[BUG] The latest master code doesn't work with pydantic 2.0a2

[ftian1]

Describe the bug when source build latest code on Ubuntu 18.04.3 LTS and run bert pruning sparse example in DeepSpeedExample, you will see crash. from the log, it's because the code in deepspeed/runtime/config_utils.py was changed recently to use pydantic BaseModel.fields which doesn't exist in latest pydantic version.

To Reproduce

1. cd DeepSpeed
2. python setup.py install
3. cd DeepSpeedExamples/compression/bert
4. source ./bash_script/pruning_sparse.sh # port number in sh file is wrong, need manual change.

Expected behavior functionality work

Screenshots

/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/pydantic-2.0a2-py3.8.egg/pydantic/_internal/_config.py:230: UserWarning: Valid config keys have changed in V2:

• 'allow_population_by_field_name' has been renamed to 'populate_by_name'
• 'validate_all' has been renamed to 'validate_default' warnings.warn(message, UserWarning) Traceback (most recent call last): File "run_glue_no_trainer.py", line 49, in from huggingface_transformer.modeling_bert import BertForSequenceClassification File "/home/wenhuach/DeepSpeedExamples/compression/bert/huggingface_transformer/modeling_bert.py", line 51, in from transformers.modeling_utils import ( File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/transformers/modeling_utils.py", line 37, in from .deepspeed import deepspeed_config, is_deepspeed_zero3_enabled File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/transformers/deepspeed.py", line 38, in from accelerate.utils.deepspeed import HfDeepSpeedConfig as DeepSpeedConfig File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/accelerate/init.py", line 7, in from .accelerator import Accelerator File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/accelerate/accelerator.py", line 29, in from .checkpointing import load_accelerator_state, load_custom_state, save_accelerator_state, save_custom_state File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/accelerate/checkpointing.py", line 24, in from .utils import ( File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/accelerate/utils/init.py", line 124, in from .other import ( File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/accelerate/utils/other.py", line 27, in from deepspeed import DeepSpeedEngine File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/init.py", line 16, in from . import module_inject File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/module_inject/init.py", line 6, in from .replace_module import replace_transformer_layer, revert_transformer_layer, ReplaceWithTensorSlicing, GroupQuantizer, generic_injection File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/module_inject/replace_module.py", line 732, in from ..pipe import PipelineModule File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/pipe/init.py", line 6, in from ..runtime.pipe import PipelineModule, LayerSpec, TiedLayerSpec File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/pipe/init.py", line 6, in from .module import PipelineModule, LayerSpec, TiedLayerSpec File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/pipe/module.py", line 19, in from ..activation_checkpointing import checkpointing File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/activation_checkpointing/checkpointing.py", line 25, in from deepspeed.runtime.config import DeepSpeedConfig File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/config.py", line 28, in from .zero.config import get_zero_config, ZeroStageEnum File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/zero/init.py", line 6, in from .partition_parameters import ZeroParamType File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/zero/partition_parameters.py", line 601, in class Init(InsertPostInitMethodToModuleSubClasses): File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/zero/partition_parameters.py", line 603, in Init param_persistence_threshold = get_config_default(DeepSpeedZeroConfig, "param_persistence_threshold") File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/deepspeed-0.9.1+cc67f22f-py3.8.egg/deepspeed/runtime/config_utils.py", line 114, in get_config_default assert field_name in config.fields, f"'{field_name}' is not a field in {config}" AttributeError: type object 'DeepSpeedZeroConfig' has no attribute 'fields' ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 2220051) of binary: /home/wenhuach/anaconda3/envs/torch/bin/python Traceback (most recent call last): File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/torch/distributed/launch.py", line 195, in main() File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/torch/distributed/launch.py", line 191, in main launch(args) File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/torch/distributed/launch.py", line 176, in launch run(args) File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/torch/distributed/run.py", line 753, in run elastic_launch( File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 132, in call return launch_agent(self._config, self._entrypoint, list(args)) File "/home/wenhuach/anaconda3/envs/torch/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 246, in launch_agent raise ChildFailedError( torch.distributed.elastic.multiprocessing.errors.ChildFailedError: ============================================================ run_glue_no_trainer.py FAILED

System info (please complete the following information):

• OS: Ubuntu 18.04
• GPU: NV P40
• Python version: 3.8

[ftian1]

PS: after manually switch pydantic version from 2.0a2 to 1.10.7. the latest code works.

I have a quick glance on that, it's because pydantic 2.0a2 has removed __field__ attributes in BaseModel.

[loadams]

It looks like we're not up to date with the pre-releases from pydantic, but things do work with their latest stable release, 1.10.7. We will work on getting things to build with their pre-release for when they release 2.0.0 officially.