/venv/main/lib/python3.12/site-packages/torch/jit/_script.py:1491: FutureWarning: `torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.
  warnings.warn(
/venv/main/lib/python3.12/site-packages/torch/jit/_script.py:1491: FutureWarning: `torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.
  warnings.warn(
[2026-09-30 06:12:54,353] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
[2026-09-30 06:12:54,353] [INFO] [runner.py:630:main] cmd = /venv/main/bin/python -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29500 --enable_each_rank_log=None --log_level=info train.py --deepspeed --config examples/anima_nextscene/gpu_20260930/smoke_b1.toml
/venv/main/lib/python3.12/site-packages/torch/jit/_script.py:1491: FutureWarning: `torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.
  warnings.warn(
/venv/main/lib/python3.12/site-packages/torch/jit/_script.py:1491: FutureWarning: `torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.
  warnings.warn(
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NCCL_P2P_DISABLE=1
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NV_LIBNCCL_DEV_PACKAGE=libnccl-dev=2.28.3-1+cuda13.0
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NV_LIBNCCL_DEV_PACKAGE_VERSION=2.28.3-1
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NCCL_VERSION=2.28.3-1
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NV_LIBNCCL_DEV_PACKAGE_NAME=libnccl-dev
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NV_LIBNCCL_PACKAGE=libnccl2=2.28.3-1+cuda13.0
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NV_LIBNCCL_PACKAGE_NAME=libnccl2
[2026-09-30 06:12:57,437] [INFO] [launch.py:155:main] 0 NV_LIBNCCL_PACKAGE_VERSION=2.28.3-1
[2026-09-30 06:12:57,437] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
[2026-09-30 06:12:57,437] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
[2026-09-30 06:12:57,437] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
[2026-09-30 06:12:57,437] [INFO] [launch.py:180:main] dist_world_size=1
[2026-09-30 06:12:57,437] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
[2026-09-30 06:12:57,437] [INFO] [launch.py:272:main] process 19657 spawned with command: ['/venv/main/bin/python', '-u', 'train.py', '--local_rank=0', '--deepspeed', '--config', 'examples/anima_nextscene/gpu_20260930/smoke_b1.toml']
/venv/main/lib/python3.12/site-packages/torch/jit/_script.py:1491: FutureWarning: `torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.
  warnings.warn(
/venv/main/lib/python3.12/site-packages/torch/jit/_script.py:1491: FutureWarning: `torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.
  warnings.warn(
[2026-09-30 06:13:01,417] [INFO] [comm.py:821:init_distributed] cdb=None
[2026-09-30 06:13:01,417] [INFO] [comm.py:852:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
using video_clip_mode=single_beginning
using framerate=16
Grouped metadata is not cached. Computing ungrouped metadata and then grouping.
Intermediate metadata is not cached. Enumerating all files.

  0%|          | 0/32 [00:00<?, ?it/s]
100%|██████████| 32/32 [00:00<00:00, 118987.35it/s]
Saving intermediate metadata dataset.

Saving the dataset (0/1 shards):   0%|          | 0/16 [00:00<?, ? examples/s]
Saving the dataset (1/1 shards): 100%|██████████| 16/16 [00:00<00:00, 9918.54 examples/s]
Saving the dataset (1/1 shards): 100%|██████████| 16/16 [00:00<00:00, 9447.96 examples/s]
Loading intermediate metadata dataset.
Caching ungrouped metadata.

Map (num_proc=8):   0%|          | 0/16 [00:00<?, ? examples/s]
Map (num_proc=8): 100%|██████████| 16/16 [00:00<00:00, 88.55 examples/s]

Grouping examples:   0%|          | 0/16 [00:00<?, ?it/s]
Grouping examples: 100%|██████████| 16/16 [00:00<00:00, 43919.41it/s]

Saving the dataset (0/1 shards):   0%|          | 0/16 [00:00<?, ? examples/s]
Saving the dataset (1/1 shards): 100%|██████████| 16/16 [00:00<00:00, 10903.15 examples/s]
Saving the dataset (1/1 shards): 100%|██████████| 16/16 [00:00<00:00, 10407.70 examples/s]
Loading grouped metadata with grouping key (1.0, 1)
caching latents: /workspace/ns_smoke/target
caching latents: (1.0, 1)

Adding size bucket:   0%|          | 0/16 [00:00<?, ? examples/s]
Adding size bucket: 100%|██████████| 16/16 [00:00<00:00, 7710.12 examples/s]
caching latents: (1.0, 512, 512, 1)
[CACHE] Initializing
[CACHE] Storing new fingerprint: 729cf0b7f391da14
[CACHE] Existing cache length: 0

  0%|          | 0/4 [00:00<?, ?it/s][CACHE] Creating new shard: shard_0

 25%|██▌       | 1/4 [00:00<00:02,  1.36it/s]
 50%|█████     | 2/4 [00:00<00:00,  2.60it/s]
100%|██████████| 4/4 [00:01<00:00,  5.32it/s]
100%|██████████| 4/4 [00:01<00:00,  3.94it/s]
Building iteration order

Saving the dataset (0/1 shards):   0%|          | 0/16 [00:00<?, ? examples/s]
Saving the dataset (1/1 shards): 100%|██████████| 16/16 [00:00<00:00, 12593.14 examples/s]
Saving the dataset (1/1 shards): 100%|██████████| 16/16 [00:00<00:00, 11831.61 examples/s]
caching text embeddings: /workspace/ns_smoke/target
caching text embeddings: (1.0, 1)

Map:   0%|          | 0/16 [00:00<?, ? examples/s]
Map: 100%|██████████| 16/16 [00:00<00:00, 10446.59 examples/s]
[CACHE] Initializing
[CACHE] Storing new fingerprint: 85f0af65ef677478
[CACHE] Existing cache length: 0

  0%|          | 0/4 [00:00<?, ?it/s][CACHE] Creating new shard: shard_0

 25%|██▌       | 1/4 [00:00<00:02,  1.08it/s]
100%|██████████| 4/4 [00:01<00:00,  3.98it/s]
[CACHE] Initializing
[CACHE] Storing new fingerprint: b072e77f045c8ad1
[CACHE] Existing cache length: 0

  0%|          | 0/1 [00:00<?, ?it/s][CACHE] Creating new shard: shard_0

100%|██████████| 1/1 [00:00<00:00, 57.87it/s]
Found grouped metadata cache. Directly loading it.
Loading grouped metadata with grouping key [1.0, 1]
caching latents: /workspace/ns_smoke/target
caching latents: [1.0, 1]
caching latents: (1.0, 512, 512, 1)
[CACHE] Initializing
[CACHE] Existing cache has fingerprint 729cf0b7f391da14
[CACHE] Existing cache length: 16
caching text embeddings: /workspace/ns_smoke/target
caching text embeddings: [1.0, 1]

Map:   0%|          | 0/16 [00:00<?, ? examples/s]
Map: 100%|██████████| 16/16 [00:00<00:00, 7067.81 examples/s]
[CACHE] Initializing
[CACHE] Existing cache has fingerprint 85f0af65ef677478
[CACHE] Existing cache length: 16
[CACHE] Initializing
[CACHE] Existing cache has fingerprint b072e77f045c8ad1
[CACHE] Existing cache length: 1
[Anima NextScene] RoPE layout=disjoint_w ref_temporal_index=1
[Anima NextScene] LoRA targets: 280 linears (112 cross_attn; adaln/llm_adapter excluded)
trainable params: 96,468,992 || all params: 2,187,537,920 || trainable%: 4.4099
SEED_LAYERS=False BASE_SEED=1234 SEED_FN=None
Using topology: {ProcessCoord(pipe=0, data=0): 0}
[2026-09-30 06:13:05,584] [INFO] [module.py:398:_partition_layers] Partitioning pipeline stages with method parameters
stage=0 layers=31
     0: InitialLayer
     1: LLMAdapterLayer
     2: TransformerLayer
     3: TransformerLayer
     4: TransformerLayer
     5: TransformerLayer
     6: TransformerLayer
     7: TransformerLayer
     8: TransformerLayer
     9: TransformerLayer
    10: TransformerLayer
    11: TransformerLayer
    12: TransformerLayer
    13: TransformerLayer
    14: TransformerLayer
    15: TransformerLayer
    16: TransformerLayer
    17: TransformerLayer
    18: TransformerLayer
    19: TransformerLayer
    20: TransformerLayer
    21: TransformerLayer
    22: TransformerLayer
    23: TransformerLayer
    24: TransformerLayer
    25: TransformerLayer
    26: TransformerLayer
    27: TransformerLayer
    28: TransformerLayer
    29: TransformerLayer
    30: FinalLayer
  loss: loss_fn
[2026-09-30 06:13:05,947] [INFO] [logging.py:123:log_dist] [Rank 0] DeepSpeed info: version=0.18.4, git-hash=unknown, git-branch=unknown
[2026-09-30 06:13:05,947] [INFO] [config.py:687:__init__] Config mesh_device None world_size = 1
[2026-09-30 06:13:05,965] [INFO] [engine.py:1415:_configure_distributed_model] ********** distributed groups summary **********
	 self.dp_world_size=1
	 self.mp_world_size=1
	 self.seq_dp_world_size=1
	 self.sequence_parallel_size=1
***********************************************
[2026-09-30 06:13:05,974] [INFO] [logging.py:123:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False
[2026-09-30 06:13:05,978] [INFO] [logging.py:123:log_dist] [Rank 0] [TorchCheckpointEngine] Initialized with serialization = True
[2026-09-30 06:13:05,978] [INFO] [config.py:958:print] DeepSpeedEngine configuration:
[2026-09-30 06:13:05,978] [INFO] [config.py:962:print]   activation_checkpointing_config  {
    "partition_activations": false, 
    "contiguous_memory_optimization": false, 
    "cpu_checkpointing": false, 
    "number_checkpoints": null, 
    "synchronize_checkpoint_boundary": false, 
    "profile": false
}
[2026-09-30 06:13:05,978] [INFO] [config.py:962:print]   aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'intra_op_parallelism': 1, 'single_submit': False, 'overlap_events': True, 'use_gds': False}
[2026-09-30 06:13:05,978] [INFO] [config.py:962:print]   amp_enabled .................. False
[2026-09-30 06:13:05,978] [INFO] [config.py:962:print]   amp_params ................... False
[2026-09-30 06:13:05,978] [INFO] [config.py:962: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
}
[2026-09-30 06:13:05,978] [INFO] [config.py:962:print]   bfloat16_config .............. enabled=False immediate_grad_update=False check_grad_overflow=False bf16_master_weights_and_grads=False bf16_optimizer_states=False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   checkpoint_config ............ {'tag_validation': 'WARN', 'checkpoint_serialization': True, 'writer': None}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   checkpoint_parallel_write_pipeline  False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   checkpoint_tag_validation_enabled  True
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   checkpoint_tag_validation_fail  False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   comms_config ................. <deepspeed.comm.config.DeepSpeedCommsConfig object at 0x7e09e56e1af0>
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   communication_data_type ...... None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   compile_config ............... deepcompile=False free_activation=False free_activation_threshold=10485760 offload_activation=False offload_opt_states=False double_buffer=True symmetric_memory=False debug_log=False offload_parameters=False sync_before_reduce=False sync_after_reduce=False sync_before_allgather=False sync_after_allgather=False keep_int_input_tensors=True keep_all_input_tensors=False
[2026-09-30 06:13:05,979] [INFO] [config.py:962: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}}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   curriculum_enabled_legacy .... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   curriculum_params_legacy ..... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   data_efficiency_config ....... {'enabled': False, 'seed': 1234, 'data_sampling': {'enabled': False, 'num_epochs': 1000, 'num_workers': 0, 'pin_memory': False, 'curriculum_learning': {'enabled': False}, 'dynamic_batching': {'enabled': False, 'lr_scaling_method': 'linear', 'min_batch_size': 1, 'max_batch_size': None, 'sequence_picking_order': 'dataloader', 'verbose': False}}, 'data_routing': {'enabled': False, 'random_ltd': {'enabled': False, 'layer_token_lr_schedule': {'enabled': False}}}}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   data_efficiency_enabled ...... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   dataloader_drop_last ......... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   datastates_config ............ {
    "enabled": false, 
    "config": null
}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   disable_allgather ............ False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   dump_state ................... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_enabled ........... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_gas_boundary_resolution  1
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_layer_name ........ bert.encoder.layer
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_layer_num ......... 0
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_max_iter .......... 100
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_stability ......... 1e-06
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_tol ............... 0.01
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   eigenvalue_verbose ........... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   elasticity_enabled ........... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   float16_config ............... enabled=False auto_cast=False loss_scale=0.0 initial_scale_power=16 loss_scale_window=1000 hysteresis=2 consecutive_hysteresis=False min_loss_scale=1 fp16_master_weights_and_grads=False
[2026-09-30 06:13:05,979] [INFO] [config.py:962: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
}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   global_rank .................. 0
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   grad_accum_dtype ............. None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   gradient_accumulation_steps .. 1
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   gradient_clipping ............ 1.0
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   gradient_predivide_factor .... 1.0
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   graph_harvesting ............. False
[2026-09-30 06:13:05,979] [INFO] [config.py:962: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
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   load_universal_checkpoint .... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   memory_breakdown ............. False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   mics_hierarchial_params_gather  False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   mics_shard_size .............. -1
[2026-09-30 06:13:05,979] [INFO] [config.py:962: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')
[2026-09-30 06:13:05,979] [INFO] [config.py:962: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
}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   optimizer_legacy_fusion ...... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   optimizer_name ............... None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   optimizer_params ............. None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True}
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   pld_enabled .................. False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   pld_params ................... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   prescale_gradients ........... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   scheduler_name ............... None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   scheduler_params ............. None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   seq_parallel_communication_data_type  torch.float32
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   sparse_attention ............. None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   sparse_gradients_enabled ..... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   steps_per_print .............. 1
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   tensor_parallel_config ....... dtype=torch.float16 autotp_size=0 tp_overlap_comm=False tensor_parallel=TPConfig(tp_size=1, tp_grain_size=1, mpu=None, tp_group=None) injection_policy_tuple=None keep_module_on_host=False replace_with_kernel_inject=False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   timers_config ................ enabled=True synchronized=True
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   torch_autocast_dtype ......... None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   torch_autocast_enabled ....... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   torch_autocast_lower_precision_safe_modules  None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   train_batch_size ............. 1
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   train_micro_batch_size_per_gpu  1
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   use_data_before_expert_parallel_  False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   use_node_local_storage ....... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   wall_clock_breakdown ......... False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   weight_quantization_config ... None
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   world_size ................... 1
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   zero_allow_untested_optimizer  False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   zero_config .................. stage=0 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=500000000 use_multi_rank_bucket_allreduce=True allgather_partitions=True allgather_bucket_size=500000000 overlap_comm=False load_from_fp32_weights=True elastic_checkpoint=False offload_param=None offload_optimizer=None zenflow=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 module_granularity_threshold=0 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=False zero_hpz_partition_size=1 zero_quantized_weights=False zero_quantized_nontrainable_weights=False zero_quantized_gradients=False zeropp_loco_param=None mics_shard_size=-1 mics_hierarchical_params_gather=False memory_efficient_linear=True pipeline_loading_checkpoint=False override_module_apply=True log_trace_cache_warnings=False enable_sanity_checks=False leaf_module=DeepSpeedZeroLeafModuleConfig(classes=['transformers.models.mixtral.modeling_mixtral.MixtralSparseMoeBlock', 'transformers.models.qwen2_moe.modeling_qwen2_moe.Qwen2MoeSparseMoeBlock', 'transformers.models.qwen3_moe.modeling_qwen3_moe.Qwen3MoeSparseMoeBlock'], names=[], name_suffixes=[])
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   zero_enabled ................. False
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   zero_force_ds_cpu_optimizer .. True
[2026-09-30 06:13:05,979] [INFO] [config.py:962:print]   zero_optimization_stage ...... 0
[2026-09-30 06:13:05,979] [INFO] [config.py:948:print_user_config]   json = {
    "train_micro_batch_size_per_gpu": 1, 
    "gradient_accumulation_steps": 1, 
    "gradient_clipping": 1.0, 
    "steps_per_print": 1
}
[2026-09-30 06:13:05,979] [INFO] [engine.py:104:__init__] CONFIG: micro_batches=1 micro_batch_size=1
[2026-09-30 06:13:05,979] [INFO] [engine.py:145:__init__] is_pipe_partitioned= False is_grad_partitioned= False
[2026-09-30 06:13:05,982] [INFO] [engine.py:164:__init__] RANK=0 STAGE=0 LAYERS=31 [0, 31) STAGE_PARAMS=96468992 (96.469M) TOTAL_PARAMS=96468992 (96.469M) UNIQUE_PARAMS=96468992 (96.469M)
Global batch size = 1
Using base_lr=5e-05, self_attn_lr=5e-05, cross_attn_lr=5e-05, mlp_lr=5e-05, mod_lr=5e-05, llm_adapter_lr=5e-05
Num base params: 0
Num self_attn params: 224
Num cross_attn params: 224
Num mlp params: 112
Num mod params: 0
Num llm_adapter params: 72
[2026-09-30 06:13:06,008] [INFO] [logging.py:123:log_dist] [Rank 0] Using client callable to create basic optimizer
[2026-09-30 06:13:06,008] [INFO] [logging.py:123:log_dist] [Rank 0] Removing param_group that has no 'params' in the basic Optimizer
[2026-09-30 06:13:06,089] [INFO] [logging.py:123:log_dist] [Rank 0] DeepSpeed Basic Optimizer = AdamW
[2026-09-30 06:13:06,089] [INFO] [logging.py:123:log_dist] [Rank 0] DeepSpeed Final Optimizer = AdamW
[rank0]: Traceback (most recent call last):
[rank0]:   File "/workspace/diffusion-pipe-easycontrol/train.py", line 1010, in <module>
[rank0]:     iterator = get_data_iterator_for_step(train_dataloader, model_engine)
[rank0]:                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[rank0]:   File "/workspace/diffusion-pipe-easycontrol/train.py", line 172, in get_data_iterator_for_step
[rank0]:     items = [next(dataloader_iter) for _ in range(num_micro_batches)]
[rank0]:              ^^^^^^^^^^^^^^^^^^^^^
[rank0]:   File "/workspace/diffusion-pipe-easycontrol/utils/dataset.py", line 1494, in __next__
[rank0]:     self.next_micro_batch = next(self.data)
[rank0]:                             ^^^^^^^^^^^^^^^
[rank0]:   File "/workspace/diffusion-pipe-easycontrol/utils/dataset.py", line 1531, in _pull_batches_from_dataloader
[rank0]:     target, mask = label
[rank0]:     ^^^^^^^^^^^^
[rank0]: ValueError: too many values to unpack (expected 2)
[2026-09-30 06:13:08,438] [INFO] [launch.py:335:sigkill_handler] Killing subprocess 19657
[2026-09-30 06:13:08,439] [ERROR] [launch.py:341:sigkill_handler] ['/venv/main/bin/python', '-u', 'train.py', '--local_rank=0', '--deepspeed', '--config', 'examples/anima_nextscene/gpu_20260930/smoke_b1.toml'] exits with return code = 1