shibing624 / MedicalGPT

MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training Pipeline. 训练医疗大模型,实现了包括增量预训练(PT)、有监督微调(SFT)、RLHF、DPO、ORPO。
Apache License 2.0
3.24k stars 492 forks source link

大佬能帮我看看loss吗,Train 的loss一直在下降,eval的loss 触底猛反弹 #260

Closed SoYuCry closed 10 months ago

SoYuCry commented 11 months ago

Describe the Question

Please provide a clear and concise description of what the question is. 如题,我应该选择哪一个位置的权重呢?谢谢大佬! image

shibing624 commented 11 months ago

loss只是参考,选测试得分最高的checkpoint。