2026-04-16 00:11:52.393711: 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) I0416 00:11:52.511178 136824743865472 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0416 00:12:51.469554 136824743865472 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 00:12:51.469666 136824743865472 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 00:12:51.469702 136824743865472 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Apr 6 2026 20:48:10 (1775533690) cl/895581894 I0416 00:12:51.469727 136824743865472 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 00:12:51.469748 136824743865472 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 00:12:51.469827 136824743865472 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 00:12:51.569921 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:12:51.967401 136824743865472 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 00:12:51.967906 136824743865472 checkpointing.py:677] Setting up checkpoint logger... I0416 00:12:51.967957 136824743865472 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True I0416 00:12:51.968037 136824743865472 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 00:12:51.969025 136824743865472 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7c70675f8a10>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 00:12:53.651937 136824743865472 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 00:12:53.652207 136824743865472 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 0x7c69eb331f40>}, handler_registry=None I0416 00:12:53.652597 136824743865472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c69eb331f40>` 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`. I0416 00:12:53.652639 136824743865472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6b9b8e6cf0>` 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`. I0416 00:12:53.652669 136824743865472 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 0x7c69eb331f40>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c69eb331f40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6b9b8e6cf0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6b9b8e6cf0>}). I0416 00:12:53.652995 136824743865472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 00:12:53.653064 136824743865472 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7c69eb3394e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 00:12:54.155608 136718640481856 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776298244308332609, 'commit_timestamp_nsecs': 1776298261968187016, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/5/_CHECKPOINT_METADATA I0416 00:12:54.236057 136718640481856 checkpoint_manager.py:1847] Missing metrics for step 5 I0416 00:12:54.282395 136824743865472 checkpoint_manager.py:1788] Found 3 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints I0416 00:12:54.282896 136824743865472 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_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c6bb53949b0> I0416 00:12:54.282993 136824743865472 checkpointing.py:301] Checkpoint manager created! W0416 00:12:54.300022 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:12:54.480669 136824743865472 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 00:12:54.480769 136824743865472 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:12:54.577673 136824743865472 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:12:54.577753 136824743865472 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:12:54.591824 136824743865472 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:12:54.591870 136824743865472 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:12:54.611978 136824743865472 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 00:12:54.612035 136824743865472 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:12:54.626148 136824743865472 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 00:12:54.626197 136824743865472 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:12:54.640272 136824743865472 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 00:12:54.640321 136824743865472 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:12:54.661709 136824743865472 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 00:12:54.661764 136824743865472 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:12:54.680520 136824743865472 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 00:12:54.680573 136824743865472 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 00:12:54.860116 136824743865472 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 00:12:54.860209 136824743865472 checkpointing.py:581] restoring from this run's directory step 9 I0416 00:12:55.027652 136824743865472 checkpointer.py:304] Restoring checkpoint from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/9. I0416 00:12:55.957650 3418394 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 00:12:55.957702 3418394 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 00:13:03.727363 136824743865472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/read/gbytes_per_sec: 1.488 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host) I0416 00:13:03.727833 136824743865472 checkpointer.py:318] Finished restoring checkpoint in 8.78 seconds from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/9. I0416 00:13:03.727928 136824743865472 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776298374.9525495, 'checkpointer_duration_secs': 8.775334596633911, 'checkpoint_manager_start_time': 1776298374.8604531, 'checkpoint_manager_duration_secs': 8.867433547973633} [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 I0416 00:13:04.099640 136824743865472 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 00:13:04.099749 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0416 00:13:04.099791 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0416 00:13:04.099834 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0416 00:13:04.099872 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0416 00:13:04.099899 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0416 00:13:04.099943 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0416 00:13:04.099983 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0416 00:13:04.100011 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0416 00:13:04.100037 136824743865472 maxtext_utils.py:1623] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0416 00:13:04.100092 136824743865472 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0416 00:13:04.100126 136824743865472 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0416 00:13:04.938846 136824743865472 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:13:04.938930 136824743865472 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 00:13:04.952383 136824743865472 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:13:04.952435 136824743865472 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 00:13:05.199138 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:13:05.280686 136824743865472 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. I0416 00:13:05.281197 136824743865472 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 00:13:05.281340 136824743865472 metric_logger.py:289] number parameters: 1.104 billion W0416 00:13:06.118822 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:13:06.190142 136824743865472 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0416 00:13:06.306441 136824743865472 checkpointing.py:776] Waited 0.11629009246826172 seconds for step 10 to finish before starting checkpointing. I0416 00:13:06.306832 136824743865472 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:13:06.306994 136824743865472 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 10 I0416 00:13:06.307301 136824743865472 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10. I0416 00:13:06.389880 136824743865472 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 00:13:06.479922 136717010994752 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10 I0416 00:13:06.486166 136824743865472 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:13:06.486255 136824743865472 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 W0416 00:13:06.699363 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.726822 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.733295 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.737878 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.742034 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.746588 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.750670 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:06.754711 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:13:07.207628 136717000508992 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/items I0416 00:13:07.339223 136716969051712 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776298387117402936, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776298387117402936, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0416 00:13:13.843029 136824743865472 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 7.357266s I0416 00:13:13.847272 136824743865472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.655 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host) I0416 00:13:13.847337 136824743865472 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 7.456539s (batch_requests_ready=0.091975s, total_serialization_initiated=7.360486s, others=0.004079s) I0416 00:13:13.847421 136824743865472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 7.457077s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=7.457058) I0416 00:13:13.848929 136717010994752 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:13:13.849028 136824743865472 async_checkpointer.py:561] Finished blocking save. Time taken: 7.541994s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10. I0416 00:13:13.849254 136824743865472 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:13:13.849459 136717021480512 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:13:13.849563 136824743865472 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776298386.3068197, 'wait_for_prev_duration_secs': 3.838539123535156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776298386.307015, 'checkpointer_blocking_duration_secs': 7.542140960693359, 'get_old_steps_start_time': 1776298393.8491867, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776298386.306738, 'checkpoint_manager_blocking_duration_secs': 7.542798042297363} I0416 00:13:13.849691 136824743865472 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0416 00:13:13.849742 136824743865472 max_utils.py:750] Memstats: After params initialized: I0416 00:13:13.849790 136824743865472 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0)) I0416 00:13:13.849818 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 00:13:13.849840 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 00:13:13.849860 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 00:13:13.849878 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 00:13:13.849894 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 00:13:13.849912 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 00:13:13.849930 136824743865472 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 00:13:13.856427 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:13.862918 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:13:13.867682 3416729 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:13:14.055663 136824743865472 metric_logger.py:185] completed step: 10, seconds: 0.908, TFLOP/s/device: 14.958, Tokens/s/device: 2254.702, total_weights: 16384, loss: 3.602 I0416 00:13:14.172650 136824743865472 metric_logger.py:185] completed step: 11, seconds: 7.864, TFLOP/s/device: 1.728, Tokens/s/device: 260.430, total_weights: 16384, loss: 3.101 I0416 00:13:14.291184 136824743865472 metric_logger.py:185] completed step: 12, seconds: 0.007, TFLOP/s/device: 2071.209, Tokens/s/device: 312195.122, total_weights: 16384, loss: 2.684 I0416 00:13:14.410114 136824743865472 metric_logger.py:185] completed step: 13, seconds: 0.120, TFLOP/s/device: 112.990, Tokens/s/device: 17031.043, total_weights: 16384, loss: 2.348 I0416 00:13:14.417949 136824743865472 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0416 00:13:14.614147 136716979537472 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_0 I0416 00:13:14.648569 136824743865472 checkpointing.py:776] Waited 0.2305915355682373 seconds for step 15 to finish before starting checkpointing. I0416 00:13:14.649021 136824743865472 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:13:21.181374 136717000508992 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 7.333305s (commit=6.899734s, array_metadata_write=0.433571s) I0416 00:13:21.182269 136717010994752 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 854.331 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 14 seconds) (per-host) I0416 00:13:21.182320 136717010994752 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 7.333232s. W0416 00:13:21.518352 136717010994752 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/items, I0416 00:13:21.606323 136717010994752 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 00:13:21.606431 136717010994752 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.144929s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/items I0416 00:13:21.607064 136717010994752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/items I0416 00:13:21.783937 136716969051712 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776298387117402936, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0416 00:13:21.854123 136717010994752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10 I0416 00:13:22.522349 136717010994752 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10`. I0416 00:13:22.522961 136717010994752 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 16.215935s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/10 I0416 00:13:22.523032 136717010994752 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.673946s. I0416 00:13:22.523202 136717021480512 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:13:22.523302 136717021480512 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:13:22.523353 136717021480512 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:13:22.523400 136717021480512 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0416 00:13:22.524437 136824743865472 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. W0416 00:13:22.524889 136824743865472 checkpoint_manager.py:1441] Waiting for previous save to complete took 7.875879 seconds. If this number is high, consider checkpointing less frequently. I0416 00:13:22.525794 136824743865472 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 15 I0416 00:13:22.526094 136824743865472 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15. I0416 00:13:22.706028 136824743865472 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:13:22.706123 136824743865472 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 I0416 00:13:22.724936 136717021480512 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15 I0416 00:13:23.420115 136717000508992 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15/items I0416 00:13:32.184139 136824743865472 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 9.478490s I0416 00:13:32.188409 136824743865472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.289 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 9 seconds) (per-host) I0416 00:13:32.188483 136824743865472 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 9.574333s (batch_requests_ready=0.088076s, total_serialization_initiated=9.482075s, others=0.004182s) I0416 00:13:32.188565 136824743865472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 9.574855s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=9.574840) I0416 00:13:32.189711 136717021480512 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:13:32.189803 136824743865472 async_checkpointer.py:561] Finished blocking save. Time taken: 9.663966s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15. I0416 00:13:32.190083 136824743865472 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:13:32.190228 136717010994752 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:13:32.190299 136824743865472 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776298394.6489942, 'wait_for_prev_duration_secs': 7.875879287719727, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776298402.5258183, 'checkpointer_blocking_duration_secs': 9.66417407989502, 'get_old_steps_start_time': 1776298412.1900086, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1776298394.6489239, 'checkpoint_manager_blocking_duration_secs': 17.541350841522217} I0416 00:13:32.190406 136824743865472 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0416 00:13:32.190874 136824743865472 metric_logger.py:185] completed step: 14, seconds: 0.118, TFLOP/s/device: 115.346, Tokens/s/device: 17386.137, total_weights: 16384, loss: 2.083 I0416 00:13:32.198372 136824743865472 metric_logger.py:185] completed step: 15, seconds: 0.119, TFLOP/s/device: 114.080, Tokens/s/device: 17195.345, total_weights: 16384, loss: 1.880 I0416 00:13:32.314328 136824743865472 metric_logger.py:185] completed step: 16, seconds: 17.779, TFLOP/s/device: 0.764, Tokens/s/device: 115.195, total_weights: 16384, loss: 1.726 I0416 00:13:32.433311 136824743865472 metric_logger.py:185] completed step: 17, seconds: 0.006, TFLOP/s/device: 2398.011, Tokens/s/device: 361454.289, total_weights: 16384, loss: 1.610 I0416 00:13:32.552276 136824743865472 metric_logger.py:185] completed step: 18, seconds: 0.117, TFLOP/s/device: 116.395, Tokens/s/device: 17544.311, total_weights: 16384, loss: 1.521 I0416 00:13:32.671159 136824743865472 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0416 00:13:32.672686 136824743865472 checkpointing.py:776] Waited 0.0015540122985839844 seconds for step 19 to finish before starting checkpointing. I0416 00:13:32.673091 136824743865472 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:13:32.944661 136716979537472 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_0 I0416 00:13:38.688788 136717000508992 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 6.499643s (commit=6.080899s, array_metadata_write=0.418743s) I0416 00:13:38.689751 136717021480512 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 786.088 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 16 seconds) (per-host) I0416 00:13:38.689865 136717021480512 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 6.500015s. W0416 00:13:39.031614 136717021480512 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15/items, I0416 00:13:39.114865 136717021480512 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 00:13:39.114976 136717021480512 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.135037s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15/items I0416 00:13:39.115620 136717021480512 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15/items I0416 00:13:39.356394 136717021480512 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15 I0416 00:13:40.037035 136717021480512 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15`. I0416 00:13:40.037695 136717021480512 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 17.511864s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/15 I0416 00:13:40.037770 136717021480512 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 7.847924s. I0416 00:13:40.037884 136717010994752 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:13:40.037933 136717010994752 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:13:40.037981 136717010994752 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:13:40.038017 136717010994752 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0416 00:13:40.038216 136824743865472 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. W0416 00:13:40.038388 136824743865472 checkpoint_manager.py:1441] Waiting for previous save to complete took 7.365311 seconds. If this number is high, consider checkpointing less frequently. I0416 00:13:40.038782 136824743865472 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 19 I0416 00:13:40.039194 136824743865472 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19. I0416 00:13:40.205879 136824743865472 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:13:40.205965 136824743865472 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 I0416 00:13:40.207555 136717010994752 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19 I0416 00:13:40.900173 136717000508992 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19/items I0416 00:13:52.200961 136824743865472 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.995453s I0416 00:13:52.206066 136824743865472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.021 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0416 00:13:52.206141 136824743865472 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.084442s (batch_requests_ready=0.080609s, total_serialization_initiated=11.998912s, others=0.004921s) I0416 00:13:52.206229 136824743865472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.084935s (all_items=0.000010s, per_item={'items': '0.00001025'}, temp_paths=12.084925) I0416 00:13:52.208271 136717010994752 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:13:52.208423 136824743865472 async_checkpointer.py:561] Finished blocking save. Time taken: 12.169516s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19. I0416 00:13:52.208683 136824743865472 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:13:52.208918 136717021480512 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:13:52.209108 136824743865472 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776298412.6730568, 'wait_for_prev_duration_secs': 7.365310907363892, 'time_between_consecutive_saves_sec': 10.149641752243042, 'checkpointer_blocking_start_time': 1776298420.0388825, 'checkpointer_blocking_duration_secs': 12.16968584060669, 'get_old_steps_start_time': 1776298432.2085843, 'get_old_steps_duration_secs': 4.863739013671875e-05, 'checkpoint_manager_blocking_start_time': 1776298412.6729941, 'checkpoint_manager_blocking_duration_secs': 19.536075353622437} I0416 00:13:52.209281 136824743865472 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0416 00:13:52.209318 136824743865472 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:13:52.974907 136716990023232 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_0 I0416 00:13:58.447077 136717000508992 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 6.239868s (commit=5.803645s, array_metadata_write=0.436223s) I0416 00:13:58.448435 136717010994752 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 689.531 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 18 seconds) (per-host) I0416 00:13:58.448611 136717010994752 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 6.240129s. W0416 00:13:58.771078 136717010994752 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19/items, I0416 00:13:58.884886 136717010994752 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 00:13:58.885040 136717010994752 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.164210s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19/items I0416 00:13:58.885947 136717010994752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19/items I0416 00:13:59.110743 136717010994752 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19 I0416 00:13:59.769725 136717010994752 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19`. I0416 00:13:59.770670 136717010994752 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 19.731773s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_false/checkpoints/19 I0416 00:13:59.770745 136717010994752 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 7.562266s. I0416 00:13:59.771021 136717021480512 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:13:59.771192 136717021480512 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:13:59.771244 136717021480512 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:13:59.771296 136717021480512 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0416 00:13:59.771489 136824743865472 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. I0416 00:13:59.771686 136824743865472 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:13:59.772455 136824743865472 metric_logger.py:185] completed step: 19, seconds: 0.123, TFLOP/s/device: 110.672, Tokens/s/device: 16681.736, total_weights: 16384, loss: 1.449 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops