Do we have any other scheduler in ds_config?

[128Ghe980]

I find WarmupCosineLR and WarmupDecayLR in https://deepspeed.readthedocs.io/en/latest/schedulers.html#warmupdecaylr
And i use these two in ds_config, but these two dont work
Do we have any other scheduler in ds_config?

[tjruwase]

@128Ghe980, can you please clarify what you mean by don't work?

You might find these tutorials useful.

https://www.deepspeed.ai/tutorials/one-cycle/
https://www.deepspeed.ai/tutorials/lrrt/

[128Ghe980]

@128Ghe980, can you please clarify what you mean by don't work?

You might find these tutorials useful.

https://www.deepspeed.ai/tutorials/one-cycle/ https://www.deepspeed.ai/tutorials/lrrt/

error:

Traceback (most recent call last):
  File "/home/tione/notebook/code/Qwen1_5/finetune.py", line 376, in <module>
    train()
  File "/home/tione/notebook/code/Qwen1_5/finetune.py", line 367, in train
    trainer.train()
  File "/usr/local/lib/python3.10/dist-packages/transformers/trainer.py", line 1539, in train
    return inner_training_loop(
  File "/usr/local/lib/python3.10/dist-packages/transformers/trainer.py", line 1690, in _inner_training_loop
    model, self.optimizer, self.lr_scheduler = self.accelerator.prepare(
  File "/usr/local/lib/python3.10/dist-packages/accelerate/accelerator.py", line 1198, in prepare
    result = self._prepare_deepspeed(*args)
  File "/usr/local/lib/python3.10/dist-packages/accelerate/accelerator.py", line 1537, in _prepare_deepspeed
    engine, optimizer, _, lr_scheduler = deepspeed.initialize(**kwargs)
  File "/usr/local/lib/python3.10/dist-packages/deepspeed/__init__.py", line 171, in initialize
    engine = DeepSpeedEngine(args=args,
  File "/usr/local/lib/python3.10/dist-packages/deepspeed/runtime/engine.py", line 311, in __init__
    self._configure_lr_scheduler(lr_scheduler)
  File "/usr/local/lib/python3.10/dist-packages/deepspeed/runtime/engine.py", line 874, in _configure_lr_scheduler
    lr_scheduler = self._scheduler_from_config(self.optimizer)
  File "/usr/local/lib/python3.10/dist-packages/deepspeed/runtime/engine.py", line 924, in _scheduler_from_config
    assert hasattr(torch.optim.lr_scheduler,
AssertionError: DeepSpeed does not recognize LR scheduler WarmupCosineLR

my ds_config:

{
    "fp16": {
        "enabled": "auto",
        "loss_scale": 0,
        "loss_scale_window": 1000,
        "initial_scale_power": 16,
        "hysteresis": 2,
        "min_loss_scale": 1
    },
    "bf16": {
        "enabled": "auto"
    },
    "optimizer": {
        "type": "AdamW",
        "params": {
            "lr": "auto",
            "betas": "auto",
            "eps": "auto",
            "weight_decay": "auto"
        }
    },

    "scheduler": {
        "type": "WarmupCosineLR",
        "params": {
            "warmup_min_lr": "auto",
            "warmup_max_lr": "auto",
            "warmup_num_steps": "auto"
        }
    },

    "zero_optimization": {
        "stage": 3,
        "offload_optimizer": {
            "device": "none",
            "pin_memory": true
        },
        "offload_param": {
            "device": "none",
            "pin_memory": true
        },
        "overlap_comm": true,
        "contiguous_gradients": true,
        "sub_group_size": 1e9,
        "reduce_bucket_size": "auto",
        "stage3_prefetch_bucket_size": "auto",
        "stage3_param_persistence_threshold": "auto",
        "stage3_max_live_parameters": 1e9,
        "stage3_max_reuse_distance": 1e9,
        "stage3_gather_16bit_weights_on_model_save": true
    },

    "gradient_accumulation_steps": "auto",
    "gradient_clipping": "auto",
    "steps_per_print": 100,
    "train_batch_size": "auto",
    "train_micro_batch_size_per_gpu": "auto",
    "wall_clock_breakdown": false
}

[tjruwase]

assert hasattr(torch.optim.lr_scheduler,
AssertionError: DeepSpeed does not recognize LR scheduler WarmupCosineLR

This is a bug. Can you please file a ticket? Thanks!