2026-04-23 01:43:59.444232: 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 00:00:1776908639.457534 2282724 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776908639.461380 2282724 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776908639.472866 2282724 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776908639.472881 2282724 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776908639.472883 2282724 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776908639.472884 2282724 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 01:44:03.164080: 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 01:44:03.299695 135404902385472 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0423 01:44:34.155269 135404902385472 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 01:44:34.155531 135404902385472 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 01:44:34.155584 135404902385472 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 01:44:34.155615 135404902385472 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system I0423 01:44:34.155637 135404902385472 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 01:44:34.155806 135404902385472 train.py:703] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0423 01:44:34.808697 135404902385472 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 01:44:34.808970 135404902385472 checkpointing.py:677] Setting up checkpoint logger... I0423 01:44:34.809016 135404902385472 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 01:44:34.809067 135404902385472 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0423 01:44:34.809489 135404902385472 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7b26221b1c40>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 01:44:36.477270 135404902385472 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0423 01:44:36.477654 135404902385472 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 0x7b245423dca0>}, handler_registry=None I0423 01:44:36.477885 135404902385472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b245423dca0>` 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 01:44:36.477927 135404902385472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b245423ca70>` 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 01:44:36.477957 135404902385472 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 0x7b245423dca0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b245423dca0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b245423ca70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b245423ca70>}). I0423 01:44:36.478246 135404902385472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0423 01:44:36.478300 135404902385472 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7b2454254860> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 01:44:36.916381 135375177647680 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776908540901841809, 'commit_timestamp_nsecs': 1776908572839679338, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/5/_CHECKPOINT_METADATA I0423 01:44:37.002904 135375188133440 checkpoint_manager.py:1847] Missing metrics for step 0 I0423 01:44:37.023493 135404902385472 checkpoint_manager.py:1788] Found 3 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints I0423 01:44:37.024010 135404902385472 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_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b245423e180> I0423 01:44:37.024127 135404902385472 checkpointing.py:301] Checkpoint manager created! I0423 01:44:37.428999 135404902385472 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 01:44:37.429154 135404902385472 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 01:44:37.716951 135404902385472 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 01:44:37.717058 135404902385472 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 01:44:37.729437 135404902385472 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 01:44:37.729487 135404902385472 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 01:44:37.748522 135404902385472 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 01:44:37.748580 135404902385472 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:44:37.760950 135404902385472 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 01:44:37.761003 135404902385472 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:44:37.773492 135404902385472 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 01:44:37.773545 135404902385472 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:44:37.793083 135404902385472 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 01:44:37.793146 135404902385472 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:44:37.808577 135404902385472 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 01:44:37.808634 135404902385472 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0423 01:44:37.958896 135404902385472 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0423 01:44:37.959002 135404902385472 checkpointing.py:581] restoring from this run's directory step 9 I0423 01:44:38.139486 135404902385472 checkpointer.py:304] Restoring checkpoint from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/9. I0423 01:44:38.721697 2284206 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0423 01:44:38.721770 2284206 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0423 01:44:45.894099 135404902385472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/read/gbytes_per_sec: 1.664 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host) I0423 01:44:45.894368 135404902385472 checkpointer.py:318] Finished restoring checkpoint in 7.84 seconds from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/9. I0423 01:44:45.894437 135404902385472 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776908678.0493367, 'checkpointer_duration_secs': 7.84507417678833, 'checkpoint_manager_start_time': 1776908677.9592588, 'checkpoint_manager_duration_secs': 7.935155630111694} [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 01:44:46.256160 135404902385472 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 01:44:46.256345 135404902385472 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 01:44:46.256389 135404902385472 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 01:44:46.256436 135404902385472 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 01:44:46.256479 135404902385472 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 01:44:46.256508 135404902385472 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 01:44:46.256548 135404902385472 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 01:44:46.256588 135404902385472 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 01:44:46.256617 135404902385472 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 01:44:46.256642 135404902385472 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 01:44:46.256676 135404902385472 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0423 01:44:46.256711 135404902385472 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0423 01:44:46.670206 135404902385472 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 01:44:46.670323 135404902385472 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0423 01:44:46.682578 135404902385472 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 01:44:46.682629 135404902385472 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0423 01:44:47.488314 135404902385472 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 01:44:47.489122 135404902385472 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0423 01:44:47.489314 135404902385472 metric_logger.py:301] number parameters: 1.104 billion I0423 01:44:49.613144 135404902385472 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0423 01:44:49.730602 135404902385472 checkpointing.py:776] Waited 0.11743927001953125 seconds for step 10 to finish before starting checkpointing. I0423 01:44:49.731272 135404902385472 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 01:44:49.731430 135404902385472 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 10 I0423 01:44:49.731744 135404902385472 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10. I0423 01:44:49.817718 135404902385472 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0423 01:44:49.895509 135404902385472 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 01:44:49.895604 135404902385472 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 01:44:49.904513 135372803671616 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10 I0423 01:44:50.635253 135372793185856 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10/items I0423 01:44:50.769106 135372761728576 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776908690537925968, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776908690537925968, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10/_CHECKPOINT_METADATA I0423 01:44:52.098320 135404902385472 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.203280s I0423 01:44:52.103855 135404902385472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.402 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0423 01:44:52.104264 135404902385472 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.285071s (batch_requests_ready=0.073175s, total_serialization_initiated=2.206180s, others=0.005716s) I0423 01:44:52.104403 135404902385472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.286131s (all_items=0.000021s, per_item={'items': '0.00002098'}, temp_paths=2.286110) I0423 01:44:52.105929 135372803671616 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 01:44:52.106016 135404902385472 async_checkpointer.py:561] Finished blocking save. Time taken: 2.374546s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10. I0423 01:44:52.106240 135404902385472 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0423 01:44:52.106443 135372952569408 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 01:44:52.106559 135404902385472 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776908689.731258, 'wait_for_prev_duration_secs': 4.315376281738281e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776908689.7314517, 'checkpointer_blocking_duration_secs': 2.374704360961914, 'get_old_steps_start_time': 1776908692.106171, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1776908689.7311678, 'checkpoint_manager_blocking_duration_secs': 2.375363349914551} I0423 01:44:52.106739 135404902385472 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0423 01:44:52.106781 135404902385472 max_utils.py:750] Memstats: After params initialized: I0423 01:44:52.106830 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0423 01:44:52.106860 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0423 01:44:52.106881 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0423 01:44:52.106900 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0423 01:44:52.106919 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0423 01:44:52.106937 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0423 01:44:52.106955 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0423 01:44:52.106980 135404902385472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0423 01:44:52.354459 135404902385472 metric_logger.py:196] completed step: 10, seconds: 2.123, TFLOP/s/device: 6.399, Tokens/s/device: 964.582, total_weights: 16384, loss: 3.602, lm_loss: 3.602, perplexity: 36.686 I0423 01:44:52.472231 135404902385472 metric_logger.py:196] completed step: 11, seconds: 2.739, TFLOP/s/device: 4.960, Tokens/s/device: 747.649, total_weights: 16384, loss: 3.101, lm_loss: 3.101, perplexity: 22.211 I0423 01:44:52.591212 135404902385472 metric_logger.py:196] completed step: 12, seconds: 0.006, TFLOP/s/device: 2182.321, Tokens/s/device: 328943.142, total_weights: 16384, loss: 2.684, lm_loss: 2.684, perplexity: 14.644 I0423 01:44:52.710028 135404902385472 metric_logger.py:196] completed step: 13, seconds: 0.119, TFLOP/s/device: 114.555, Tokens/s/device: 17266.963, total_weights: 16384, loss: 2.348, lm_loss: 2.348, perplexity: 10.461 I0423 01:44:52.715428 135404902385472 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0423 01:44:52.874492 135372772214336 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_08_checkpoint_async_true/checkpoints/10/items/array_metadatas/process_0 I0423 01:44:52.947312 135404902385472 checkpointing.py:776] Waited 0.2318401336669922 seconds for step 15 to finish before starting checkpointing. I0423 01:44:52.948275 135404902385472 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 01:45:31.606701 135372793185856 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 39.501676s (commit=39.033218s, array_metadata_write=0.468458s) I0423 01:45:31.613524 135372803671616 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 302.356 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 41 seconds) (per-host) I0423 01:45:31.613586 135372803671616 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.507500s. I0423 01:45:32.103170 135372803671616 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 01:45:32.327431 135372761728576 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776908690537925968, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10/_CHECKPOINT_METADATA I0423 01:45:32.480356 135372803671616 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0423 01:45:32.480550 135372803671616 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.546566s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10/items I0423 01:45:32.481445 135372803671616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10/items I0423 01:45:32.720344 135372803671616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10 I0423 01:45:33.430274 135372803671616 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_08_checkpoint_async_true/checkpoints/10`. I0423 01:45:33.431012 135372803671616 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 43.699548s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/10 I0423 01:45:33.431102 135372803671616 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 41.325017s. I0423 01:45:33.431305 135372952569408 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 01:45:33.431424 135372952569408 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 01:45:33.431488 135372952569408 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0423 01:45:33.431535 135372952569408 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0423 01:45:33.431704 135404902385472 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 01:45:33.431845 135404902385472 checkpoint_manager.py:1441] Waiting for previous save to complete took 40.483572 seconds. If this number is high, consider checkpointing less frequently. I0423 01:45:33.432922 135404902385472 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 15 I0423 01:45:33.433252 135404902385472 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15. I0423 01:45:33.602455 135404902385472 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 01:45:33.602629 135404902385472 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 01:45:33.603779 135372803671616 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15 I0423 01:45:34.278326 135372772214336 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15/items I0423 01:45:35.244947 135404902385472 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.643078s I0423 01:45:35.245520 135404902385472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 7.139 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0423 01:45:35.245574 135404902385472 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.728629s (batch_requests_ready=0.081064s, total_serialization_initiated=1.647132s, others=0.000433s) I0423 01:45:35.245642 135404902385472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.729192s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=1.729176) I0423 01:45:35.250716 135372803671616 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 01:45:35.250824 135404902385472 async_checkpointer.py:561] Finished blocking save. Time taken: 1.817858s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15. I0423 01:45:35.251020 135404902385472 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0423 01:45:35.251196 135372952569408 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 01:45:35.251313 135404902385472 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776908692.94825, 'wait_for_prev_duration_secs': 40.48357176780701, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776908733.4329453, 'checkpointer_blocking_duration_secs': 1.8179867267608643, 'get_old_steps_start_time': 1776908735.250949, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776908692.9481819, 'checkpoint_manager_blocking_duration_secs': 42.303107023239136} I0423 01:45:35.251406 135404902385472 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0423 01:45:35.252101 135404902385472 metric_logger.py:196] completed step: 14, seconds: 0.119, TFLOP/s/device: 114.098, Tokens/s/device: 17198.089, total_weights: 16384, loss: 2.083, lm_loss: 2.083, perplexity: 8.030 I0423 01:45:35.261182 135404902385472 metric_logger.py:196] completed step: 15, seconds: 0.119, TFLOP/s/device: 114.251, Tokens/s/device: 17221.227, total_weights: 16384, loss: 1.880, lm_loss: 1.880, perplexity: 6.551 I0423 01:45:35.377852 135404902385472 metric_logger.py:196] completed step: 16, seconds: 42.544, TFLOP/s/device: 0.319, Tokens/s/device: 48.139, total_weights: 16384, loss: 1.726, lm_loss: 1.726, perplexity: 5.616 I0423 01:45:35.496647 135404902385472 metric_logger.py:196] completed step: 17, seconds: 0.006, TFLOP/s/device: 2140.717, Tokens/s/device: 322672.129, total_weights: 16384, loss: 1.610, lm_loss: 1.610, perplexity: 5.002 I0423 01:45:35.616025 135404902385472 metric_logger.py:196] completed step: 18, seconds: 0.118, TFLOP/s/device: 115.013, Tokens/s/device: 17335.952, total_weights: 16384, loss: 1.521, lm_loss: 1.521, perplexity: 4.577 I0423 01:45:35.734092 135404902385472 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0423 01:45:35.735227 135404902385472 checkpointing.py:776] Waited 0.0012099742889404297 seconds for step 19 to finish before starting checkpointing. I0423 01:45:35.735483 135404902385472 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 01:45:35.976577 135375018264128 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_08_checkpoint_async_true/checkpoints/15/items/array_metadatas/process_0 I0423 01:46:05.880474 135372793185856 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 30.630865s (commit=30.171520s, array_metadata_write=0.459345s) I0423 01:46:05.881598 135372803671616 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 390.451 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 32 seconds) (per-host) I0423 01:46:05.881713 135372803671616 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.630842s. I0423 01:46:06.362765 135372803671616 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 01:46:06.787934 135372803671616 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0423 01:46:06.788138 135372803671616 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.595268s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15/items I0423 01:46:06.788975 135372803671616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15/items I0423 01:46:07.015394 135372803671616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15 I0423 01:46:07.709966 135372803671616 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_08_checkpoint_async_true/checkpoints/15`. I0423 01:46:07.710695 135372803671616 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 34.277736s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/15 I0423 01:46:07.710762 135372803671616 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 32.459894s. I0423 01:46:07.710962 135372952569408 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 01:46:07.711094 135372952569408 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 01:46:07.711147 135372952569408 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0423 01:46:07.711185 135372952569408 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0423 01:46:07.711369 135404902385472 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 01:46:07.711492 135404902385472 checkpoint_manager.py:1441] Waiting for previous save to complete took 31.976020 seconds. If this number is high, consider checkpointing less frequently. I0423 01:46:07.711776 135404902385472 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 19 I0423 01:46:07.712122 135404902385472 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19. I0423 01:46:07.876451 135372803671616 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19 I0423 01:46:07.881384 135404902385472 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 01:46:07.881483 135404902385472 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 01:46:08.583872 135291700512320 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19/items I0423 01:46:10.487456 135404902385472 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.606543s I0423 01:46:10.491889 135404902385472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.572 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0423 01:46:10.491950 135404902385472 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.699080s (batch_requests_ready=0.085140s, total_serialization_initiated=2.609628s, others=0.004312s) I0423 01:46:10.492025 135404902385472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.699600s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=2.699587) I0423 01:46:10.493962 135372952569408 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 01:46:10.494093 135404902385472 async_checkpointer.py:561] Finished blocking save. Time taken: 2.782263s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19. I0423 01:46:10.494325 135404902385472 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0423 01:46:10.494482 135372803671616 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 01:46:10.494583 135404902385472 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776908735.7354584, 'wait_for_prev_duration_secs': 31.976019859313965, 'time_between_consecutive_saves_sec': 2.303908586502075, 'checkpointer_blocking_start_time': 1776908767.7118006, 'checkpointer_blocking_duration_secs': 2.78244948387146, 'get_old_steps_start_time': 1776908770.4942608, 'get_old_steps_duration_secs': 2.1457672119140625e-05, 'checkpoint_manager_blocking_start_time': 1776908735.7354279, 'checkpoint_manager_blocking_duration_secs': 34.75913071632385} I0423 01:46:10.494700 135404902385472 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0423 01:46:10.494735 135404902385472 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 01:46:11.227515 135375018264128 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_08_checkpoint_async_true/checkpoints/19/items/array_metadatas/process_0 I0423 01:46:48.672256 135291700512320 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.179711s (commit=37.728660s, array_metadata_write=0.451051s) I0423 01:46:48.673433 135372952569408 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 309.115 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 40 seconds) (per-host) I0423 01:46:48.673554 135372952569408 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.179401s. I0423 01:46:49.162287 135372952569408 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 01:46:49.556680 135372952569408 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0423 01:46:49.556875 135372952569408 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.557552s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19/items I0423 01:46:49.557658 135372952569408 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19/items I0423 01:46:49.797887 135372952569408 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19 I0423 01:46:50.482466 135372952569408 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_08_checkpoint_async_true/checkpoints/19`. I0423 01:46:50.483191 135372952569408 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 42.771379s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_08_checkpoint_async_true/checkpoints/19 I0423 01:46:50.483257 135372952569408 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 39.989105s. I0423 01:46:50.483463 135372803671616 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 01:46:50.483586 135372803671616 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 01:46:50.483635 135372803671616 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0423 01:46:50.483673 135372803671616 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0423 01:46:50.483786 135404902385472 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 01:46:50.483920 135404902385472 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 01:46:50.484747 135404902385472 metric_logger.py:196] completed step: 19, seconds: 0.119, TFLOP/s/device: 114.309, Tokens/s/device: 17229.920, 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