训练中被强行终止：args.max_epochs <= math.ceil(state.epoch)，TypeError: '<=' not supported between instances of 'NoneType' and 'int'

[YUANMU227]

[rank0]: File "/home/ms-swift/swift/trainers/callback.py", line 95, in on_epoch_end
[rank0]: if args.max_epochs <= math.ceil(state.epoch):
[rank0]: TypeError: '<=' not supported between instances of 'NoneType' and 'int'
Train: 50%|███████████████████████████████ | 269/538 [18:24<18:24, 4.11s/it]
[rank0]:[W511 00:19:28.037565019 ProcessGroupNCCL.cpp:1496] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())

训练中被强行终止
--num_train_epochs设置为2，已经保存两个文件：checkpoint-100、checkpoint-200

发现最近有更新，是否和这个更新有关？
#4125

正在支持 max_epochs参数，强制在对应epochs时终止训练并保存权重
megatron是不是只支持传入train_iters, 不支持epoch，另外如果使用了packing，怎么看packing后实际有多少样本

[Jintao-Huang]

已经修复了

[YUANMU227]

已经修复了

很强，速度很快啊