bigscience

/

bloomz-7b1-optimizer-states

like 1

Follow

BigScience Workshop 1.29k

• "collections.OrderedDict",
• "torch.Size",
• "numpy.ndarray",
• "numpy.dtype",
• "_codecs.encode",
• "torch._utils._rebuild_tensor_v2",
• "megatron.enums.PositionEmbeddingType",
• "numpy.core.multiarray._reconstruct",
• "argparse.Namespace",
• "__builtin__.set",
• "torch.float16",
• "torch.ByteStorage"

How to fix it?

• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict",
• "torch.FloatStorage",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum"

How to fix it?

• "torch._utils._rebuild_tensor_v2",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch.FloatStorage",
• "collections.OrderedDict"

How to fix it?

• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "torch.FloatStorage",
• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict",
• "deepspeed.runtime.zero.config.ZeroStageEnum"

How to fix it?

• "torch.FloatStorage",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "collections.OrderedDict",
• "torch._utils._rebuild_tensor_v2",
• "deepspeed.runtime.zero.config.ZeroStageEnum"

How to fix it?

• "torch.FloatStorage",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict"

How to fix it?

• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "torch.FloatStorage"

How to fix it?

• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch.FloatStorage",
• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict"

How to fix it?

• "collections.OrderedDict",
• "torch.FloatStorage",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch._utils._rebuild_tensor_v2",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler"

How to fix it?

• "collections.OrderedDict",
• "torch._utils._rebuild_tensor_v2",
• "torch.FloatStorage",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler"

How to fix it?

• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict",
• "torch.FloatStorage",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum"

How to fix it?

• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "collections.OrderedDict",
• "torch.FloatStorage",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch._utils._rebuild_tensor_v2"

How to fix it?

• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "torch.FloatStorage",
• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict",
• "deepspeed.runtime.zero.config.ZeroStageEnum"

How to fix it?

• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "collections.OrderedDict",
• "torch._utils._rebuild_tensor_v2",
• "torch.FloatStorage",
• "deepspeed.runtime.zero.config.ZeroStageEnum"

How to fix it?

• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "collections.OrderedDict",
• "torch._utils._rebuild_tensor_v2",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "torch.FloatStorage"

How to fix it?

• "torch._utils._rebuild_tensor_v2",
• "collections.OrderedDict",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch.FloatStorage"

How to fix it?

• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "collections.OrderedDict",
• "torch._utils._rebuild_tensor_v2",
• "deepspeed.runtime.fp16.loss_scaler.DynamicLossScaler",
• "torch.FloatStorage"

How to fix it?