distribute as package

Author: marsggbo

Diff for: LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2019 marsggbo
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

Diff for: README.md

@@ -1,7 +1,20 @@
 # torchline
 Easy to use Pytorch
 
+# Dependences
 
+- Python>=3.6
+- Pytorch==1.3.1
+- torchvision==0.4.2
+- yacs==0.1.6
+- pytorch-lightning==0.5.3.2
+
+
+# Install
+
+```bash
+pip install torchline
+```
 
 # Structures
 

Diff for: release.md

@@ -3,15 +3,22 @@
 
 ## v0.1
 
-### 更新信息
+### 2019.12.12 更新信息
 - 基本框架搭建完成
 - 可正常运行cifar10_demo
 
+
+### 2019.12.13 更新信息
+- 实现setup安装
+- 完善包之间的引用关系
+
 ### todo list
 
 - [ ] 弄清楚logging机制
 - [ ] save和load模型，优化器参数
 - [ ] skin数据集读取测试
 - [ ] 构建skin project
 - [ ] 能否预测单张图片？
-- [ ] 构建一个简单地API接口
+- [ ] 构建一个简单地API接口
+- [ ] 进一步完善包导入
+- [ ] 完善使用文档

Diff for: requirements.txt

@@ -0,0 +1,4 @@
+torch==1.3.1
+torchvision==0.4.2
+yacs==0.1.6
+pytorch-lightning==0.5.3.2

Diff for: setup.py

@@ -0,0 +1,23 @@
+from setuptools import setup, find_packages
+
+with open('requirements.txt', 'r') as f:
+    requirements = f.read().splitlines()
+
+setup(
+    name="torchline", # Replace with your own username
+    version="0.1",
+    author="marsggbo",
+    author_email="[email protected]",
+    description="A framework for easy to use Pytorch",
+    long_description='...',
+    long_description_content_type="text/markdown",
+    url="https://github.com/marsggbo/torchline",
+    packages=find_packages(exclude=("tests", "projects")),
+    install_requires=requirements,
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+    ],
+    python_requires='>=3.6',
+)

Diff for: torchline/__init__.py

@@ -0,0 +1,6 @@
+from .config import *
+from .data import *
+from .engine import *
+from .losses import *
+from .models import *
+from .utils import *

Diff for: torchline/config/skin10.yaml

@@ -1,5 +1,6 @@
-from .build import build_data
+from .build import build_data, DATASET_REGISTRY
 from .skin100 import Skin100Dataset
 from .skin10 import Skin10Dataset
 from .common_datasets import MNIST, CIFAR10
-from .transforms import DefaultTransforms, build_transforms
+from .transforms import DefaultTransforms, build_transforms
+from .autoaugment import ImageNetPolicy, CIFAR10Policy, SVHNPolicy, SubPolicy

Diff for: torchline/data/__init__.py

@@ -0,0 +1 @@
+from .lightning_module_template import LightningTemplateModel

Diff for: torchline/engine/__init__.py

@@ -1,2 +1,2 @@
-from .build import build_loss_fn
+from .build import build_loss_fn, LOSS_FN_REGISTRY
 from .loss import CrossEntropy

Diff for: torchline/losses/__init__.py

@@ -1,2 +1,2 @@
-from .build import build_model
+from .build import build_model, META_ARCH_REGISTRY
 from .resnet_models import *

Diff for: torchline/models/__init__.py

@@ -36,13 +36,8 @@ def __init__(self, cfg):
         self.layer3 = nn.Sequential(*model[6])
         self.layer4 = nn.Sequential(*model[7])
 
-        self.clf = nn.Sequential(
-            nn.Conv2d(2048, 512, kernel_size=1),
-            nn.ReLU(inplace=True),
-            nn.Conv2d(512, self.num_classes, kernel_size=1),
-            nn.ReLU(inplace=True),
-            nn.AdaptiveAvgPool2d((1,1))
-        )
+        self.avg_pool = nn.AdaptiveAvgPool2d((1,1))
+        self.clf = nn.Linear(in_features=2048, out_features=self.num_classes)
 
     def extract_features(self, x):
         assert len(self.cfg.MODEL.FEATURES) >= 1
@@ -67,13 +62,14 @@ def forward(self, x):
             x (tensor): N*c*h*w
 
         return:
-            img_cls_preds (tensor): N*classes
+            predictions (tensor): N*classes
         '''
         bs= x.shape[0]
         features = self.extract_features(x)
-        img_cls_preds = self.clf(features).view(bs, -1)
+        predictions = self.avg_pool(features).view(bs, -1)
+        predictions = self.clf(predictions).view(bs, -1)
 
-        return img_cls_preds
+        return predictions