torch.distributed.elastic.multiprocessing.errors.ChildFailedError:

(hallo) egg@egg:~/ls/hallo/hallo-main$ accelerate launch -m
--config_file accelerate_config.yaml
--machine_rank 0
--main_process_ip 0.0.0.0
--main_process_port 20055
--num_machines 1
--num_processes 1
scripts.train_stage1 --config ./configs/train/stage1.yaml [2025-04-02 20:02:59,612] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2025-04-02 20:02:59,834] [INFO] [comm.py:637:init_distributed] cdb=None [2025-04-02 20:02:59,834] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl 04/02/2025 20:02:59 - INFO - main - Distributed environment: DEEPSPEED Backend: nccl Num processes: 1 Process index: 0 Local process index: 0 Device: cuda:0

Mixed precision type: no ds_config: {'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'gradient_accumulation_steps': 1, 'zero_optimization': {'stage': 2, 'offload_optimizer': {'device': 'none', 'nvme_path': None}, 'offload_param': {'device': 'none', 'nvme_path': None}, 'stage3_gather_16bit_weights_on_model_save': False}, 'steps_per_print': inf, 'fp16': {'enabled': False}, 'bf16': {'enabled': False}}

{'scaling_factor', 'force_upcast', 'latents_mean', 'latents_std'} was not found in config. Values will be initialized to default values. The config attributes {'center_input_sample': False, 'out_channels': 4} were passed to UNet2DConditionModel, but are not expected and will be ignored. Please verify your config.json configuration file. {'mid_block_type', 'upcast_attention', 'encoder_hid_dim_type', 'addition_embed_type', 'use_linear_projection', 'projection_class_embeddings_input_dim', 'resnet_time_scale_shift', 'dual_cross_attention', 'conv_in_kernel', 'class_embed_type', 'mid_block_only_cross_attention', 'encoder_hid_dim', 'addition_time_embed_dim', 'only_cross_attention', 'num_class_embeds', 'time_embedding_act_fn', 'time_embedding_type', '_landmark_net', '_center_input_sample', 'time_embedding_dim', 'timestep_post_act', 'out_channels', 'num_attention_heads', 'attention_type', 'class_embeddings_concat', 'reverse_transformer_layers_per_block', 'time_cond_proj_dim', 'addition_embed_type_num_heads', 'transformer_layers_per_block', 'dropout'} was not found in config. Values will be initialized to default values. Some weights of the model checkpoint were not used when initializing UNet2DConditionModel: ['conv_norm_out.bias, conv_norm_out.weight, conv_out.bias, conv_out.weight'] 04/02/2025 20:03:01 - INFO - hallo.models.unet_3d - loaded temporal unet's pretrained weights from /home/egg/ls/hallo/hallo-main/pretrained_models/stable-diffusion-v1-5/unet ... The config attributes {'center_input_sample': False} were passed to UNet3DConditionModel, but are not expected and will be ignored. Please verify your config.json configuration file. {'use_inflated_groupnorm', 'motion_module_decoder_only', 'use_audio_module', 'upcast_attention', 'motion_module_mid_block', 'motion_module_type', 'only_cross_attention', 'use_linear_projection', 'num_class_embeds', 'resnet_time_scale_shift', 'dual_cross_attention', 'motion_module_resolutions', 'stack_enable_blocks_name', 'unet_use_cross_frame_attention', 'class_embed_type', 'audio_attention_dim', 'motion_module_kwargs', 'stack_enable_blocks_depth'} was not found in config. Values will be initialized to default values. 04/02/2025 20:03:08 - INFO - hallo.models.unet_3d - Loaded 0.0M-parameter motion module [2025-04-02 20:03:09,648] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed info: version=0.13.5, git-hash=unknown, git-branch=unknown [2025-04-02 20:03:10,047] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False [2025-04-02 20:03:10,051] [INFO] [logging.py:96:log_dist] [Rank 0] Using client Optimizer as basic optimizer [2025-04-02 20:03:10,051] [INFO] [logging.py:96:log_dist] [Rank 0] Removing param_group that has no 'params' in the basic Optimizer [2025-04-02 20:03:10,321] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed Basic Optimizer = AdamW [2025-04-02 20:03:10,321] [INFO] [utils.py:56:is_zero_supported_optimizer] Checking ZeRO support for optimizer=AdamW type=<class 'torch.optim.adamw.AdamW'> [2025-04-02 20:03:10,321] [INFO] [logging.py:96:log_dist] [Rank 0] Creating torch.float32 ZeRO stage 2 optimizer [2025-04-02 20:03:10,321] [INFO] [stage_1_and_2.py:149:init] Reduce bucket size 500,000,000 [2025-04-02 20:03:10,321] [INFO] [stage_1_and_2.py:150:init] Allgather bucket size 500,000,000 [2025-04-02 20:03:10,321] [INFO] [stage_1_and_2.py:151:init] CPU Offload: False [2025-04-02 20:03:10,321] [INFO] [stage_1_and_2.py:152:init] Round robin gradient partitioning: False [2025-04-02 20:03:13,796] [INFO] [utils.py:800:see_memory_usage] Before initializing optimizer states [2025-04-02 20:03:13,797] [INFO] [utils.py:801:see_memory_usage] MA 9.78 GB Max_MA 12.98 GB CA 13.04 GB Max_CA 13 GB [2025-04-02 20:03:13,797] [INFO] [utils.py:808:see_memory_usage] CPU Virtual Memory: used = 6.84 GB, percent = 5.4% [2025-04-02 20:03:14,000] [INFO] [utils.py:800:see_memory_usage] After initializing optimizer states [2025-04-02 20:03:14,000] [INFO] [utils.py:801:see_memory_usage] MA 9.78 GB Max_MA 16.19 GB CA 19.45 GB Max_CA 19 GB [2025-04-02 20:03:14,001] [INFO] [utils.py:808:see_memory_usage] CPU Virtual Memory: used = 6.84 GB, percent = 5.4% [2025-04-02 20:03:14,001] [INFO] [stage_1_and_2.py:539:init] optimizer state initialized [2025-04-02 20:03:14,235] [INFO] [utils.py:800:see_memory_usage] After initializing ZeRO optimizer [2025-04-02 20:03:14,235] [INFO] [utils.py:801:see_memory_usage] MA 9.78 GB Max_MA 9.78 GB CA 19.45 GB Max_CA 19 GB [2025-04-02 20:03:14,236] [INFO] [utils.py:808:see_memory_usage] CPU Virtual Memory: used = 6.84 GB, percent = 5.4% [2025-04-02 20:03:14,267] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed Final Optimizer = AdamW [2025-04-02 20:03:14,267] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed using client LR scheduler [2025-04-02 20:03:14,267] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed LR Scheduler = None [2025-04-02 20:03:14,267] [INFO] [logging.py:96:log_dist] [Rank 0] step=0, skipped=0, lr=[1e-05], mom=[(0.9, 0.999)] [2025-04-02 20:03:14,270] [INFO] [config.py:996:print] DeepSpeedEngine configuration: [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] activation_checkpointing_config { "partition_activations": false, "contiguous_memory_optimization": false, "cpu_checkpointing": false, "number_checkpoints": null, "synchronize_checkpoint_boundary": false, "profile": false } [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'thread_count': 1, 'single_submit': False, 'overlap_events': True} [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] amp_enabled .................. False [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] amp_params ................... False [2025-04-02 20:03:14,271] [INFO] [config.py:1000: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 } [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] bfloat16_enabled ............. False [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] bfloat16_immediate_grad_update False [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] checkpoint_parallel_write_pipeline False [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] checkpoint_tag_validation_enabled True [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] checkpoint_tag_validation_fail False [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] comms_config ................. <deepspeed.comm.config.DeepSpeedCommsConfig object at 0x769c28da1d80> [2025-04-02 20:03:14,271] [INFO] [config.py:1000:print] communication_data_type ...... None [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] compile_config ............... enabled=False backend='inductor' kwargs={} [2025-04-02 20:03:14,272] [INFO] [config.py:1000: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}} [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] curriculum_enabled_legacy .... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] curriculum_params_legacy ..... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000: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}}}} [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] data_efficiency_enabled ...... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] dataloader_drop_last ......... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] disable_allgather ............ False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] dump_state ................... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] dynamic_loss_scale_args ...... None [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_enabled ........... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_gas_boundary_resolution 1 [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_layer_name ........ bert.encoder.layer [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_layer_num ......... 0 [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_max_iter .......... 100 [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_stability ......... 1e-06 [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_tol ............... 0.01 [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] eigenvalue_verbose ........... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] elasticity_enabled ........... False [2025-04-02 20:03:14,272] [INFO] [config.py:1000: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 } [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] fp16_auto_cast ............... None [2025-04-02 20:03:14,272] [INFO] [config.py:1000:print] fp16_enabled ................. False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] fp16_master_weights_and_gradients False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] global_rank .................. 0 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] grad_accum_dtype ............. None [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] gradient_accumulation_steps .. 1 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] gradient_clipping ............ 0.0 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] gradient_predivide_factor .... 1.0 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] graph_harvesting ............. False [2025-04-02 20:03:14,273] [INFO] [config.py:1000: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 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] initial_dynamic_scale ........ 65536 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] load_universal_checkpoint .... False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] loss_scale ................... 0 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] memory_breakdown ............. False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] mics_hierarchial_params_gather False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] mics_shard_size .............. -1 [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] monitor_config ............... tensorboard=TensorBoardConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') wandb=WandbConfig(enabled=False, group=None, team=None, project='deepspeed') csv_monitor=CSVConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') enabled=False [2025-04-02 20:03:14,273] [INFO] [config.py:1000: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 } [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] optimizer_legacy_fusion ...... False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] optimizer_name ............... None [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] optimizer_params ............. None [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True} [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] pld_enabled .................. False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] pld_params ................... False [2025-04-02 20:03:14,273] [INFO] [config.py:1000:print] prescale_gradients ........... False [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] scheduler_name ............... None [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] scheduler_params ............. None [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] seq_parallel_communication_data_type torch.float32 [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] sparse_attention ............. None [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] sparse_gradients_enabled ..... False [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] steps_per_print .............. inf [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] train_batch_size ............. 1 [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] train_micro_batch_size_per_gpu 1 [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] use_data_before_expert_parallel False [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] use_node_local_storage ....... False [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] wall_clock_breakdown ......... False [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] weight_quantization_config ... None [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] world_size ................... 1 [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] zero_allow_untested_optimizer True [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] zero_config .................. stage=2 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=500,000,000 use_multi_rank_bucket_allreduce=True allgather_partitions=True allgather_bucket_size=500,000,000 overlap_comm=False load_from_fp32_weights=True elastic_checkpoint=False offload_param=DeepSpeedZeroOffloadParamConfig(device='none', nvme_path=None, buffer_count=5, buffer_size=100,000,000, max_in_cpu=1,000,000,000, pin_memory=False) offload_optimizer=DeepSpeedZeroOffloadOptimizerConfig(device='none', nvme_path=None, buffer_count=4, pin_memory=False, pipeline=False, pipeline_read=False, pipeline_write=False, fast_init=False, ratio=1.0) sub_group_size=1,000,000,000 cpu_offload_param=None cpu_offload_use_pin_memory=None cpu_offload=None prefetch_bucket_size=50,000,000 param_persistence_threshold=100,000 model_persistence_threshold=sys.maxsize max_live_parameters=1,000,000,000 max_reuse_distance=1,000,000,000 gather_16bit_weights_on_model_save=False stage3_gather_fp16_weights_on_model_save=False ignore_unused_parameters=True legacy_stage1=False round_robin_gradients=False 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 [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] zero_enabled ................. True [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] zero_force_ds_cpu_optimizer .. True [2025-04-02 20:03:14,274] [INFO] [config.py:1000:print] zero_optimization_stage ...... 2 [2025-04-02 20:03:14,274] [INFO] [config.py:986:print_user_config] json = { "train_batch_size": 1, "train_micro_batch_size_per_gpu": 1, "gradient_accumulation_steps": 1, "zero_optimization": { "stage": 2, "offload_optimizer": { "device": "none", "nvme_path": null }, "offload_param": { "device": "none", "nvme_path": null }, "stage3_gather_16bit_weights_on_model_save": false }, "steps_per_print": inf, "fp16": { "enabled": false }, "bf16": { "enabled": false }, "zero_allow_untested_optimizer": true } 04/02/2025 20:03:14 - INFO - main - save config to ./exp_output/stage1 04/02/2025 20:03:14 - INFO - main - ***** Running training ***** 04/02/2025 20:03:14 - INFO - main - Num examples = 3 04/02/2025 20:03:14 - INFO - main - Num Epochs = 10000 04/02/2025 20:03:14 - INFO - main - Instantaneous batch size per device = 1 04/02/2025 20:03:14 - INFO - main - Total train batch size (w. parallel, distributed & accumulation) = 1 04/02/2025 20:03:14 - INFO - main - Gradient Accumulation steps = 1 04/02/2025 20:03:14 - INFO - main - Total optimization steps = 30000 04/02/2025 20:03:14 - INFO - main - Loading checkpoint from ./exp_output/stage1/checkpoints Could not find checkpoint under ./exp_output/stage1/checkpoints, start training from scratch Steps: 0%| | 0/30000 [00:00<?, ?it/s][2025-04-02 20:03:17,825] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: -11) local_rank: 0 (pid: 1601371) of binary: /home/egg/anaconda3/envs/hallo/bin/python Traceback (most recent call last):