2026-04-16 00:25:58.764641: 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 00:25:58.880852 137179156421760 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0416 00:26:55.890850 137179156421760 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 00:26:55.890955 137179156421760 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 00:26:55.890989 137179156421760 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 00:26:55.891013 137179156421760 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 00:26:55.891038 137179156421760 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 00:26:55.891126 137179156421760 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 00:26:55.981535 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:26:56.379532 137179156421760 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 00:26:56.379929 137179156421760 checkpointing.py:677] Setting up checkpoint logger... I0416 00:26:56.379983 137179156421760 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0416 00:26:56.380022 137179156421760 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 00:26:56.380655 137179156421760 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7cc2ec9fcce0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 00:26:58.045145 137179156421760 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 00:26:58.045318 137179156421760 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 0x7cbc6fd39430>}, handler_registry=None I0416 00:26:58.045567 137179156421760 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cbc6fd39430>` 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 00:26:58.045608 137179156421760 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cbc6fd39850>` 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 00:26:58.045639 137179156421760 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 0x7cbc6fd39430>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cbc6fd39430>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cbc6fd39850>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cbc6fd39850>}). I0416 00:26:58.045943 137179156421760 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 00:26:58.045995 137179156421760 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7cbc6fd3d4e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 00:26:58.475387 137073042392640 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776299041887443546, 'commit_timestamp_nsecs': 1776299080176258681, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA I0416 00:26:58.545374 137073042392640 checkpoint_manager.py:1847] Missing metrics for step 5 I0416 00:26:58.634407 137179156421760 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_16_zarr3_false/checkpoints I0416 00:26:58.634905 137179156421760 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_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cbc70527380> I0416 00:26:58.634990 137179156421760 checkpointing.py:301] Checkpoint manager created! W0416 00:26:58.651568 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:26:58.833641 137179156421760 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 00:26:58.833746 137179156421760 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:26:58.920928 137179156421760 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:26:58.921009 137179156421760 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:26:58.935237 137179156421760 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:26:58.935286 137179156421760 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:26:58.955667 137179156421760 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 00:26:58.955725 137179156421760 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:26:58.969876 137179156421760 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 00:26:58.969924 137179156421760 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:26:58.983918 137179156421760 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 00:26:58.983966 137179156421760 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:26:59.017568 137179156421760 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 00:26:59.017722 137179156421760 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:26:59.037128 137179156421760 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 00:26:59.037179 137179156421760 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 00:26:59.220227 137179156421760 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 00:26:59.220332 137179156421760 checkpointing.py:581] restoring from this run's directory step 9 I0416 00:26:59.376842 137179156421760 checkpointer.py:304] Restoring checkpoint from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9. I0416 00:26:59.895199 3430863 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 00:26:59.895265 3430863 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 00:27:07.974437 137179156421760 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/read/gbytes_per_sec: 1.493 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host) I0416 00:27:07.974741 137179156421760 checkpointer.py:318] Finished restoring checkpoint in 8.67 seconds from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9. I0416 00:27:07.974833 137179156421760 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_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776299219.301346, 'checkpointer_duration_secs': 8.67344617843628, 'checkpoint_manager_start_time': 1776299219.220636, 'checkpoint_manager_duration_secs': 8.754160165786743} [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 00:27:08.350872 137179156421760 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 00:27:08.350981 137179156421760 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 00:27:08.351022 137179156421760 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 00:27:08.351084 137179156421760 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 00:27:08.351127 137179156421760 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 00:27:08.351155 137179156421760 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 00:27:08.351193 137179156421760 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 00:27:08.351231 137179156421760 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 00:27:08.351261 137179156421760 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 00:27:08.351286 137179156421760 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 00:27:08.351320 137179156421760 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0416 00:27:08.351353 137179156421760 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0416 00:27:08.785728 137179156421760 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:27:08.785808 137179156421760 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 00:27:08.799197 137179156421760 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:27:08.799245 137179156421760 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 00:27:09.459508 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:27:09.527972 137179156421760 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 00:27:09.528481 137179156421760 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 00:27:09.528620 137179156421760 metric_logger.py:289] number parameters: 1.104 billion W0416 00:27:10.360656 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:27:10.428651 137179156421760 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0416 00:27:10.544695 137179156421760 checkpointing.py:776] Waited 0.11603140830993652 seconds for step 10 to finish before starting checkpointing. I0416 00:27:10.545124 137179156421760 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:27:10.545294 137179156421760 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 10 I0416 00:27:10.545579 137179156421760 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_16_zarr3_false/checkpoints/10. I0416 00:27:10.633029 137179156421760 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 00:27:10.718571 137071482111552 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10 I0416 00:27:10.723078 137179156421760 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:27:10.723168 137179156421760 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 00:27:10.935761 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.951088 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.959929 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.967611 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.971838 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.975880 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.979855 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:10.983964 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:27:11.452963 137071471625792 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10/items I0416 00:27:11.599092 137071440168512 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776299231361694533, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776299231361694533, "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_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA I0416 00:27:14.537354 137179156421760 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 3.814702s I0416 00:27:14.541275 137179156421760 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.159 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 3 seconds) (per-host) I0416 00:27:14.541335 137179156421760 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 3.907112s (batch_requests_ready=0.085466s, total_serialization_initiated=3.817829s, others=0.003818s) I0416 00:27:14.541402 137179156421760 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 3.907905s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=3.907886) I0416 00:27:14.543065 137071461140032 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:27:14.543163 137179156421760 async_checkpointer.py:561] Finished blocking save. Time taken: 3.997828s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10. I0416 00:27:14.543356 137179156421760 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:27:14.543518 137071492597312 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:27:14.543608 137179156421760 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776299230.5451086, 'wait_for_prev_duration_secs': 4.363059997558594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776299230.545316, 'checkpointer_blocking_duration_secs': 3.9979615211486816, 'get_old_steps_start_time': 1776299234.5432954, 'get_old_steps_duration_secs': 2.3126602172851562e-05, 'checkpoint_manager_blocking_start_time': 1776299230.5449915, 'checkpoint_manager_blocking_duration_secs': 3.998591423034668} I0416 00:27:14.543714 137179156421760 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0416 00:27:14.543753 137179156421760 max_utils.py:750] Memstats: After params initialized: I0416 00:27:14.543799 137179156421760 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0)) I0416 00:27:14.543823 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 00:27:14.543843 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 00:27:14.543860 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 00:27:14.543884 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 00:27:14.543900 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 00:27:14.543917 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 00:27:14.543933 137179156421760 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 00:27:14.549866 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:14.555833 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:27:14.560031 3429228 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:27:14.747335 137179156421760 metric_logger.py:185] completed step: 10, seconds: 0.900, TFLOP/s/device: 15.104, Tokens/s/device: 2276.704, total_weights: 16384, loss: 3.602 I0416 00:27:14.863815 137179156421760 metric_logger.py:185] completed step: 11, seconds: 4.317, TFLOP/s/device: 3.147, Tokens/s/device: 474.387, total_weights: 16384, loss: 3.101 I0416 00:27:14.982107 137179156421760 metric_logger.py:185] completed step: 12, seconds: 0.005, TFLOP/s/device: 2777.418, Tokens/s/device: 418642.682, total_weights: 16384, loss: 2.684 I0416 00:27:15.100676 137179156421760 metric_logger.py:185] completed step: 13, seconds: 0.117, TFLOP/s/device: 115.825, Tokens/s/device: 17458.464, total_weights: 16384, loss: 2.348 I0416 00:27:15.105164 137179156421760 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0416 00:27:15.312739 137071450654272 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_16_zarr3_false/checkpoints/10/items/array_metadatas/process_0 I0416 00:27:15.337433 137179156421760 checkpointing.py:776] Waited 0.23223447799682617 seconds for step 15 to finish before starting checkpointing. I0416 00:27:15.337834 137179156421760 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:27:43.792553 137071471625792 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.250214s (commit=28.791688s, array_metadata_write=0.458526s) I0416 00:27:43.793722 137071461140032 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 381.092 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 33 seconds) (per-host) I0416 00:27:43.793830 137071461140032 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.250621s. I0416 00:27:44.215202 137071461140032 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 00:27:44.400842 137071440168512 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776299231361694533, '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_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA I0416 00:27:44.808875 137071461140032 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.737812s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10/items I0416 00:27:44.809545 137071461140032 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10/items I0416 00:27:45.068331 137071461140032 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10 I0416 00:27:45.729003 137071461140032 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_16_zarr3_false/checkpoints/10`. I0416 00:27:45.729624 137071461140032 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 35.184297s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/10 I0416 00:27:45.729711 137071461140032 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 31.186503s. I0416 00:27:45.729888 137071492597312 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:27:45.729998 137071492597312 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:27:45.730079 137071492597312 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:27:45.730129 137071492597312 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0416 00:27:45.730467 137179156421760 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 00:27:45.731284 137179156421760 checkpoint_manager.py:1441] Waiting for previous save to complete took 30.393449 seconds. If this number is high, consider checkpointing less frequently. I0416 00:27:45.732372 137179156421760 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 15 I0416 00:27:45.732670 137179156421760 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_16_zarr3_false/checkpoints/15. I0416 00:27:45.907905 137179156421760 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:27:45.907996 137179156421760 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 00:27:45.915715 137071492597312 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15 I0416 00:27:46.616160 137071429682752 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15/items I0416 00:27:51.910152 137179156421760 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 6.002649s I0416 00:27:51.913988 137179156421760 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.024 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 6 seconds) (per-host) I0416 00:27:51.914062 137179156421760 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 6.096153s (batch_requests_ready=0.086464s, total_serialization_initiated=6.005955s, others=0.003734s) I0416 00:27:51.914135 137179156421760 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 6.096687s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=6.096672) I0416 00:27:51.915426 137071471625792 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:27:51.915508 137179156421760 async_checkpointer.py:561] Finished blocking save. Time taken: 6.183089s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15. I0416 00:27:51.915704 137179156421760 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:27:51.915904 137072958506560 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:27:51.916008 137179156421760 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776299235.3378088, 'wait_for_prev_duration_secs': 30.393449068069458, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776299265.732399, 'checkpointer_blocking_duration_secs': 6.1832098960876465, 'get_old_steps_start_time': 1776299271.9156303, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776299235.3377464, 'checkpoint_manager_blocking_duration_secs': 36.57823443412781} I0416 00:27:51.916156 137179156421760 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0416 00:27:51.916587 137179156421760 metric_logger.py:185] completed step: 14, seconds: 0.118, TFLOP/s/device: 114.694, Tokens/s/device: 17287.952, total_weights: 16384, loss: 2.083 I0416 00:27:51.924006 137179156421760 metric_logger.py:185] completed step: 15, seconds: 0.119, TFLOP/s/device: 114.599, Tokens/s/device: 17273.663, total_weights: 16384, loss: 1.880 I0416 00:27:52.039711 137179156421760 metric_logger.py:185] completed step: 16, seconds: 36.817, TFLOP/s/device: 0.369, Tokens/s/device: 55.626, total_weights: 16384, loss: 1.726 I0416 00:27:52.158494 137179156421760 metric_logger.py:185] completed step: 17, seconds: 0.005, TFLOP/s/device: 2505.926, Tokens/s/device: 377720.398, total_weights: 16384, loss: 1.610 I0416 00:27:52.276624 137179156421760 metric_logger.py:185] completed step: 18, seconds: 0.117, TFLOP/s/device: 116.416, Tokens/s/device: 17547.467, total_weights: 16384, loss: 1.521 I0416 00:27:52.394479 137179156421760 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0416 00:27:52.395384 137179156421760 checkpointing.py:776] Waited 0.0009241104125976562 seconds for step 19 to finish before starting checkpointing. I0416 00:27:52.395632 137179156421760 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:27:52.663271 137073010935360 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_16_zarr3_false/checkpoints/15/items/array_metadatas/process_0 I0416 00:28:17.375161 137071429682752 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 25.460471s (commit=25.054978s, array_metadata_write=0.405493s) I0416 00:28:17.376264 137071471625792 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 400.427 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 31 seconds) (per-host) I0416 00:28:17.376378 137071471625792 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 25.460827s. I0416 00:28:17.780524 137071471625792 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 00:28:18.335447 137071471625792 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.689623s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15/items I0416 00:28:18.336190 137071471625792 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15/items I0416 00:28:18.570651 137071471625792 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15 I0416 00:28:19.253282 137071471625792 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_16_zarr3_false/checkpoints/15`. I0416 00:28:19.253900 137071471625792 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 33.521486s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/15 I0416 00:28:19.253986 137071471625792 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 27.338437s. I0416 00:28:19.254156 137072958506560 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:28:19.254261 137072958506560 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:28:19.254325 137072958506560 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:28:19.254368 137072958506560 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0416 00:28:19.254505 137179156421760 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 00:28:19.254620 137179156421760 checkpoint_manager.py:1441] Waiting for previous save to complete took 26.858993 seconds. If this number is high, consider checkpointing less frequently. I0416 00:28:19.254836 137179156421760 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 19 I0416 00:28:19.255127 137179156421760 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_16_zarr3_false/checkpoints/19. I0416 00:28:19.432226 137072958506560 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19 I0416 00:28:19.440538 137179156421760 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:28:19.440632 137179156421760 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 00:28:20.163769 137072968992320 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19/items I0416 00:28:26.468086 137179156421760 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 7.027931s I0416 00:28:26.471849 137179156421760 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.732 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host) I0416 00:28:26.471909 137179156421760 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 7.123704s (batch_requests_ready=0.088719s, total_serialization_initiated=7.031326s, others=0.003659s) I0416 00:28:26.471980 137179156421760 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 7.124225s (all_items=0.000014s, per_item={'items': '0.00001431'}, temp_paths=7.124211) I0416 00:28:26.473337 137072958506560 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:28:26.473441 137179156421760 async_checkpointer.py:561] Finished blocking save. Time taken: 7.218563s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19. I0416 00:28:26.473665 137179156421760 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:28:26.473870 137072937535040 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:28:26.474099 137179156421760 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776299272.395601, 'wait_for_prev_duration_secs': 26.85899305343628, 'time_between_consecutive_saves_sec': 6.66545557975769, 'checkpointer_blocking_start_time': 1776299299.2548597, 'checkpointer_blocking_duration_secs': 7.218729734420776, 'get_old_steps_start_time': 1776299306.473601, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776299272.3955696, 'checkpoint_manager_blocking_duration_secs': 34.07850480079651} I0416 00:28:26.474227 137179156421760 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0416 00:28:26.474261 137179156421760 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:28:27.217748 137073010935360 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_16_zarr3_false/checkpoints/19/items/array_metadatas/process_0 I0416 00:29:02.980008 137072968992320 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.507695s (commit=36.058708s, array_metadata_write=0.448987s) I0416 00:29:02.980972 137072958506560 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 289.617 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 43 seconds) (per-host) I0416 00:29:02.981096 137072958506560 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.507606s. I0416 00:29:03.421058 137072958506560 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 00:29:03.910617 137072958506560 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.639650s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19/items I0416 00:29:03.911357 137072958506560 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19/items I0416 00:29:04.139375 137072958506560 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19 I0416 00:29:04.814242 137072958506560 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_16_zarr3_false/checkpoints/19`. I0416 00:29:04.814870 137072958506560 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 45.559997s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/19 I0416 00:29:04.814946 137072958506560 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 38.341454s. I0416 00:29:04.815135 137072937535040 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:29:04.815249 137072937535040 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:29:04.815316 137072937535040 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:29:04.815366 137072937535040 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0416 00:29:04.815523 137179156421760 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 00:29:04.815662 137179156421760 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:29:04.816367 137179156421760 metric_logger.py:185] completed step: 19, seconds: 0.119, TFLOP/s/device: 114.378, Tokens/s/device: 17240.363, 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