@ -6,7 +6,7 @@ MindSpore lite is a high-performance, lightweight open source reasoning framewor
<imgsrc="../../docs/MindSpore-Lite-architecture.png"alt="MindSpore Lite Architecture"width="600"/>
For more details please check out our [MindSpore Lite Architecture Guide](https://www.mindspore.cn/lite/docs/en/master/architecture.html).
For more details please check out our [MindSpore Lite Architecture Guide](https://www.mindspore.cn/doc/note/en/master/design/mindspore/architecture_lite.html).
### MindSpore Lite features
@ -41,7 +41,7 @@ For more details please check out our [MindSpore Lite Architecture Guide](https:
2. Model converter and optimization
If you use MindSpore or a third-party model, you need to use [MindSpore Lite Model Converter Tool](https://www.mindspore.cn/lite/tutorial/en/master/use/converter_tool.html) to convert the model into MindSpore Lite model. The MindSpore Lite model converter tool provides the converter of TensorFlow Lite, Caffe, ONNX to MindSpore Lite model, fusion and quantization could be introduced during convert procedure.
If you use MindSpore or a third-party model, you need to use [MindSpore Lite Model Converter Tool](https://www.mindspore.cn/tutorial/lite/en/master/use/convert_model.html) to convert the model into MindSpore Lite model. The MindSpore Lite model converter tool provides the converter of TensorFlow Lite, Caffe, ONNX to MindSpore Lite model, fusion and quantization could be introduced during convert procedure.
MindSpore also provides a tool to convert models running on IoT devices .
@ -51,7 +51,7 @@ For more details please check out our [MindSpore Lite Architecture Guide](https:
4. Inference
Load the model and perform inference. [Inference](https://www.mindspore.cn/lite/tutorial/en/master/use/runtime.html) is the process of running input data through the model to get output.
Load the model and perform inference. [Inference](https://www.mindspore.cn/tutorial/lite/en/master/use/runtime.html) is the process of running input data through the model to get output.
MindSpore provides pre-trained model that can be deployed on mobile device [example](https://www.mindspore.cn/lite/examples/en).
@ -41,7 +41,7 @@ MDP requires MindSpore version 0.7.0-beta or later. MDP is actively evolving. In
### Tutorial
**Bayesian Neural Network**
1. Process the required dataset. The MNIST dateset is used in the example. Data processing is consistent with [Implementing an Image Classification Application](https://www.mindspore.cn/tutorial/en/master/quick_start/quick_start.html) in Tutorial.
1. Process the required dataset. The MNIST dateset is used in the example. Data processing is consistent with [Implementing an Image Classification Application](https://www.mindspore.cn/tutorial/training/en/master/quick_start/quick_start.html) in Tutorial.
2. Define a Bayesian Neural Network. The bayesian LeNet is used in this example.
3. Process the required dataset. The MNIST dateset is used in the example. Data processing is consistent with [Implementing an Image Classification Application](https://www.mindspore.cn/tutorial/en/master/quick_start/quick_start.html) in Tutorial.
3. Process the required dataset. The MNIST dateset is used in the example. Data processing is consistent with [Implementing an Image Classification Application](https://www.mindspore.cn/tutorial/training/en/master/quick_start/quick_start.html) in Tutorial.
4. Use SVI interface to train VAE network. vi.run can return the trained network, get_train_loss can get the loss after training.
```
@ -423,7 +423,7 @@ if __name__ == "__main__":
**Uncertainty Evaluation**
The uncertainty estimation toolbox is based on MindSpore Deep Probabilistic Programming (MDP), and it is suitable for mainstream deep learning models, such as regression, classification, target detection and so on. In the inference stage, with the uncertainy estimation toolbox, developers only need to pass in the trained model and training dataset, specify the task and the samples to be estimated, then can obtain the aleatoric uncertainty and epistemic uncertainty. Based the uncertainty information, developers can understand the model and the dataset better.
In classification task, for example, the model is lenet model. The MNIST dateset is used in the example. Data processing is consistent with [Implementing an Image Classification Application](https://www.mindspore.cn/tutorial/en/master/quick_start/quick_start.html) in Tutorial. For evaluating the uncertainty of test examples, the use of the toolbox is as follows:
In classification task, for example, the model is lenet model. The MNIST dateset is used in the example. Data processing is consistent with [Implementing an Image Classification Application](https://www.mindspore.cn/tutorial/training/en/master/quick_start/quick_start.html) in Tutorial. For evaluating the uncertainty of test examples, the use of the toolbox is as follows:
```
from mindspore.nn.probability.toolbox.uncertainty_evaluation import UncertaintyEvaluation
from mindspore.train.serialization import load_checkpoint, load_param_into_net
@ -57,7 +57,7 @@ The default configuration of the Dataset are as follows:
## Mixed Precision
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
@ -69,8 +69,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
@ -67,8 +67,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
@ -334,7 +334,7 @@ Parameters for both training and evaluation can be set in config.py
## [How to use](#contents)
### Inference
If you need to use the trained model to perform inference on multiple hardware platforms, such as GPU, Ascend 910 or Ascend 310, you can refer to this [Link](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/network_migration.html). Following the steps below, this is a simple example:
If you need to use the trained model to perform inference on multiple hardware platforms, such as GPU, Ascend 910 or Ascend 310, you can refer to this [Link](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/migrate_3rd_scripts.html). Following the steps below, this is a simple example:
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
@ -56,8 +56,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
@ -132,7 +132,7 @@ sh run_distribute_train.sh RANK_TABLE_FILE DATA_PATH
sh run_standalone_train.sh DEVICE_ID DATA_PATH
```
> Notes:
RANK_TABLE_FILE can refer to [Link](https://www.mindspore.cn/tutorial/en/master/advanced_use/distributed_training_ascend.html) , and the device_ip can be got as [Link]https://gitee.com/mindspore/mindspore/tree/master/model_zoo/utils/hccl_tools.
RANK_TABLE_FILE can refer to [Link](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/distributed_training_ascend.html) , and the device_ip can be got as [Link]https://gitee.com/mindspore/mindspore/tree/master/model_zoo/utils/hccl_tools.
> This is processor cores binding operation regarding the `device_num` and total processor numbers. If you are not expect to do it, remove the operations `taskset` in `scripts/run_distribute_train.sh`
@ -307,7 +307,7 @@ Usage: sh run_standalone_train.sh [PRETRAINED_MODEL]
## [Training Process](#contents)
- Set options in `config.py`, including loss_scale, learning rate and network hyperparameters. Click [here](https://www.mindspore.cn/tutorial/zh-CN/master/use/data_preparation/loading_the_datasets.html#mindspore) for more information about dataset.
- Set options in `config.py`, including loss_scale, learning rate and network hyperparameters. Click [here](https://www.mindspore.cn/tutorial/training/zh-CN/master/use/data_preparation.html) for more information about dataset.
### [Training](#content)
- Run `run_standalone_train.sh` for non-distributed training of MaskRCNN model.
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
# [Environment Requirements](#contents)
@ -53,8 +53,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
# [Environment Requirements](#contents)
@ -57,8 +57,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
@ -43,10 +43,10 @@ A testing set containing about 2000 readable words
- Hardware（Ascend）
- Prepare hardware environment with Ascend processor. If you want to try Ascend , please send the [application form](https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/file/other/Ascend%20Model%20Zoo%E4%BD%93%E9%AA%8C%E8%B5%84%E6%BA%90%E7%94%B3%E8%AF%B7%E8%A1%A8.docx) to ascend@huawei.com. Once approved, you can get the resources.
If you need to use the trained model to perform inference on multiple hardware platforms, such as GPU, Ascend 910 or Ascend 310, you can refer to this [Link](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/network_migration.html). Following the steps below, this is a simple example:
If you need to use the trained model to perform inference on multiple hardware platforms, such as GPU, Ascend 910 or Ascend 310, you can refer to this [Link](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/migrate_3rd_scripts.html). Following the steps below, this is a simple example:
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data types, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data types, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
# [Environment Requirements](#contents)
@ -82,8 +82,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
# [Environment Requirements](#contents)
@ -56,8 +56,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
@ -58,8 +58,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
@ -147,7 +147,7 @@ sh run_eval.sh [DATASET] [CHECKPOINT_PATH] [DEVICE_ID]
### Training on Ascend
To train the model, run `train.py`. If the `mindrecord_dir` is empty, it will generate [mindrecord](https://www.mindspore.cn/tutorial/en/master/use/data_preparation/converting_datasets.html) files by `coco_root`(coco dataset) or `iamge_dir` and `anno_path`(own dataset). **Note if mindrecord_dir isn't empty, it will use mindrecord_dir instead of raw images.**
To train the model, run `train.py`. If the `mindrecord_dir` is empty, it will generate [mindrecord](https://www.mindspore.cn/tutorial/training/zh-CN/master/advanced_use/converse_datasets.html) files by `coco_root`(coco dataset) or `iamge_dir` and `anno_path`(own dataset). **Note if mindrecord_dir isn't empty, it will use mindrecord_dir instead of raw images.**
@ -235,7 +235,7 @@ step: 300, loss is 0.18949677, fps is 57.63118508760329
## [How to use](#contents)
### Inference
If you need to use the trained model to perform inference on multiple hardware platforms, such as Ascend 910 or Ascend 310, you can refer to this [Link](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/network_migration.html). Following the steps below, this is a simple example:
If you need to use the trained model to perform inference on multiple hardware platforms, such as Ascend 910 or Ascend 310, you can refer to this [Link](https://www.mindspore.cn/tutorial/training/master/advanced_use/migrate_3rd_scripts.html). Following the steps below, this is a simple example:
@ -78,7 +78,7 @@ here basic modules mainly include basic operation like: **3×3 conv** and **2×
## Mixed Precision
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
@ -90,8 +90,8 @@ For FP16 operators, if the input data type is FP32, the backend of MindSpore wil
@ -290,7 +290,7 @@ train_parallel1/log:epcoh: 2 step: 97, loss is 1.7133579
...
...
```
> About rank_table.json, you can refer to the [distributed training tutorial](https://www.mindspore.cn/tutorial/en/master/advanced_use/distributed_training.html).
> About rank_table.json, you can refer to the [distributed training tutorial](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/distributed_training_tutorials.html).
> **Attention** This will bind the processor cores according to the `device_num` and total processor numbers. If you don't expect to run pretraining with binding processor cores, remove the operations about `taskset` in `scripts/run_distribute_train.sh`
Some files were not shown because too many files have changed in this diff
Show More
Ting Wang
4 years ago