iicd/autodl-projects
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

master -> main

Browse Source
This commit is contained in:
D-X-Y 2021-03-01 21:02:29 +08:00
parent af3ce50a65
commit 493478800c
13 changed files with 32 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
.github/CONTRIBUTING.md vendored
View File

@ -19,7 +19,7 @@ This section guides you through submitting a bug report for AutoDL-Projects.
Following these guidelines helps maintainers and the community understand your report :pencil:, reproduce the behavior :computer: :computer:, and find related reports :mag_right:.
When you are creating a bug report, please include as many details as possible.
Fill out [the required template](https://github.com/D-X-Y/AutoDL-Projects/blob/master/.github/ISSUE_TEMPLATE/bug-report.md). The information it asks for helps us resolve issues faster.
Fill out [the required template](https://github.com/D-X-Y/AutoDL-Projects/blob/main/.github/ISSUE_TEMPLATE/bug-report.md). The information it asks for helps us resolve issues faster.
> **Note:** If you find a **Closed** issue that seems like it is the same thing that you're experiencing, open a new issue and include a link to the original issue in the body of your new one.

3
.gitmodules vendored
View File

@ -1,6 +1,3 @@
[submodule "qlib-git"]
path = .latent-data/qlib
url = git@github.com:microsoft/qlib.git
[submodule ".latent-data/qlib"]
path = .latent-data/qlib
url = git@github.com:microsoft/qlib.git

18
README.md
View File

@ -7,7 +7,7 @@
Automated Deep Learning Projects (AutoDL-Projects) is an open source, lightweight, but useful project for everyone.
This project implemented several neural architecture search (NAS) and hyper-parameter optimization (HPO) algorithms.
中文介绍见[README_CN.md](README_CN.md)
中文介绍见[README_CN.md](https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/README_CN.md)
**Who should consider using AutoDL-Projects**
@ -36,38 +36,38 @@ At this moment, this project provides the following algorithms and scripts to ru
<td rowspan="6" align="center" valign="middle" halign="middle"> NAS </td>
<td align="center" valign="middle"> TAS </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1905.09717">Network Pruning via Transformable Architecture Search</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NeurIPS-2019-TAS.md">NeurIPS-2019-TAS.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NeurIPS-2019-TAS.md">NeurIPS-2019-TAS.md</a> </td>
</tr>
<tr> <!-- (2-nd row) -->
<td align="center" valign="middle"> DARTS </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1806.09055">DARTS: Differentiable Architecture Search</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/ICLR-2019-DARTS.md">ICLR-2019-DARTS.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/ICLR-2019-DARTS.md">ICLR-2019-DARTS.md</a> </td>
</tr>
<tr> <!-- (3-nd row) -->
<td align="center" valign="middle"> GDAS </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1910.04465">Searching for A Robust Neural Architecture in Four GPU Hours</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/CVPR-2019-GDAS.md">CVPR-2019-GDAS.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/CVPR-2019-GDAS.md">CVPR-2019-GDAS.md</a> </td>
</tr>
<tr> <!-- (4-rd row) -->
<td align="center" valign="middle"> SETN </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1910.05733">One-Shot Neural Architecture Search via Self-Evaluated Template Network</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/ICCV-2019-SETN.md">ICCV-2019-SETN.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/ICCV-2019-SETN.md">ICCV-2019-SETN.md</a> </td>
</tr>
<tr> <!-- (5-th row) -->
<td align="center" valign="middle"> NAS-Bench-201 </td>
<td align="center" valign="middle"> <a href="https://openreview.net/forum?id=HJxyZkBKDr"> NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> </td>
</tr>
<tr> <!-- (6-th row) -->
<td align="center" valign="middle"> NATS-Bench </td>
<td align="center" valign="middle"> <a href="https://xuanyidong.com/assets/projects/NATS-Bench"> NATS-Bench: Benchmarking NAS Algorithms for Architecture Topology and Size</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
</tr>
<tr> <!-- (7-th row) -->
<td align="center" valign="middle"> ... </td>
<td align="center" valign="middle"> ENAS / REA / REINFORCE / BOHB </td>
<td align="center" valign="middle"> Please check the original papers </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
</tr>
<tr> <!-- (start second block) -->
<td rowspan="1" align="center" valign="middle" halign="middle"> HPO </td>
@ -79,7 +79,7 @@ At this moment, this project provides the following algorithms and scripts to ru
<td rowspan="1" align="center" valign="middle" halign="middle"> Basic </td>
<td align="center" valign="middle"> ResNet </td>
<td align="center" valign="middle"> Deep Learning-based Image Classification </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/BASELINE.md">BASELINE.md</a> </a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/BASELINE.md">BASELINE.md</a> </a> </td>
</tr>
</tbody>
</table>

4
docs/CVPR-2019-GDAS.md
View File

@ -22,7 +22,7 @@ from utils import get_model_infos
flop, param = get_model_infos(net, (1,3,32,32))
```
2. Different NAS-searched architectures are defined [here](https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/nas_infer_model/DXYs/genotypes.py).
2. Different NAS-searched architectures are defined [here](https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/nas_infer_model/DXYs/genotypes.py).
## Usage
@ -34,7 +34,7 @@ CUDA_VISIBLE_DEVICES=0 bash ./scripts/nas-infer-train.sh cifar10 GDAS_V1 96 -1
CUDA_VISIBLE_DEVICES=0 bash ./scripts/nas-infer-train.sh cifar100 GDAS_V1 96 -1
CUDA_VISIBLE_DEVICES=0,1,2,3 bash ./scripts/nas-infer-train.sh imagenet-1k GDAS_V1 256 -1
```
If you are interested in the configs of each NAS-searched architecture, they are defined at [genotypes.py](https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/nas_infer_model/DXYs/genotypes.py).
If you are interested in the configs of each NAS-searched architecture, they are defined at [genotypes.py](https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/nas_infer_model/DXYs/genotypes.py).
### Searching on the NASNet search space

2
docs/ICCV-2019-SETN.md
View File

@ -18,7 +18,7 @@ from utils import get_model_infos
flop, param = get_model_infos(net, (1,3,32,32))
```
2. Different NAS-searched architectures are defined [here](https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/nas_infer_model/DXYs/genotypes.py).
2. Different NAS-searched architectures are defined [here](https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/nas_infer_model/DXYs/genotypes.py).
## Usage

2
docs/ICLR-2019-DARTS.md
View File

@ -16,7 +16,7 @@ This command will start to use the first-order DARTS to search architectures on
CUDA_VISIBLE_DEVICES=0 bash ./scripts-search/DARTS1V-search-NASNet-space.sh cifar10 -1
```
After searching, if you want to train the searched architecture found by the above scripts, you need to add the config of that architecture (will be printed in log) in [genotypes.py](https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/nas_infer_model/DXYs/genotypes.py).
After searching, if you want to train the searched architecture found by the above scripts, you need to add the config of that architecture (will be printed in log) in [genotypes.py](https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/nas_infer_model/DXYs/genotypes.py).
In future, I will add a more eligent way to train the searched architecture from the DARTS search space.

6
docs/NAS-Bench-201-PURE.md
View File

@ -1,6 +1,6 @@
# [NAS-BENCH-201: Extending the Scope of Reproducible Neural Architecture Search](https://openreview.net/forum?id=HJxyZkBKDr)
**Since our NAS-BENCH-201 has been extended to NATS-Bench, this `README` is deprecated and not maintained. Please use [NATS-Bench](https://github.com/D-X-Y/AutoDL-Projects/blob/master/docs/NATS-Bench.md), which has 5x more architecture information and faster API than NAS-BENCH-201.**
**Since our NAS-BENCH-201 has been extended to NATS-Bench, this `README` is deprecated and not maintained. Please use [NATS-Bench](https://github.com/D-X-Y/AutoDL-Projects/blob/main/docs/NATS-Bench.md), which has 5x more architecture information and faster API than NAS-BENCH-201.**
We propose an algorithm-agnostic NAS benchmark (NAS-Bench-201) with a fixed search space, which provides a unified benchmark for almost any up-to-date NAS algorithms.
The design of our search space is inspired by that used in the most popular cell-based searching algorithms, where a cell is represented as a directed acyclic graph.
@ -44,7 +44,7 @@ It is recommended to put these data into `$TORCH_HOME` (`~/.torch/` by default).
## How to Use NAS-Bench-201
**More usage can be found in [our test codes](https://github.com/D-X-Y/AutoDL-Projects/blob/master/exps/NAS-Bench-201/test-nas-api.py)**.
**More usage can be found in [our test codes](https://github.com/D-X-Y/AutoDL-Projects/blob/main/exps/NAS-Bench-201/test-nas-api.py)**.
1. Creating an API instance from a file:
```
@ -161,7 +161,7 @@ api.reload('{:}/{:}'.format(os.environ['TORCH_HOME'], 'NAS-BENCH-201-4-v1.0-arch
weights = api.get_net_param(3, 'cifar10', None) # Obtaining the weights of all trials for the 3-th architecture on cifar10. It will returns a dict, where the key is the seed and the value is the trained weights.
```
To obtain the training and evaluation information (please see the comments [here](https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/nas_201_api/api_201.py#L142)):
To obtain the training and evaluation information (please see the comments [here](https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/nas_201_api/api_201.py#L142)):
```
api.get_more_info(112, 'cifar10', None, hp='200', is_random=True)
# Query info of last training epoch for 112-th architecture

6
docs/NAS-Bench-201.md
View File

@ -1,6 +1,6 @@
# [NAS-BENCH-201: Extending the Scope of Reproducible Neural Architecture Search](https://openreview.net/forum?id=HJxyZkBKDr)
**Since our NAS-BENCH-201 has been extended to NATS-Bench, this README is deprecated and not maintained. Please use [NATS-Bench](https://github.com/D-X-Y/AutoDL-Projects/blob/master/docs/NATS-Bench.md), which has 5x more architecture information and faster API than NAS-BENCH-201.**
**Since our NAS-BENCH-201 has been extended to NATS-Bench, this README is deprecated and not maintained. Please use [NATS-Bench](https://github.com/D-X-Y/AutoDL-Projects/blob/main/docs/NATS-Bench.md), which has 5x more architecture information and faster API than NAS-BENCH-201.**
We propose an algorithm-agnostic NAS benchmark (NAS-Bench-201) with a fixed search space, which provides a unified benchmark for almost any up-to-date NAS algorithms.
The design of our search space is inspired by that used in the most popular cell-based searching algorithms, where a cell is represented as a directed acyclic graph.
@ -42,7 +42,7 @@ It is recommended to put these data into `$TORCH_HOME` (`~/.torch/` by default).
## How to Use NAS-Bench-201
**More usage can be found in [our test codes](https://github.com/D-X-Y/AutoDL-Projects/blob/master/exps/NAS-Bench-201/test-nas-api.py)**.
**More usage can be found in [our test codes](https://github.com/D-X-Y/AutoDL-Projects/blob/main/exps/NAS-Bench-201/test-nas-api.py)**.
1. Creating an API instance from a file:
```
@ -159,7 +159,7 @@ api.reload('{:}/{:}'.format(os.environ['TORCH_HOME'], 'NAS-BENCH-201-4-v1.0-arch
weights = api.get_net_param(3, 'cifar10', None) # Obtaining the weights of all trials for the 3-th architecture on cifar10. It will returns a dict, where the key is the seed and the value is the trained weights.
```
To obtain the training and evaluation information (please see the comments [here](https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/nas_201_api/api_201.py#L142)):
To obtain the training and evaluation information (please see the comments [here](https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/nas_201_api/api_201.py#L142)):
```
api.get_more_info(112, 'cifar10', None, hp='200', is_random=True)
# Query info of last training epoch for 112-th architecture

2
docs/NeurIPS-2019-TAS.md
View File

@ -31,7 +31,7 @@ args: `cifar10` indicates the dataset name, `ResNet56` indicates the basemodel n
**Model Configuration**
The searched shapes for ResNet-20/32/56/110/164 and ResNet-18/50 in Table 3/4 in the original paper are listed in [`configs/NeurIPS-2019`](https://github.com/D-X-Y/AutoDL-Projects/tree/master/configs/NeurIPS-2019).
The searched shapes for ResNet-20/32/56/110/164 and ResNet-18/50 in Table 3/4 in the original paper are listed in [`configs/NeurIPS-2019`](https://github.com/D-X-Y/AutoDL-Projects/tree/main/configs/NeurIPS-2019).
**Search for the depth configuration of ResNet**
```

16
README_CN.md → docs/README_CN.md
View File

@ -37,38 +37,38 @@
<td rowspan="6" align="center" valign="middle" halign="middle"> NAS </td>
<td align="center" valign="middle"> TAS </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1905.09717">Network Pruning via Transformable Architecture Search</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NeurIPS-2019-TAS.md">NeurIPS-2019-TAS.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NeurIPS-2019-TAS.md">NeurIPS-2019-TAS.md</a> </td>
</tr>
<tr> <!-- (2-nd row) -->
<td align="center" valign="middle"> DARTS </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1806.09055">DARTS: Differentiable Architecture Search</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/ICLR-2019-DARTS.md">ICLR-2019-DARTS.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/ICLR-2019-DARTS.md">ICLR-2019-DARTS.md</a> </td>
</tr>
<tr> <!-- (3-nd row) -->
<td align="center" valign="middle"> GDAS </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1910.04465">Searching for A Robust Neural Architecture in Four GPU Hours</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/CVPR-2019-GDAS.md">CVPR-2019-GDAS.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/CVPR-2019-GDAS.md">CVPR-2019-GDAS.md</a> </td>
</tr>
<tr> <!-- (4-rd row) -->
<td align="center" valign="middle"> SETN </td>
<td align="center" valign="middle"> <a href="https://arxiv.org/abs/1910.05733">One-Shot Neural Architecture Search via Self-Evaluated Template Network</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/ICCV-2019-SETN.md">ICCV-2019-SETN.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/ICCV-2019-SETN.md">ICCV-2019-SETN.md</a> </td>
</tr>
<tr> <!-- (5-th row) -->
<td align="center" valign="middle"> NAS-Bench-201 </td>
<td align="center" valign="middle"> <a href="https://openreview.net/forum?id=HJxyZkBKDr"> NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> </td>
</tr>
<tr> <!-- (6-th row) -->
<td align="center" valign="middle"> NATS-Bench </td>
<td align="center" valign="middle"> <a href="https://xuanyidong.com/assets/projects/NATS-Bench"> NATS-Bench: Benchmarking NAS Algorithms for Architecture Topology and Size</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
</tr>
<tr> <!-- (7-th row) -->
<td align="center" valign="middle"> ... </td>
<td align="center" valign="middle"> ENAS / REA / REINFORCE / BOHB </td>
<td align="center" valign="middle"> Please check the original papers. </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NAS-Bench-201.md">NAS-Bench-201.md</a> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/NATS-Bench.md">NATS-Bench.md</a> </td>
</tr>
<tr> <!-- (start second block) -->
<td rowspan="1" align="center" valign="middle" halign="middle"> HPO </td>
@ -80,7 +80,7 @@
<td rowspan="1" align="center" valign="middle" halign="middle"> Basic </td>
<td align="center" valign="middle"> ResNet </td>
<td align="center" valign="middle"> Deep Learning-based Image Classification </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/master/docs/BASELINE.md">BASELINE.md</a> </a> </td>
<td align="center" valign="middle"> <a href="https://github.com/D-X-Y/AutoDL-Projects/tree/main/docs/BASELINE.md">BASELINE.md</a> </a> </td>
</tr>
</tbody>
</table>

2
exps/NATS-algos/search-size.py
View File

@ -8,7 +8,7 @@
# - masking + sampling (mask_rl) from "Can Weight Sharing Outperform Random Architecture Search? An Investigation With TuNAS, CVPR 2020"
#
# For simplicity, we use tas, mask_gumbel, and mask_rl to refer these three strategies. Their official implementations are at the following links:
# - TAS: https://github.com/D-X-Y/AutoDL-Projects/blob/master/docs/NeurIPS-2019-TAS.md
# - TAS: https://github.com/D-X-Y/AutoDL-Projects/blob/main/docs/NeurIPS-2019-TAS.md
# - FBNetV2: https://github.com/facebookresearch/mobile-vision
# - TuNAS: https://github.com/google-research/google-research/tree/master/tunas
####

2
lib/nas_201_api/api_201.py
View File

@ -244,7 +244,7 @@ class NASBench201API(NASBenchMetaAPI):
arch_str: the input is a string indicates the architecture topology, such as
|nor_conv_1x1~0|+|none~0|none~1|+|none~0|none~1|skip_connect~2|
search_space: a list of operation string, the default list is the search space for NAS-Bench-201
the default value should be be consistent with this line https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/models/cell_operations.py#L24
the default value should be be consistent with this line https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/models/cell_operations.py#L24
:return
the numpy matrix (2-D np.ndarray) representing the DAG of this architecture topology
:usage

2
lib/nats_bench/api_topology.py
View File

@ -306,7 +306,7 @@ class NATStopology(NASBenchMetaAPI):
arch_str: the input is a string indicates the architecture topology, such as
|nor_conv_1x1~0|+|none~0|none~1|+|none~0|none~1|skip_connect~2|
search_space: a list of operation string, the default list is the topology search space for NATS-BENCH.
the default value should be be consistent with this line https://github.com/D-X-Y/AutoDL-Projects/blob/master/lib/models/cell_operations.py#L24
the default value should be be consistent with this line https://github.com/D-X-Y/AutoDL-Projects/blob/main/lib/models/cell_operations.py#L24
Returns:
the numpy matrix (2-D np.ndarray) representing the DAG of this architecture topology