MaxView

2026-04-23 02:38:08.831140: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1776911888.844461 2331918 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776911888.848373 2331918 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776911888.859891 2331918 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776911888.859908 2331918 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776911888.859910 2331918 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776911888.859911 2331918 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
`rope_scaling`'s factor field must be a float >= 1, got 40
`rope_scaling`'s beta_fast field must be a float, got 32
`rope_scaling`'s beta_slow field must be a float, got 1
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
2026-04-23 02:38:12.819513: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0423 02:38:12.954742 125640258283328 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0423 02:39:09.703220 125640258283328 max_utils.py:800] System Information: Jax Version: 0.8.1
I0423 02:39:09.703330 125640258283328 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0423 02:39:09.703369 125640258283328 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709
I0423 02:39:09.703395 125640258283328 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system
I0423 02:39:09.703418 125640258283328 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 02:39:09.703537 125640258283328 train.py:703] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
I0423 02:39:10.363617 125640258283328 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 02:39:10.363890 125640258283328 checkpointing.py:677] Setting up checkpoint logger...
I0423 02:39:10.363937 125640258283328 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0423 02:39:10.363973 125640258283328 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0423 02:39:10.364421 125640258283328 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x72449ffad760>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 02:39:12.042797 125640258283328 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0423 02:39:12.043178 125640258283328 checkpoint_manager.py:702] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7242d01c8e60>}, handler_registry=None
I0423 02:39:12.043428 125640258283328 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7242d01c8e60>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0423 02:39:12.043470 125640258283328 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7242d01ca210>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0423 02:39:12.043500 125640258283328 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7242d01c8e60>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7242d01c8e60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7242d01ca210>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7242d01ca210>}).
I0423 02:39:12.043795 125640258283328 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0423 02:39:12.043846 125640258283328 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7242c6f00900> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0423 02:39:12.490782 125610579985984 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776911777167490390, 'commit_timestamp_nsecs': 1776911826581664225, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA
I0423 02:39:12.568494 125610579985984 checkpoint_manager.py:1847] Missing metrics for step 5
I0423 02:39:12.599204 125640258283328 checkpoint_manager.py:1788] Found 3 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints
I0423 02:39:12.599682 125640258283328 checkpoint_manager.py:921] [process=0][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_hns=False, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False), root_directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7242d0121ac0>
I0423 02:39:12.599808 125640258283328 checkpointing.py:301] Checkpoint manager created!
I0423 02:39:12.989198 125640258283328 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 02:39:12.989290 125640258283328 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0423 02:39:13.272577 125640258283328 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 02:39:13.272663 125640258283328 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0423 02:39:13.284996 125640258283328 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 02:39:13.285056 125640258283328 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0423 02:39:13.304146 125640258283328 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 02:39:13.304207 125640258283328 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 02:39:13.316613 125640258283328 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 02:39:13.316662 125640258283328 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 02:39:13.329009 125640258283328 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 02:39:13.329071 125640258283328 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 02:39:13.348544 125640258283328 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 02:39:13.348604 125640258283328 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 02:39:13.364019 125640258283328 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 02:39:13.364082 125640258283328 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0423 02:39:13.535082 125640258283328 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 02:39:13.535174 125640258283328 checkpointing.py:581] restoring from this run's directory step 9
I0423 02:39:13.697130 125640258283328 checkpointer.py:304] Restoring checkpoint from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/9.
I0423 02:39:14.262283 2333473 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0423 02:39:14.262348 2333473 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0423 02:39:22.326506 125640258283328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/read/gbytes_per_sec: 1.492 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host)
I0423 02:39:22.326797 125640258283328 checkpointer.py:318] Finished restoring checkpoint in 8.71 seconds from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/9.
I0423 02:39:22.326873 125640258283328 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776911953.612968, 'checkpointer_duration_secs': 8.713873863220215, 'checkpoint_manager_start_time': 1776911953.5353785, 'checkpoint_manager_duration_secs': 8.791464567184448}
[DECOUPLED NO-OP] gcs_storage: using stubs.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] workload_monitor: using stub.
[DECOUPLED NO-OP] vertex_tensorboard: using stub.
fsdp: 8
I0423 02:39:22.699901 125640258283328 maxtext_utils.py:1707]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   PartitionSpec('norm',)
    Physical:  (None,)
I0423 02:39:22.700010 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 02:39:22.700069 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 02:39:22.700118 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   PartitionSpec('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0423 02:39:22.700159 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0423 02:39:22.700187 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0423 02:39:22.700227 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0423 02:39:22.700268 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   PartitionSpec('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0423 02:39:22.700298 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0423 02:39:22.700324 125640258283328 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0423 02:39:22.700358 125640258283328 maxtext_utils.py:1707]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   PartitionSpec('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0423 02:39:22.700393 125640258283328 maxtext_utils.py:1707]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   PartitionSpec('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0423 02:39:23.103293 125640258283328 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0423 02:39:23.103375 125640258283328 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0423 02:39:23.115062 125640258283328 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0423 02:39:23.115111 125640258283328 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
I0423 02:39:23.923078 125640258283328 max_utils.py:791] Total memory size: 3.6 GB, Output size: 1.5 GB, Temp size: 2.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB.
I0423 02:39:23.923730 125640258283328 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0423 02:39:23.923887 125640258283328 metric_logger.py:301] number parameters: 1.104 billion
I0423 02:39:25.876671 125640258283328 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0423 02:39:25.994324 125640258283328 checkpointing.py:776] Waited 0.11764383316040039 seconds for step 10 to finish before starting checkpointing.
I0423 02:39:25.994933 125640258283328 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 02:39:25.995101 125640258283328 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 10
I0423 02:39:25.995391 125640258283328 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10.
I0423 02:39:26.077254 125640258283328 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0423 02:39:26.157473 125608386364992 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10
I0423 02:39:26.166627 125640258283328 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0423 02:39:26.166717 125640258283328 replica_slices.py:410] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0423 02:39:26.929141 125608375879232 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10/items
I0423 02:39:27.084536 125608344421952 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776911966838640664, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776911966838640664, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA
I0423 02:39:27.920171 125640258283328 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.753945s
I0423 02:39:27.924842 125640258283328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 6.684 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host)
I0423 02:39:27.925013 125640258283328 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.846353s (batch_requests_ready=0.084975s, total_serialization_initiated=1.756858s, others=0.004520s)
I0423 02:39:27.925375 125640258283328 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.847638s (all_items=0.000022s, per_item={'items': '0.00002193'}, temp_paths=1.847616)
I0423 02:39:27.926932 125608396850752 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0423 02:39:27.927020 125640258283328 async_checkpointer.py:561] Finished blocking save. Time taken: 1.931874s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10.
I0423 02:39:27.927221 125640258283328 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 02:39:27.927412 125608386364992 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 02:39:27.927523 125640258283328 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776911965.9949207, 'wait_for_prev_duration_secs': 4.0531158447265625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776911965.9951239, 'checkpointer_blocking_duration_secs': 1.9320204257965088, 'get_old_steps_start_time': 1776911967.9271605, 'get_old_steps_duration_secs': 2.2649765014648438e-05, 'checkpoint_manager_blocking_start_time': 1776911965.9948344, 'checkpoint_manager_blocking_duration_secs': 1.9326610565185547}
I0423 02:39:27.927636 125640258283328 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0423 02:39:27.927677 125640258283328 max_utils.py:750] 
Memstats: After params initialized:
I0423 02:39:27.927736 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0))
I0423 02:39:27.927763 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0423 02:39:27.927785 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0423 02:39:27.927805 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0423 02:39:27.927825 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0423 02:39:27.927843 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0423 02:39:27.927861 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0423 02:39:27.927878 125640258283328 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
I0423 02:39:28.173270 125640258283328 metric_logger.py:196] completed step: 10, seconds: 1.952, TFLOP/s/device: 6.960, Tokens/s/device: 1049.041, total_weights: 16384, loss: 3.602, lm_loss: 3.602, perplexity: 36.686
I0423 02:39:28.289619 125640258283328 metric_logger.py:196] completed step: 11, seconds: 2.295, TFLOP/s/device: 5.922, Tokens/s/device: 892.565, total_weights: 16384, loss: 3.101, lm_loss: 3.101, perplexity: 22.211
I0423 02:39:28.408800 125640258283328 metric_logger.py:196] completed step: 12, seconds: 0.006, TFLOP/s/device: 2257.373, Tokens/s/device: 340255.856, total_weights: 16384, loss: 2.684, lm_loss: 2.684, perplexity: 14.644
I0423 02:39:28.527784 125640258283328 metric_logger.py:196] completed step: 13, seconds: 0.117, TFLOP/s/device: 115.830, Tokens/s/device: 17459.208, total_weights: 16384, loss: 2.348, lm_loss: 2.348, perplexity: 10.461
I0423 02:39:28.533070 125640258283328 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0423 02:39:28.657242 125608354907712 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10/items/array_metadatas/process_0
I0423 02:39:28.765710 125640258283328 checkpointing.py:776] Waited 0.2326047420501709 seconds for step 15 to finish before starting checkpointing.
I0423 02:39:28.766713 125640258283328 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 02:40:04.435428 125608375879232 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.509334s (commit=36.001417s, array_metadata_write=0.507917s)
I0423 02:40:04.437148 125608396850752 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 329.439 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 38 seconds) (per-host)
I0423 02:40:04.437206 125608396850752 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.510130s.
I0423 02:40:04.909359 125608396850752 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0423 02:40:05.087306 125608344421952 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776911966838640664, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA
I0423 02:40:05.387356 125608396850752 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.665566s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10/items
I0423 02:40:05.388266 125608396850752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10/items
I0423 02:40:05.625998 125608396850752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10
I0423 02:40:06.313650 125608396850752 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10`.
I0423 02:40:06.314326 125608396850752 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 40.319189s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/10
I0423 02:40:06.314404 125608396850752 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 38.387328s.
I0423 02:40:06.314596 125608386364992 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 02:40:06.314716 125608386364992 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 02:40:06.314789 125608386364992 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0423 02:40:06.314838 125608386364992 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0423 02:40:06.314999 125640258283328 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0423 02:40:06.315144 125640258283328 checkpoint_manager.py:1441] Waiting for previous save to complete took 37.548424 seconds. If this number is high, consider checkpointing less frequently.
I0423 02:40:06.316266 125640258283328 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 15
I0423 02:40:06.316560 125640258283328 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15.
I0423 02:40:06.461676 125608396850752 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15
I0423 02:40:06.466272 125640258283328 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0423 02:40:06.466365 125640258283328 replica_slices.py:410] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0423 02:40:07.196436 125608354907712 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15/items
I0423 02:40:10.910377 125640258283328 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 4.444467s
I0423 02:40:10.914381 125640258283328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.719 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host)
I0423 02:40:10.914459 125640258283328 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.539519s (batch_requests_ready=0.087320s, total_serialization_initiated=4.448306s, others=0.003893s)
I0423 02:40:10.914535 125640258283328 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.540280s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=4.540266)
I0423 02:40:10.916036 125608396850752 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0423 02:40:10.916183 125640258283328 async_checkpointer.py:561] Finished blocking save. Time taken: 4.599873s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15.
I0423 02:40:10.916335 125640258283328 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 02:40:10.916598 125608386364992 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 02:40:10.916713 125640258283328 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776911968.7666898, 'wait_for_prev_duration_secs': 37.548424243927, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776912006.3162913, 'checkpointer_blocking_duration_secs': 4.599951505661011, 'get_old_steps_start_time': 1776912010.9162612, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776911968.7666166, 'checkpoint_manager_blocking_duration_secs': 42.15007257461548}
I0423 02:40:10.916839 125640258283328 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0423 02:40:10.917607 125640258283328 metric_logger.py:196] completed step: 14, seconds: 0.120, TFLOP/s/device: 113.425, Tokens/s/device: 17096.586, total_weights: 16384, loss: 2.083, lm_loss: 2.083, perplexity: 8.030
I0423 02:40:10.926079 125640258283328 metric_logger.py:196] completed step: 15, seconds: 0.119, TFLOP/s/device: 114.596, Tokens/s/device: 17273.226, total_weights: 16384, loss: 1.880, lm_loss: 1.880, perplexity: 6.551
I0423 02:40:11.042789 125640258283328 metric_logger.py:196] completed step: 16, seconds: 42.391, TFLOP/s/device: 0.321, Tokens/s/device: 48.312, total_weights: 16384, loss: 1.726, lm_loss: 1.726, perplexity: 5.616
I0423 02:40:11.161910 125640258283328 metric_logger.py:196] completed step: 17, seconds: 0.006, TFLOP/s/device: 2224.117, Tokens/s/device: 335243.084, total_weights: 16384, loss: 1.610, lm_loss: 1.610, perplexity: 5.002
I0423 02:40:11.280945 125640258283328 metric_logger.py:196] completed step: 18, seconds: 0.118, TFLOP/s/device: 115.348, Tokens/s/device: 17386.580, total_weights: 16384, loss: 1.521, lm_loss: 1.521, perplexity: 4.577
I0423 02:40:11.398760 125640258283328 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0423 02:40:11.399785 125640258283328 checkpointing.py:776] Waited 0.0010426044464111328 seconds for step 19 to finish before starting checkpointing.
I0423 02:40:11.400118 125640258283328 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 02:40:11.664978 125610559014464 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15/items/array_metadatas/process_0
I0423 02:40:40.710891 125608354907712 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.796057s (commit=29.315299s, array_metadata_write=0.480759s)
I0423 02:40:40.711918 125608396850752 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 368.023 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 34 seconds) (per-host)
I0423 02:40:40.712025 125608396850752 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.795880s.
I0423 02:40:41.187926 125608396850752 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0423 02:40:41.674571 125608396850752 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.647541s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15/items
I0423 02:40:41.675318 125608396850752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15/items
I0423 02:40:41.904203 125608396850752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15
I0423 02:40:42.606187 125608396850752 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15`.
I0423 02:40:42.606848 125608396850752 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 36.290545s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/15
I0423 02:40:42.606916 125608396850752 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 31.690774s.
I0423 02:40:42.607113 125608386364992 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 02:40:42.607221 125608386364992 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 02:40:42.607274 125608386364992 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0423 02:40:42.607314 125608386364992 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0423 02:40:42.607460 125640258283328 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0423 02:40:42.607574 125640258283328 checkpoint_manager.py:1441] Waiting for previous save to complete took 31.207470 seconds. If this number is high, consider checkpointing less frequently.
I0423 02:40:42.607875 125640258283328 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 19
I0423 02:40:42.608181 125640258283328 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19.
I0423 02:40:42.772347 125640258283328 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0423 02:40:42.772487 125608396850752 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19
I0423 02:40:42.772572 125640258283328 replica_slices.py:410] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0423 02:40:43.453771 125610559014464 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19/items
I0423 02:40:47.380373 125640258283328 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 4.608587s
I0423 02:40:47.383959 125640258283328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.628 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host)
I0423 02:40:47.384030 125640258283328 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.695066s (batch_requests_ready=0.079746s, total_serialization_initiated=4.611831s, others=0.003488s)
I0423 02:40:47.384111 125640258283328 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.695832s (all_items=0.000014s, per_item={'items': '0.00001407'}, temp_paths=4.695818)
I0423 02:40:47.385875 125608386364992 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0423 02:40:47.386022 125640258283328 async_checkpointer.py:561] Finished blocking save. Time taken: 4.778102s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19.
I0423 02:40:47.386153 125640258283328 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 02:40:47.386320 125608396850752 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 02:40:47.386419 125640258283328 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776912011.4000885, 'wait_for_prev_duration_secs': 31.207470178604126, 'time_between_consecutive_saves_sec': 5.085235595703125, 'checkpointer_blocking_start_time': 1776912042.6079, 'checkpointer_blocking_duration_secs': 4.778192758560181, 'get_old_steps_start_time': 1776912047.3861015, 'get_old_steps_duration_secs': 1.9788742065429688e-05, 'checkpoint_manager_blocking_start_time': 1776912011.4000351, 'checkpoint_manager_blocking_duration_secs': 35.986361503601074}
I0423 02:40:47.386534 125640258283328 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0423 02:40:47.386566 125640258283328 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 02:40:48.138253 125610410116672 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19/items/array_metadatas/process_0
I0423 02:41:28.147143 125610559014464 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.762574s (commit=40.267097s, array_metadata_write=0.495477s)
I0423 02:41:28.148317 125608386364992 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 277.980 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 45 seconds) (per-host)
I0423 02:41:28.148432 125608386364992 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.762451s.
I0423 02:41:28.669032 125608386364992 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0423 02:41:29.140903 125608386364992 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.647146s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19/items
I0423 02:41:29.141757 125608386364992 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19/items
I0423 02:41:29.383649 125608386364992 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19
I0423 02:41:30.083230 125608386364992 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19`.
I0423 02:41:30.083834 125608386364992 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 47.475924s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_16_zarr3_false/checkpoints/19
I0423 02:41:30.083910 125608386364992 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 42.697931s.
I0423 02:41:30.084095 125608396850752 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 02:41:30.084207 125608396850752 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 02:41:30.084271 125608396850752 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0423 02:41:30.084313 125608396850752 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0423 02:41:30.084471 125640258283328 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0423 02:41:30.084625 125640258283328 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 02:41:30.085704 125640258283328 metric_logger.py:196] completed step: 19, seconds: 0.119, TFLOP/s/device: 113.924, Tokens/s/device: 17171.844, total_weights: 16384, loss: 1.449, lm_loss: 1.449, perplexity: 4.260
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops