2026-04-16 00:01:51.541452: 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 00:01:51.657937 131032678485120 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0416 00:02:46.708449 131032678485120 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 00:02:46.708587 131032678485120 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 00:02:46.708628 131032678485120 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 00:02:46.708659 131032678485120 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 00:02:46.708683 131032678485120 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 00:02:46.708774 131032678485120 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 00:02:46.803766 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:02:47.208421 131032678485120 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 00:02:47.208796 131032678485120 checkpointing.py:677] Setting up checkpoint logger... I0416 00:02:47.208842 131032678485120 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 00:02:47.208906 131032678485120 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 00:02:47.209586 131032678485120 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x772bd55fbd40>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 00:02:49.551955 131032678485120 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 00:02:49.552202 131032678485120 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 0x772559332000>}, handler_registry=None I0416 00:02:49.552541 131032678485120 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x772559332000>` 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 00:02:49.552585 131032678485120 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7725593324e0>` 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 00:02:49.552616 131032678485120 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 0x772559332000>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x772559332000>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7725593324e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7725593324e0>}). I0416 00:02:49.552976 131032678485120 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 00:02:49.553032 131032678485120 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7725593394e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 00:02:49.673592 131032678485120 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints I0416 00:02:49.673832 131032678485120 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_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7725593309b0> I0416 00:02:49.673918 131032678485120 checkpointing.py:301] Checkpoint manager created! W0416 00:02:49.690728 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:02:49.871476 131032678485120 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 00:02:49.871650 131032678485120 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:02:49.966780 131032678485120 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:02:49.966859 131032678485120 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:02:49.981235 131032678485120 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:02:49.981283 131032678485120 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:02:50.002074 131032678485120 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 00:02:50.002129 131032678485120 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:02:50.016480 131032678485120 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 00:02:50.016530 131032678485120 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:02:50.030848 131032678485120 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 00:02:50.030901 131032678485120 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:02:50.052769 131032678485120 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 00:02:50.052826 131032678485120 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:02:50.071278 131032678485120 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 00:02:50.071331 131032678485120 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 00:02:50.255459 131032678485120 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 00:02:50.255563 131032678485120 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0416 00:02:50.305544 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. [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 00:02:50.686403 131032678485120 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 00:02:50.686512 131032678485120 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 00:02:50.686553 131032678485120 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 00:02:50.686597 131032678485120 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 00:02:50.686640 131032678485120 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 00:02:50.686668 131032678485120 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 00:02:50.686713 131032678485120 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 00:02:50.686755 131032678485120 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 00:02:50.686784 131032678485120 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 00:02:50.686814 131032678485120 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 00:02:50.686849 131032678485120 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0416 00:02:50.686882 131032678485120 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0416 00:02:51.109111 131032678485120 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:02:51.109192 131032678485120 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 00:02:51.122858 131032678485120 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:02:51.122911 131032678485120 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 00:02:51.807539 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:02:51.934852 131032678485120 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 00:02:51.935310 131032678485120 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 00:02:51.935449 131032678485120 metric_logger.py:289] number parameters: 1.104 billion W0416 00:02:52.788328 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:02:52.886152 131032678485120 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0416 00:02:53.003198 131032678485120 checkpointing.py:776] Waited 0.11703228950500488 seconds for step 0 to finish before starting checkpointing. I0416 00:02:53.003627 131032678485120 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:02:53.003787 131032678485120 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0416 00:02:53.004165 131032678485120 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_true/checkpoints/0. I0416 00:02:53.092980 131032678485120 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 00:02:53.170760 131032678485120 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:02:53.170843 131032678485120 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 00:02:53.184450 130925161612864 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_true/checkpoints/0 W0416 00:02:53.384883 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.396625 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.411889 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.420538 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.436289 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.443475 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.448857 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.453785 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:02:53.845744 131032678485120 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.675426s I0416 00:02:53.846177 131032678485120 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 16.405 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 752 milliseconds) (per-host) I0416 00:02:53.846227 131032678485120 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.752324s (batch_requests_ready=0.072718s, total_serialization_initiated=0.679255s, others=0.000351s) I0416 00:02:53.846284 131032678485120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.752848s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.752830) I0416 00:02:53.846950 130925098698304 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:02:53.847039 131032678485120 async_checkpointer.py:561] Finished blocking save. Time taken: 0.843213s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/0. I0416 00:02:53.847233 131032678485120 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:02:53.847430 131032678485120 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776297773.0036151, 'wait_for_prev_duration_secs': 3.933906555175781e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776297773.003809, 'checkpointer_blocking_duration_secs': 0.8433547019958496, 'get_old_steps_start_time': 1776297773.8471794, 'get_old_steps_duration_secs': 1.9550323486328125e-05, 'checkpoint_manager_blocking_start_time': 1776297773.003536, 'checkpoint_manager_blocking_duration_secs': 0.8438708782196045} I0416 00:02:53.847498 131032678485120 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0416 00:02:53.847536 131032678485120 max_utils.py:750] Memstats: After params initialized: I0416 00:02:53.847612 131032678485120 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0)) I0416 00:02:53.847638 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 00:02:53.847659 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 00:02:53.847676 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 00:02:53.847693 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 00:02:53.847709 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 00:02:53.847726 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 00:02:53.847743 131032678485120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 00:02:53.853460 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.858715 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:02:53.863587 3407695 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:02:53.876530 130925151127104 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_true/checkpoints/0/items I0416 00:02:54.020861 130925119669824 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776297773791735436, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776297773791735436, "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_true/checkpoints/0/_CHECKPOINT_METADATA I0416 00:02:54.021111 130925172098624 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:02:54.056126 131032678485120 metric_logger.py:185] completed step: 0, seconds: 0.950, TFLOP/s/device: 14.299, Tokens/s/device: 2155.227, total_weights: 16384, loss: 10.874 I0416 00:02:54.057033 131032678485120 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/tensorboard/' I0416 00:02:54.172602 131032678485120 metric_logger.py:185] completed step: 1, seconds: 1.168, TFLOP/s/device: 11.630, Tokens/s/device: 1752.976, total_weights: 16384, loss: 10.874 I0416 00:02:54.291510 131032678485120 metric_logger.py:185] completed step: 2, seconds: 0.005, TFLOP/s/device: 2515.669, Tokens/s/device: 379189.039, total_weights: 16384, loss: 10.276 I0416 00:02:54.351066 3409939 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 00:02:54.351127 3409939 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 00:02:54.409635 131032678485120 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 115.854, Tokens/s/device: 17462.781, total_weights: 16384, loss: 9.282 I0416 00:02:54.414008 131032678485120 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0416 00:02:54.646176 131032678485120 checkpointing.py:776] Waited 0.23213601112365723 seconds for step 5 to finish before starting checkpointing. I0416 00:02:54.646566 131032678485120 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:02:55.098140 130925161612864 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_true/checkpoints/0/items/array_metadatas/process_0 I0416 00:03:20.286594 130925109184064 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 25.939120s (commit=25.510977s, array_metadata_write=0.428144s) I0416 00:03:20.287472 130925098698304 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 464.698 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 27 seconds) (per-host) I0416 00:03:20.287523 130925098698304 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 26.440426s. I0416 00:03:20.502490 130925098698304 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776297773791735436, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0416 00:03:20.675095 130925098698304 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 00:03:20.864965 130925119669824 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776297773791735436, '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_true/checkpoints/0/_CHECKPOINT_METADATA I0416 00:03:21.109797 130925098698304 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 00:03:21.109965 130925098698304 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.573422s. 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_true/checkpoints/0/items I0416 00:03:21.110634 130925098698304 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/0/items I0416 00:03:21.361656 130925098698304 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/0 I0416 00:03:22.004625 130925098698304 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_true/checkpoints/0`. I0416 00:03:22.005261 130925098698304 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 29.001441s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/0 I0416 00:03:22.005331 130925098698304 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 28.158236s. I0416 00:03:22.005471 130925172098624 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:03:22.005569 130925172098624 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:03:22.005623 130925172098624 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:03:22.005665 130925172098624 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 00:03:22.005856 131032678485120 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0416 00:03:22.005972 131032678485120 checkpoint_manager.py:1441] Waiting for previous save to complete took 27.359405 seconds. If this number is high, consider checkpointing less frequently. I0416 00:03:22.006627 131032678485120 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0416 00:03:22.006895 131032678485120 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_true/checkpoints/5. I0416 00:03:22.190129 131032678485120 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:03:22.190217 131032678485120 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 00:03:22.191684 130925172098624 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_true/checkpoints/5 I0416 00:03:22.899609 130925058852416 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_true/checkpoints/5/items I0416 00:03:35.177844 131032678485120 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.988108s I0416 00:03:35.182955 131032678485120 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 965.833 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 13 seconds) (per-host) I0416 00:03:35.183018 131032678485120 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 13.083923s (batch_requests_ready=0.087920s, total_serialization_initiated=12.991068s, others=0.004936s) I0416 00:03:35.183110 131032678485120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 13.084452s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=13.084438) I0416 00:03:35.184656 130925048366656 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:03:35.184789 131032678485120 async_checkpointer.py:561] Finished blocking save. Time taken: 13.178117s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/5. I0416 00:03:35.185025 131032678485120 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:03:35.185215 130925098698304 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:03:35.185325 131032678485120 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776297774.6465445, 'wait_for_prev_duration_secs': 27.359405040740967, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776297802.0066535, 'checkpointer_blocking_duration_secs': 13.178267478942871, 'get_old_steps_start_time': 1776297815.1849456, 'get_old_steps_duration_secs': 3.4332275390625e-05, 'checkpoint_manager_blocking_start_time': 1776297774.6464796, 'checkpoint_manager_blocking_duration_secs': 40.538817405700684} I0416 00:03:35.185477 131032678485120 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0416 00:03:35.186136 131032678485120 metric_logger.py:185] completed step: 4, seconds: 0.119, TFLOP/s/device: 114.245, Tokens/s/device: 17220.214, total_weights: 16384, loss: 8.327 I0416 00:03:35.194986 131032678485120 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 115.318, Tokens/s/device: 17382.005, total_weights: 16384, loss: 7.367 I0416 00:03:35.310289 131032678485120 metric_logger.py:185] completed step: 6, seconds: 40.779, TFLOP/s/device: 0.333, Tokens/s/device: 50.222, total_weights: 16384, loss: 6.459 I0416 00:03:35.428321 131032678485120 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 2210.368, Tokens/s/device: 333170.652, total_weights: 16384, loss: 5.622 I0416 00:03:35.546501 131032678485120 metric_logger.py:185] completed step: 8, seconds: 0.116, TFLOP/s/device: 117.311, Tokens/s/device: 17682.458, total_weights: 16384, loss: 4.864 I0416 00:03:35.664435 131032678485120 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0416 00:03:35.665278 131032678485120 checkpointing.py:776] Waited 0.0008640289306640625 seconds for step 9 to finish before starting checkpointing. I0416 00:03:35.665518 131032678485120 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:03:35.922785 130925161612864 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_true/checkpoints/5/items/array_metadatas/process_0 I0416 00:04:05.493066 130925058852416 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 30.309387s (commit=29.818938s, array_metadata_write=0.490449s) I0416 00:04:05.494040 130925048366656 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 291.204 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 43 seconds) (per-host) I0416 00:04:05.494156 130925048366656 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.309321s. I0416 00:04:05.916013 130925048366656 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 00:04:06.349301 130925048366656 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 00:04:06.349895 130925048366656 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.570979s. 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_true/checkpoints/5/items I0416 00:04:06.350541 130925048366656 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/5/items I0416 00:04:06.609473 130925048366656 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/5 I0416 00:04:07.297533 130925048366656 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_true/checkpoints/5`. I0416 00:04:07.298190 130925048366656 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 45.291525s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/5 I0416 00:04:07.298259 130925048366656 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 32.113427s. I0416 00:04:07.298405 130925098698304 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:04:07.298504 130925098698304 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:04:07.298560 130925098698304 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:04:07.298599 130925098698304 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0416 00:04:07.299117 131032678485120 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0416 00:04:07.299228 131032678485120 checkpoint_manager.py:1441] Waiting for previous save to complete took 31.633722 seconds. If this number is high, consider checkpointing less frequently. I0416 00:04:07.299432 131032678485120 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0416 00:04:07.299723 131032678485120 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_true/checkpoints/9. I0416 00:04:07.471427 130925098698304 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_true/checkpoints/9 I0416 00:04:07.475598 131032678485120 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:04:07.475695 131032678485120 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 00:04:08.226523 130925109184064 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_true/checkpoints/9/items I0416 00:04:13.508063 131032678485120 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 6.032814s I0416 00:04:13.512668 131032678485120 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.012 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 6 seconds) (per-host) I0416 00:04:13.512747 131032678485120 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 6.132758s (batch_requests_ready=0.092360s, total_serialization_initiated=6.035966s, others=0.004432s) I0416 00:04:13.512861 131032678485120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 6.133289s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=6.133274) I0416 00:04:13.514639 130925048366656 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:04:13.514735 131032678485120 async_checkpointer.py:561] Finished blocking save. Time taken: 6.215254s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/9. I0416 00:04:13.514942 131032678485120 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:04:13.515116 130925098698304 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:04:13.515223 131032678485120 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776297815.6654913, 'wait_for_prev_duration_secs': 31.633721828460693, 'time_between_consecutive_saves_sec': 13.659812688827515, 'checkpointer_blocking_start_time': 1776297847.2994547, 'checkpointer_blocking_duration_secs': 6.215398550033569, 'get_old_steps_start_time': 1776297853.5148766, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1776297815.6654606, 'checkpoint_manager_blocking_duration_secs': 37.84973335266113} I0416 00:04:13.515375 131032678485120 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0416 00:04:13.515409 131032678485120 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:04:14.271243 130925172098624 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_true/checkpoints/9/items/array_metadatas/process_0 I0416 00:04:49.513494 130925109184064 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.999835s (commit=35.533535s, array_metadata_write=0.466300s) I0416 00:04:49.514567 130925048366656 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 299.915 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 42 seconds) (per-host) I0416 00:04:49.514665 130925048366656 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.999887s. I0416 00:04:49.935004 130925048366656 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 00:04:50.369637 130925048366656 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 00:04:50.370272 130925048366656 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.579586s. 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_true/checkpoints/9/items I0416 00:04:50.370929 130925048366656 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/9/items I0416 00:04:50.611225 130925048366656 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/9 I0416 00:04:51.293021 130925048366656 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_true/checkpoints/9`. I0416 00:04:51.293647 130925048366656 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 43.994174s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_15_ocdbt_true/checkpoints/9 I0416 00:04:51.293706 130925048366656 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 37.778930s. I0416 00:04:51.293889 130925098698304 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:04:51.293994 130925098698304 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:04:51.294057 130925098698304 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:04:51.294098 130925098698304 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 00:04:51.294609 131032678485120 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0416 00:04:51.294740 131032678485120 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:04:51.295631 131032678485120 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.974, Tokens/s/device: 17330.084, total_weights: 16384, loss: 4.189 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops