知识卡片
提供验证集不代表训练会自动评估
Trainer接收eval_dataset只说明有可用材料,实际何时评价仍由训练参数控制。需显式核对评估策略、保存与最佳模型选择,连同优化器及学习率默认值;设置轮数和批量不等于所有训练设计都已确定。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- · 查看原书依据
>>> trainer = Trainer( ... ft_model, ... training_args, ... data_collator=collator, #2 ... train_dataset=train_dataset, #3 ... eval_dataset=eval_dataset) >>> trainer.train() #1 Reloads a fresh pretrained GPT-2 base model #2 Your ΔataCollatorForLanguageModeling configured for left-to-right causal models #3 The training subset of the NLPiAΔataset from torch.random_split This training run can take a couple of hours on a CPU, so if you have access to a GPU (on which
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第10章《真实世界中的大语言模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-10.html)
- 结论依据:摘录中关于提供验证集不代表训练会自动评估的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:>>> trainer = Trainer( ... ft_model, ... training_args, ... data_collator=collator, #2 ... train_dataset=train_dataset, #3 ... eval_dataset=eval_dataset) >>> trainer.train() #1 Reloads a fresh pretrained GPT-2 base model #2 Your ΔataCollatorForLanguageModeling configured for left-to-right causal models #3 The training subset of the NLPiAΔataset from torch.random_split This training run can take a couple of hours on a CPU, so if you have access to a GPU (on which