快速开始

备注

阅读本篇前,请确保已按照 安装指南 准备好昇腾环境及 Accelerate

本教程以一个简单的 NLP 模型为例,讲述如何使用 Accelerate 在昇腾 NPU 上进行模型的训练。

前置准备

本篇将使用到 HuggingFace 其他工具链及 scikit-learn 库,请使用以下指令安装:

pip install datasets evaluate transformers scikit-learn -i https://pypi.tuna.tsinghua.edu.cn/simple

本篇样例代码为 Accelrate 官方样例,需提前进行下载

git clone https://github.com/huggingface/accelerate.git

模型训练

1# 替换HF域名,方便国内用户进行数据及模型的下载
2export HF_ENDPOINT=https://hf-mirror.com
3# 进入项目目录
4cd accelerate/examples
5# 模型训练
6python nlp_example.py

出现如下日志代表训练成功:

Downloading builder script: 5.75kB [00:00, 7.99kB/s]
tokenizer_config.json: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 49.0/49.0 [00:00<00:00, 345kB/s]
config.json: 570B [00:00, 3.12MB/s]
vocab.txt: 213kB [00:00, 477kB/s]
tokenizer.json: 436kB [00:06, 69.4kB/s]
README.md: 35.3kB [00:00, 67.4kB/s]
train-00000-of-00001.parquet: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 649k/649k [00:00<00:00, 6.23MB/s]
validation-00000-of-00001.parquet: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 75.7k/75.7k [00:00<00:00, 142MB/s]
test-00000-of-00001.parquet: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 308k/308k [00:00<00:00, 39.0MB/s]
Generating train split: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3668/3668 [00:00<00:00, 247461.91 examples/s]
Generating validation split: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 408/408 [00:00<00:00, 128262.33 examples/s]
Generating test split: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1725/1725 [00:00<00:00, 369499.74 examples/s]
Map: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3668/3668 [00:00<00:00, 13791.03 examples/s]
Map: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 408/408 [00:00<00:00, 11843.97 examples/s]
Map: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1725/1725 [00:00<00:00, 14229.78 examples/s]
model.safetensors: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 436M/436M [00:06<00:00, 66.7MB/s]
Some weights of BertForSequenceClassification were not initialized from the model checkpoint at bert-base-cased and are newly initialized: ['classifier.bias', 'classifier.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
You're using a BertTokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.
=====  tensor([1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 1,
        1, 0, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([0, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1,
        0, 1, 1, 1, 0, 0, 1, 0], device='npu:0')
=====  tensor([1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0,
        1, 1, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1,
        1, 0, 0, 1, 1, 1, 1, 0], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1,
        0, 1, 0, 1, 1, 0, 0, 1], device='npu:0')
=====  tensor([1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1,
        0, 1, 0, 0, 0, 1, 1, 0], device='npu:0')
=====  tensor([0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1,
        0, 1, 0, 1, 0, 1, 1, 0], device='npu:0')
=====  tensor([1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0,
        0, 0, 1, 1, 0, 1, 0, 0], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 0,
        1, 0, 0, 0, 1, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 1,
        0, 0, 1, 0, 0, 1, 0, 0], device='npu:0')
=====  tensor([0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 0,
        1, 1, 1, 1, 0, 1, 0, 1], device='npu:0')
=====  tensor([0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0,
        1, 0, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 1],
      device='npu:0')
epoch 0: {'accuracy': 0.7990196078431373, 'f1': 0.8392156862745098}
=====  tensor([1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 0, 1,
        1, 0, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1,
        0, 1, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1,
        1, 1, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1,
        1, 0, 0, 1, 1, 1, 1, 0], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1,
        1, 1, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1,
        1, 1, 0, 1, 0, 1, 1, 0], device='npu:0')
=====  tensor([0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1,
        1, 1, 0, 1, 0, 1, 1, 0], device='npu:0')
=====  tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0,
        1, 1, 1, 0, 1, 1, 0, 0], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0,
        1, 1, 0, 0, 0, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1,
        1, 0, 1, 0, 0, 1, 0, 0], device='npu:0')
=====  tensor([0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 0,
        1, 1, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1,
        1, 0, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1],
      device='npu:0')
epoch 1: {'accuracy': 0.8602941176470589, 'f1': 0.8998242530755711}
=====  tensor([1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 0, 1,
        1, 0, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1,
        0, 1, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0,
        1, 1, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1,
        1, 0, 0, 1, 1, 1, 1, 0], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 1,
        1, 1, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1,
        1, 1, 0, 1, 0, 1, 1, 0], device='npu:0')
=====  tensor([0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1,
        1, 1, 0, 1, 1, 1, 1, 0], device='npu:0')
=====  tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0,
        1, 1, 1, 0, 1, 1, 0, 0], device='npu:0')
=====  tensor([1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0,
        1, 1, 1, 0, 0, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1,
        1, 0, 1, 0, 1, 1, 0, 0], device='npu:0')
=====  tensor([0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 0,
        1, 1, 1, 1, 1, 1, 0, 1], device='npu:0')
=====  tensor([1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1,
        1, 0, 1, 1, 1, 1, 1, 1], device='npu:0')
=====  tensor([1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0, 1],
      device='npu:0')
epoch 2: {'accuracy': 0.8627450980392157, 'f1': 0.903114186851211}