CS224n SQuAD2.0 项目数据集

该模型的目标是为默认期末项目建立基准时，节省CS224n学生的GPU时间。用于微调此模型的训练集与官方数据集相同；然而，评估和模型选择使用了官方开发集中随机选取的大约一半样本（共6078个示例）。数据文件可在https://github.com/elgeish/squad/tree/master/data找到——这是2020年冬季版本。由于官方SQuAD2.0开发集包含项目的测试集，学生必须确保不以任何方式使用官方SQuAD2.0开发集——包括使用在官方SQuAD2.0上微调的模型，因为这些模型使用了官方SQuAD2.0开发集进行模型选择。

结果

{
  "exact": 65.16946363935504,
  "f1": 67.87348075352251,
  "total": 6078,
  "HasAns_exact": 69.51890034364261,
  "HasAns_f1": 75.16667217179045,
  "HasAns_total": 2910,
  "NoAns_exact": 61.17424242424242,
  "NoAns_f1": 61.17424242424242,
  "NoAns_total": 3168,
  "best_exact": 65.16946363935504,
  "best_exact_thresh": 0.0,
  "best_f1": 67.87348075352243,
  "best_f1_thresh": 0.0
}

重要参数

{
  "do_lower_case": true,
  "doc_stride": 128,
  "fp16": false,
  "fp16_opt_level": "O1",
  "gradient_accumulation_steps": 24,
  "learning_rate": 3e-05,
  "max_answer_length": 30,
  "max_grad_norm": 1,
  "max_query_length": 64,
  "max_seq_length": 384,
  "model_name_or_path": "distilbert-base-uncased-distilled-squad",
  "model_type": "distilbert",
  "num_train_epochs": 4,
  "per_gpu_train_batch_size": 32,
  "save_steps": 5000,
  "seed": 42,
  "train_batch_size": 32,
  "version_2_with_negative": true,
  "warmup_steps": 0,
  "weight_decay": 0
}

环境配置

{
  "transformers": "2.5.1",
  "pytorch": "1.4.0=py3.6_cuda10.1.243_cudnn7.6.3_0",
  "python": "3.6.5=hc3d631a_2",
  "os": "Linux 4.15.0-1060-aws #62-Ubuntu SMP Tue Feb 11 21:23:22 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux",
  "gpu": "Tesla V100-SXM2-16GB"
}

引用方式

@misc{elgeish2020gestalt,
  title={Gestalt: a Stacking Ensemble for SQuAD2.0},
  author={Mohamed El-Geish},
  journal={arXiv e-prints},
  archivePrefix={arXiv},
  eprint={2004.07067},
  year={2020},
}

相关模型

• elgeish/cs224n-squad2.0-albert-base-v2
• elgeish/cs224n-squad2.0-albert-large-v2
• elgeish/cs224n-squad2.0-albert-xxlarge-v1
• elgeish/cs224n-squad2.0-roberta-base