806
train epoch[57/64] loss:0.579: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:03<00:00,  1.77s/it]
[epoch 57] train_loss: 0.652  val_accuracy: 0.806
train epoch[58/64] loss:0.640: 100% 16/16 [00:25<00:00,  1.56s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 58] train_loss: 0.680  val_accuracy: 0.782
train epoch[59/64] loss:0.799: 100% 16/16 [00:22<00:00,  1.39s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 59] train_loss: 0.669  val_accuracy: 0.787
train epoch[60/64] loss:0.784: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.38s/it]
[epoch 60] train_loss: 0.718  val_accuracy: 0.796
train epoch[61/64] loss:0.454: 100% 16/16 [00:22<00:00,  1.40s/it]
100% 2/2 [00:02<00:00,  1.33s/it]
[epoch 61] train_loss: 0.666  val_accuracy: 0.801
train epoch[62/64] loss:0.754: 100% 16/16 [00:22<00:00,  1.44s/it]
100% 2/2 [00:02<00:00,  1.34s/it]
[epoch 62] train_loss: 0.667  val_accuracy: 0.838
train epoch[63/64] loss:0.257: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.37s/it]
[epoch 63] train_loss: 0.611  val_accuracy: 0.820
train epoch[64/64] loss:0.315: 100% 16/16 [00:22<00:00,  1.42s/it]
100% 2/2 [00:02<00:00,  1.35s/it]
[epoch 64] train_loss: 0.618  val_accuracy: 0.787
Finished Training