Fix missing link in document

docs/specification/config.md

@@ -57,10 +57,10 @@ Config yaml file has following Unstandard YAML [pyyaml](https://pyyaml.org/) def
 |`!!python/object:`|module.cls|module.cls instance|
 
 You should remove these tags before using without python (pyyaml).
-i.e. [classification.yaml](/blueoil/configs/example/classification.yaml)
+i.e. [classification.yaml](https://github.com/blue-oil/blueoil/blob/master/blueoil/configs/example/classification.yaml)
 
 # Meta yaml
 The meta yaml file is created from config python file, that is only include `export` column parameter.
 For application, the exported meta yaml file has following standard yaml, it means it is not include `!!python...` generated by pyyaml.
 
-i.e. [classification_meta.yaml](/blueoil/configs/example/classification_meta.yaml)
+i.e. [classification_meta.yaml](https://github.com/blue-oil/blueoil/blob/master/blueoil/configs/example/classification_meta.yaml)

docs/usage/export.md

@@ -5,8 +5,8 @@ In the case with `images` option, create each layer output value npy files in `e
 
 * Load config file from saved experiment dir.
 * Export config file to yaml. See also [Config specification](../specification/config.md).
-  * `config.yaml` can be used for training and evaluation in python. i.e. [classification.yaml](../../blueoil/configs/example/classification.yaml) is exported from [classification.py](../../blueoil/configs/example/classification.py)
-  * `meta.yaml` include only few parameter for application such as demo. i.e. [classification_meta.yaml](../../blueoil/configs/example/classification_meta.yaml) is exported from [classification.py](../../blueoil/configs/example/classification.py)
+  * `config.yaml` can be used for training and evaluation in python. i.e. [classification.yaml](https://github.com/blue-oil/blueoil/blob/master/blueoil/configs/example/classification.yaml) is exported from [classification.py](https://github.com/blue-oil/blueoil/blob/master/blueoil/configs/example/classification.py)
+  * `meta.yaml` include only few parameter for application such as demo. i.e. [classification_meta.yaml](https://github.com/blue-oil/blueoil/blob/master/blueoil/configs/example/classification_meta.yaml) is exported from [classification.py](https://github.com/blue-oil/blueoil/blob/master/blueoil/configs/example/classification.py)
 * Save the model protocol buffer files (tf) for DLK converter.
 * Output each layer npy files for DLK converter debug.
 * Write summary in tensorboard `export` dir.

docs/usage/init.md

@@ -48,7 +48,7 @@ common:
         - probability: 0.5
 ```
 
-Please see <a href="../reference/data_augmentor.html">the data augmentor reference page</a>, when you want to know about augmentation methods. All of the augmentation methods and the required parameters are explained there. Method names in the generated yaml file correspond to class names under `blueoil.data_augmentor` in the reference.
+Please see <a href="../autoapi/blueoil/data_augmentor/index.html">the data augmentor reference page</a>, when you want to know about augmentation methods. All of the augmentation methods and the required parameters are explained there. Method names in the generated yaml file correspond to class names under `blueoil.data_augmentor` in the reference.
 
 
 #### optimizer