Train batch size errors

[lckkkk02]

[rank0]: AssertionError: Check batch related parameters. train_batch_size is not equal to micro_batch_per_gpu * gradient_acc_step * world_size 8 != 2 * 1 * 1
在前面的dist.world_size打印获取=4，但是依然不行

[lx-Meteors]

很明显 wordsize虽然是打印是4，但是系统获取的还是8？ 可以改一下后面的参数啊 例如 2*4 * 1

[loadams]

@lckkkk02 - did changing the parameters work for you? Or can you share your ds_config and code that reproduces the failure and more information about your setup?