The following values were not passed to `accelerate launch` and had defaults used instead: More than one GPU was found, enabling multi-GPU training. If this was unintended please pass in `--num_processes=1`. `--dynamo_backend` was set to a value of `'no'` To avoid this warning pass in values for each of the problematic parameters or run `accelerate config`. [2024-12-30 11:39:23,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) W1230 11:39:28.455000 139686276646720 torch/distributed/run.py:779] W1230 11:39:28.455000 139686276646720 torch/distributed/run.py:779] ***************************************** W1230 11:39:28.455000 139686276646720 torch/distributed/run.py:779] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. W1230 11:39:28.455000 139686276646720 torch/distributed/run.py:779] ***************************************** [2024-12-30 11:39:43,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,133] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,269] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:44,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-30 11:39:46,252] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-30 11:39:46,389] [INFO] [comm.py:652:init_distributed] cdb=None 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 7 Local process index: 7 Device: cuda:7 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} [2024-12-30 11:39:46,427] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-30 11:39:46,466] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-30 11:39:46,497] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-30 11:39:46,515] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-30 11:39:46,532] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-30 11:39:46,532] [INFO] [comm.py:683:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl [2024-12-30 11:39:46,550] [INFO] [comm.py:652:init_distributed] cdb=None 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 6 Local process index: 6 Device: cuda:6 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} Detected kernel version 4.19.91, which is below the recommended minimum of 5.5.0; this can cause the process to hang. It is recommended to upgrade the kernel to the minimum version or higher. 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 0 Local process index: 0 Device: cuda:0 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 1 Local process index: 1 Device: cuda:1 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 2 Local process index: 2 Device: cuda:2 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 3 Local process index: 3 Device: cuda:3 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 5 Local process index: 5 Device: cuda:5 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} 12/30/2024 11:39:46 - INFO - __main__ - Distributed environment: DEEPSPEED Backend: nccl Num processes: 8 Process index: 4 Local process index: 4 Device: cuda:4 Mixed precision type: bf16 ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'zero_allow_untested_optimizer': True, 'bf16': {'enabled': True}, 'zero_optimization': {'stage': 2, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 'auto', 'contiguous_gradients': True, 'round_robin_gradients': True}, 'steps_per_print': inf, 'fp16': {'enabled': False}} Loading checkpoint shards: 0%| | 0/4 [00:00 [2024-12-30 11:42:23,739] [INFO] [logging.py:128:log_dist] [Rank 0] Creating torch.bfloat16 ZeRO stage 2 optimizer [2024-12-30 11:42:23,739] [INFO] [stage_1_and_2.py:149:__init__] Reduce bucket size 16777216 [2024-12-30 11:42:23,739] [INFO] [stage_1_and_2.py:150:__init__] Allgather bucket size 200000000 [2024-12-30 11:42:23,739] [INFO] [stage_1_and_2.py:151:__init__] CPU Offload: False [2024-12-30 11:42:23,739] [INFO] [stage_1_and_2.py:152:__init__] Round robin gradient partitioning: True [2024-12-30 11:42:37,700] [INFO] [utils.py:781:see_memory_usage] Before initializing optimizer states [2024-12-30 11:42:37,701] [INFO] [utils.py:782:see_memory_usage] MA 18.7 GB Max_MA 20.57 GB CA 20.57 GB Max_CA 21 GB [2024-12-30 11:42:37,701] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 33.41 GB, percent = 6.7% [2024-12-30 11:42:37,867] [INFO] [utils.py:781:see_memory_usage] After initializing optimizer states [2024-12-30 11:42:37,868] [INFO] [utils.py:782:see_memory_usage] MA 18.7 GB Max_MA 22.44 GB CA 24.31 GB Max_CA 24 GB [2024-12-30 11:42:37,868] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 33.41 GB, percent = 6.7% [2024-12-30 11:42:37,868] [INFO] [stage_1_and_2.py:544:__init__] optimizer state initialized [2024-12-30 11:42:38,018] [INFO] [utils.py:781:see_memory_usage] After initializing ZeRO optimizer [2024-12-30 11:42:38,019] [INFO] [utils.py:782:see_memory_usage] MA 18.7 GB Max_MA 18.7 GB CA 24.31 GB Max_CA 24 GB [2024-12-30 11:42:38,019] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 33.41 GB, percent = 6.7% [2024-12-30 11:42:38,020] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed Final Optimizer = DeepSpeedZeroOptimizer [2024-12-30 11:42:38,020] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed using configured LR scheduler = None [2024-12-30 11:42:38,020] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed LR Scheduler = None [2024-12-30 11:42:38,020] [INFO] [logging.py:128:log_dist] [Rank 0] step=0, skipped=0, lr=[0.0], mom=[(0.9, 0.999)] [2024-12-30 11:42:38,021] [INFO] [config.py:999:print] DeepSpeedEngine configuration: [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] activation_checkpointing_config { "partition_activations": false, "contiguous_memory_optimization": false, "cpu_checkpointing": false, "number_checkpoints": null, "synchronize_checkpoint_boundary": false, "profile": false } [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'thread_count': 1, 'single_submit': False, 'overlap_events': True, 'use_gds': False} [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] amp_enabled .................. False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] amp_params ................... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] autotuning_config ............ { "enabled": false, "start_step": null, "end_step": null, "metric_path": null, "arg_mappings": null, "metric": "throughput", "model_info": null, "results_dir": "autotuning_results", "exps_dir": "autotuning_exps", "overwrite": true, "fast": true, "start_profile_step": 3, "end_profile_step": 5, "tuner_type": "gridsearch", "tuner_early_stopping": 5, "tuner_num_trials": 50, "model_info_path": null, "mp_size": 1, "max_train_batch_size": null, "min_train_batch_size": 1, "max_train_micro_batch_size_per_gpu": 1.024000e+03, "min_train_micro_batch_size_per_gpu": 1, "num_tuning_micro_batch_sizes": 3 } [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] bfloat16_enabled ............. True [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] bfloat16_immediate_grad_update False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] checkpoint_parallel_write_pipeline False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] checkpoint_tag_validation_enabled True [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] checkpoint_tag_validation_fail False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] comms_config ................. [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] communication_data_type ...... None [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] compression_config ........... {'weight_quantization': {'shared_parameters': {'enabled': False, 'quantizer_kernel': False, 'schedule_offset': 0, 'quantize_groups': 1, 'quantize_verbose': False, 'quantization_type': 'symmetric', 'quantize_weight_in_forward': False, 'rounding': 'nearest', 'fp16_mixed_quantize': False, 'quantize_change_ratio': 0.001}, 'different_groups': {}}, 'activation_quantization': {'shared_parameters': {'enabled': False, 'quantization_type': 'symmetric', 'range_calibration': 'dynamic', 'schedule_offset': 1000}, 'different_groups': {}}, 'sparse_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'row_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'head_pruning': {'shared_parameters': {'enabled': False, 'method': 'topk', 'schedule_offset': 1000}, 'different_groups': {}}, 'channel_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'layer_reduction': {'enabled': False}} [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] curriculum_enabled_legacy .... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] curriculum_params_legacy ..... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] data_efficiency_config ....... {'enabled': False, 'seed': 1234, 'data_sampling': {'enabled': False, 'num_epochs': 1000, 'num_workers': 0, 'curriculum_learning': {'enabled': False}}, 'data_routing': {'enabled': False, 'random_ltd': {'enabled': False, 'layer_token_lr_schedule': {'enabled': False}}}} [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] data_efficiency_enabled ...... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] dataloader_drop_last ......... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] disable_allgather ............ False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] dump_state ................... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] dynamic_loss_scale_args ...... None [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] eigenvalue_enabled ........... False [2024-12-30 11:42:38,022] [INFO] [config.py:1003:print] eigenvalue_gas_boundary_resolution 1 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] eigenvalue_layer_name ........ bert.encoder.layer [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] eigenvalue_layer_num ......... 0 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] eigenvalue_max_iter .......... 100 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] eigenvalue_stability ......... 1e-06 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] eigenvalue_tol ............... 0.01 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] eigenvalue_verbose ........... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] elasticity_enabled ........... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] flops_profiler_config ........ { "enabled": false, "recompute_fwd_factor": 0.0, "profile_step": 1, "module_depth": -1, "top_modules": 1, "detailed": true, "output_file": null } [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] fp16_auto_cast ............... None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] fp16_enabled ................. False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] fp16_master_weights_and_gradients False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] global_rank .................. 0 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] grad_accum_dtype ............. None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] gradient_accumulation_steps .. 4 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] gradient_clipping ............ 1.0 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] gradient_predivide_factor .... 1.0 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] graph_harvesting ............. False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] hybrid_engine ................ enabled=False max_out_tokens=512 inference_tp_size=1 release_inference_cache=False pin_parameters=True tp_gather_partition_size=8 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] initial_dynamic_scale ........ 1 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] load_universal_checkpoint .... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] loss_scale ................... 1.0 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] memory_breakdown ............. False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] mics_hierarchial_params_gather False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] mics_shard_size .............. -1 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] monitor_config ............... tensorboard=TensorBoardConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') comet=CometConfig(enabled=False, samples_log_interval=100, project=None, workspace=None, api_key=None, experiment_name=None, experiment_key=None, online=None, mode=None) wandb=WandbConfig(enabled=False, group=None, team=None, project='deepspeed') csv_monitor=CSVConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] nebula_config ................ { "enabled": false, "persistent_storage_path": null, "persistent_time_interval": 100, "num_of_version_in_retention": 2, "enable_nebula_load": true, "load_path": null } [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] optimizer_legacy_fusion ...... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] optimizer_name ............... None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] optimizer_params ............. None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True} [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] pld_enabled .................. False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] pld_params ................... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] prescale_gradients ........... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] scheduler_name ............... None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] scheduler_params ............. None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] seq_parallel_communication_data_type torch.float32 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] sparse_attention ............. None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] sparse_gradients_enabled ..... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] steps_per_print .............. inf [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] timers_config ................ enabled=True synchronized=True [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] train_batch_size ............. 32 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] train_micro_batch_size_per_gpu 1 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] use_data_before_expert_parallel_ False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] use_node_local_storage ....... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] wall_clock_breakdown ......... False [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] weight_quantization_config ... None [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] world_size ................... 8 [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] zero_allow_untested_optimizer True [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] zero_config .................. stage=2 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=16777216 use_multi_rank_bucket_allreduce=True allgather_partitions=True allgather_bucket_size=200000000 overlap_comm=True load_from_fp32_weights=True elastic_checkpoint=False offload_param=None offload_optimizer=None sub_group_size=1000000000 cpu_offload_param=None cpu_offload_use_pin_memory=None cpu_offload=None prefetch_bucket_size=50000000 param_persistence_threshold=100000 model_persistence_threshold=9223372036854775807 max_live_parameters=1000000000 max_reuse_distance=1000000000 gather_16bit_weights_on_model_save=False use_all_reduce_for_fetch_params=False stage3_gather_fp16_weights_on_model_save=False ignore_unused_parameters=True legacy_stage1=False round_robin_gradients=True zero_hpz_partition_size=1 zero_quantized_weights=False zero_quantized_nontrainable_weights=False zero_quantized_gradients=False mics_shard_size=-1 mics_hierarchical_params_gather=False memory_efficient_linear=True pipeline_loading_checkpoint=False override_module_apply=True [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] zero_enabled ................. True [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] zero_force_ds_cpu_optimizer .. True [2024-12-30 11:42:38,023] [INFO] [config.py:1003:print] zero_optimization_stage ...... 2 [2024-12-30 11:42:38,024] [INFO] [config.py:989:print_user_config] json = { "train_batch_size": 32, "train_micro_batch_size_per_gpu": 1, "gradient_accumulation_steps": 4, "gradient_clipping": 1.0, "zero_allow_untested_optimizer": true, "bf16": { "enabled": true }, "zero_optimization": { "stage": 2, "allgather_partitions": true, "allgather_bucket_size": 2.000000e+08, "overlap_comm": true, "reduce_scatter": true, "reduce_bucket_size": 1.677722e+07, "contiguous_gradients": true, "round_robin_gradients": true }, "steps_per_print": inf, "fp16": { "enabled": false } } 12/30/2024 11:42:38 - INFO - __main__ - ***** Running training ***** 12/30/2024 11:42:38 - INFO - __main__ - Num examples = 40821 12/30/2024 11:42:38 - INFO - __main__ - Num Epochs = 1 12/30/2024 11:42:38 - INFO - __main__ - Instantaneous batch size per device = 1 12/30/2024 11:42:38 - INFO - __main__ - Total train batch size (w. parallel, distributed & accumulation) = 32 12/30/2024 11:42:38 - INFO - __main__ - Gradient Accumulation steps = 4 12/30/2024 11:42:38 - INFO - __main__ - Total optimization steps = 1276 0%| | 0/1276 [00:00