XPK Start: Mon Apr 20 20:21:42 UTC 2026 2026-04-20 20:21:46.523087: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1776716506.535655 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776716506.539338 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776716506.550517 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776716506.550538 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776716506.550540 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776716506.550542 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 20:22:05.733627: 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) I0420 20:22:06.257936 133508389783360 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 20:22:15,298:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 20:22:15.298569 133508389783360 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 20:22:15,300:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-0nsj7-slice-job-0-0.mt-15-ocdbt-true-rs-0nsj7:8482 I0420 20:22:15.300915 133508389783360 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-0nsj7-slice-job-0-0.mt-15-ocdbt-true-rs-0nsj7:8482 I0420 20:22:16.279574 133508389783360 max_utils.py:284] Jax distributed system initialized! I0420 20:22:23.161180 133508389783360 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 20:22:23.161286 133508389783360 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 20:22:23.161326 133508389783360 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0420 20:22:23.161359 133508389783360 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 20:22:23.764917 133508389783360 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:22:23.765520 133508389783360 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:22:23.765702 133508389783360 checkpointing.py:688] Setting up checkpoint logger... I0420 20:22:23.765753 133508389783360 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 20:22:23.765809 133508389783360 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 20:22:23.766171 133508389783360 base_pytree_checkpoint_handler.py:415] 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 0x796c58d75370>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 20:22:25.116316 133508389783360 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 20:22:25.116653 133508389783360 checkpoint_manager.py:709] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79594939f1d0>}, handler_registry=None I0420 20:22:25.116906 133508389783360 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79594939f1d0>` 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`. I0420 20:22:25.116956 133508389783360 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7959493a6c00>` 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`. I0420 20:22:25.116993 133508389783360 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 0x79594939f1d0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79594939f1d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7959493a6c00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7959493a6c00>}). I0420 20:22:25.117506 133508389783360 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 20:22:25.117585 133508389783360 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79587b51c180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 20:22:26.753283 133386022668032 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776716377866898877, 'commit_timestamp_nsecs': 1776716425034017261, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9/_CHECKPOINT_METADATA I0420 20:22:27.003223 133386014275328 checkpoint_manager.py:1877] Missing metrics for step 0 I0420 20:22:27.036802 133508389783360 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints I0420 20:22:27.717666 133508389783360 checkpoint_manager.py:929] [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_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, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79594a35f020> I0420 20:22:27.717881 133508389783360 checkpointing.py:302] Checkpoint manager created! I0420 20:22:28.597167 133508389783360 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 20:22:28.597273 133508389783360 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:22:28.969982 133508389783360 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:22:28.970075 133508389783360 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:22:28.985769 133508389783360 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:22:28.985840 133508389783360 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:22:29.008433 133508389783360 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:22:29.008497 133508389783360 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:22:29.024258 133508389783360 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:22:29.024317 133508389783360 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:22:29.039920 133508389783360 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:22:29.039980 133508389783360 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:22:29.064581 133508389783360 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 20:22:29.064656 133508389783360 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:22:29.084189 133508389783360 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 20:22:29.084256 133508389783360 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 20:22:29.272218 133508389783360 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 20:22:29.272328 133508389783360 checkpointing.py:582] restoring from this run's directory step 9 I0420 20:22:29.763758 133508389783360 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9. I0420 20:22:31.337494 1436 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 20:22:38.611063 133508389783360 jax_array_handlers.py:837] [process=0] /jax/orbax/read/worker/io/requested throughput: 217.107 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.278793811798096 s) (per-host) I0420 20:22:38.616337 133508389783360 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/read/gbytes_per_sec: 202.465 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.816633939743042 s) (per-host) I0420 20:22:42.994773 133508389783360 checkpointer.py:318] Finished restoring checkpoint in 13.47 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9. I0420 20:22:42.994974 133508389783360 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776716549.5204933, 'checkpointer_duration_secs': 13.474448919296265, 'checkpoint_manager_start_time': 1776716549.2725794, 'checkpoint_manager_duration_secs': 13.722364902496338} fsdp: 32 I0420 20:22:43.395154 133508389783360 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0420 20:22:43.395285 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0420 20:22:43.395342 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0420 20:22:43.395405 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0420 20:22:43.395458 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0420 20:22:43.395499 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0420 20:22:43.395552 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0420 20:22:43.395605 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0420 20:22:43.395646 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0420 20:22:43.395684 133508389783360 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0420 20:22:43.395730 133508389783360 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0420 20:22:43.395776 133508389783360 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0420 20:22:43.888897 133508389783360 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:22:43.888987 133508389783360 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:22:43.903824 133508389783360 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:22:43.903884 133508389783360 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:22:54.655812 133508389783360 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0420 20:22:54.662019 133508389783360 metric_logger.py:289] number parameters: 1.104 billion I0420 20:23:06.328498 133508389783360 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 20:23:06.482215 133508389783360 checkpointing.py:798] Waited 0.15369772911071777 seconds for step 10 to finish before starting checkpointing. I0420 20:23:06.484817 133508389783360 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:23:06.486650 133508389783360 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 10 I0420 20:23:06.489387 133508389783360 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10. I0420 20:23:07.220183 133508389783360 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 20:23:07.465051 133381719308032 atomicity.py:137] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10 I0420 20:23:07.500304 133508389783360 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:23:07.500402 133508389783360 replica_slices.py:424] 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 I0420 20:23:07.738403 133508389783360 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.239072s I0420 20:23:07.739176 133508389783360 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.017 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5112168788909912 s) (per-host) I0420 20:23:07.739256 133508389783360 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.511319s (batch_requests_ready=0.268946s, total_serialization_initiated=0.241672s, others=0.000700s) I0420 20:23:07.739340 133508389783360 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.515350s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.515330) I0420 20:23:07.740217 133382818227968 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 20:23:07.740376 133508389783360 async_checkpointer.py:561] Finished blocking save. Time taken: 1.253654s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10. I0420 20:23:08.195957 133508389783360 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:23:08.196346 133381710915328 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:23:08.196510 133508389783360 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776716586.484798, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776716586.4866898, 'checkpointer_blocking_duration_secs': 1.2537930011749268, 'get_old_steps_start_time': 1776716587.740502, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776716586.4828563, 'checkpoint_manager_blocking_duration_secs': 1.713615894317627} I0420 20:23:08.196615 133508389783360 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 20:23:08.196668 133508389783360 max_utils.py:750] Memstats: After params initialized: I0420 20:23:08.196727 133508389783360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 20:23:08.196762 133508389783360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 20:23:08.196806 133508389783360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 20:23:08.196835 133508389783360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 20:23:08.486167 133508389783360 metric_logger.py:185] completed step: 10, seconds: 11.298, TFLOP/s/device: 1.203, Tokens/s/device: 181.279, total_weights: 65536, loss: 6.833 I0420 20:23:08.631624 133508389783360 metric_logger.py:185] completed step: 11, seconds: 2.156, TFLOP/s/device: 6.302, Tokens/s/device: 949.874, total_weights: 65536, loss: 6.528 I0420 20:23:08.778856 133508389783360 metric_logger.py:185] completed step: 12, seconds: 0.010, TFLOP/s/device: 1409.892, Tokens/s/device: 212514.268, total_weights: 65536, loss: 6.265 I0420 20:23:08.926430 133508389783360 metric_logger.py:185] completed step: 13, seconds: 0.146, TFLOP/s/device: 93.188, Tokens/s/device: 14046.254, total_weights: 65536, loss: 6.045 I0420 20:23:08.934882 133508389783360 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0420 20:23:09.205189 133384923748096 atomicity.py:137] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/items I0420 20:23:09.221282 133508389783360 checkpointing.py:798] Waited 0.2862675189971924 seconds for step 15 to finish before starting checkpointing. I0420 20:23:09.224113 133508389783360 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:23:09.804756 133382793049856 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776716588958911301, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776716588958911301, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA I0420 20:23:12.200445 133383313135360 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_0 I0420 20:23:52.413831 133382826620672 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.083282s (commit=40.938421s, array_metadata_write=1.144861s) I0420 20:23:52.415062 133382818227968 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 34.957 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.18708896636963 s) (per-host) I0420 20:23:52.415156 133382818227968 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.674824s. I0420 20:23:53.948111 133382818227968 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000409s. I0420 20:23:53.976742 133382793049856 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776716588958911301, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA I0420 20:23:54.595203 133382818227968 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 20:23:55.587863 133382818227968 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.251893s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/items I0420 20:23:55.590250 133382818227968 atomicity.py:608] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/items I0420 20:23:56.015188 133382818227968 atomicity.py:608] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10 I0420 20:23:58.088278 133382818227968 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10`. I0420 20:23:58.089406 133382818227968 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 51.602684s. directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10 I0420 20:23:58.090742 133382818227968 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 50.350407s. I0420 20:23:58.091093 133381710915328 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:23:58.091270 133381710915328 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:23:58.091421 133381710915328 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:23:58.093065 133381710915328 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 20:23:58.093402 133508389783360 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0420 20:23:58.093575 133508389783360 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.869469 seconds. If this number is high, consider checkpointing less frequently. I0420 20:23:58.095455 133508389783360 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 15 I0420 20:23:58.099474 133508389783360 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15. I0420 20:23:58.652531 133382818227968 atomicity.py:137] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15 I0420 20:23:58.674369 133508389783360 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:23:58.674511 133508389783360 replica_slices.py:424] 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 I0420 20:23:58.721856 133508389783360 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048700s I0420 20:23:58.722713 133508389783360 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.914 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31392598152160645 s) (per-host) I0420 20:23:58.722833 133508389783360 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314070s (batch_requests_ready=0.260495s, total_serialization_initiated=0.052807s, others=0.000768s) I0420 20:23:58.722961 133508389783360 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318666s (all_items=0.000023s, per_item={'items': '0.00002289'}, temp_paths=0.318643) I0420 20:23:58.724141 133381710915328 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 20:23:58.724342 133508389783360 async_checkpointer.py:561] Finished blocking save. Time taken: 0.628783s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15. I0420 20:23:58.730282 133508389783360 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:23:58.730676 133382784657152 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:23:58.730925 133508389783360 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776716589.2240713, 'wait_for_prev_duration_secs': 48.86946892738342, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776716638.0955203, 'checkpointer_blocking_duration_secs': 0.6289761066436768, 'get_old_steps_start_time': 1776716638.7245219, 'get_old_steps_duration_secs': 3.910064697265625e-05, 'checkpoint_manager_blocking_start_time': 1776716589.2219257, 'checkpoint_manager_blocking_duration_secs': 49.50894618034363} I0420 20:23:58.731075 133508389783360 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0420 20:24:16.323321 133508389783360 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.305, Tokens/s/device: 13913.233, total_weights: 65536, loss: 5.864 I0420 20:24:16.332160 133508389783360 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.089, Tokens/s/device: 13880.605, total_weights: 65536, loss: 5.719 I0420 20:24:16.477901 133508389783360 metric_logger.py:185] completed step: 16, seconds: 67.396, TFLOP/s/device: 0.202, Tokens/s/device: 30.388, total_weights: 65536, loss: 5.605 I0420 20:24:16.625320 133508389783360 metric_logger.py:185] completed step: 17, seconds: 0.007, TFLOP/s/device: 1961.191, Tokens/s/device: 295612.009, total_weights: 65536, loss: 5.517 I0420 20:24:16.772818 133508389783360 metric_logger.py:185] completed step: 18, seconds: 0.147, TFLOP/s/device: 92.689, Tokens/s/device: 13971.035, total_weights: 65536, loss: 5.448 I0420 20:24:16.919950 133508389783360 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 20:24:16.920561 133508389783360 checkpointing.py:798] Waited 0.0006310939788818359 seconds for step 19 to finish before starting checkpointing. I0420 20:24:16.922661 133508389783360 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:24:17.855433 133382231877376 atomicity.py:137] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15/items I0420 20:24:20.683689 133383313135360 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_0 I0420 20:25:00.185248 133382826620672 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.174057s (commit=40.094720s, array_metadata_write=1.079337s) I0420 20:25:00.186376 133381710915328 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 25.569 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.777570962905884 s) (per-host) I0420 20:25:00.186493 133381710915328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.462208s. I0420 20:25:01.724564 133381710915328 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000245s. I0420 20:25:02.047669 133381710915328 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 20:25:02.707438 133381710915328 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.586613s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15/items I0420 20:25:02.709378 133381710915328 atomicity.py:608] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15/items I0420 20:25:03.154690 133381710915328 atomicity.py:608] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15 I0420 20:25:04.640132 133381710915328 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15`. I0420 20:25:04.640996 133381710915328 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 66.545449s. directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15 I0420 20:25:04.642396 133381710915328 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 65.918110s. I0420 20:25:04.642562 133382784657152 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:25:04.642654 133382784657152 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:25:04.642730 133382784657152 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:25:04.644688 133382784657152 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0420 20:25:04.644871 133508389783360 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0420 20:25:04.644997 133508389783360 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.722352 seconds. If this number is high, consider checkpointing less frequently. I0420 20:25:04.646477 133508389783360 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 19 I0420 20:25:04.649988 133508389783360 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19. I0420 20:25:05.194516 133381710915328 atomicity.py:137] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19 I0420 20:25:05.195239 133508389783360 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:25:05.195339 133508389783360 replica_slices.py:424] 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 I0420 20:25:05.221377 133508389783360 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027771s I0420 20:25:05.222123 133508389783360 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.461 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28249382972717285 s) (per-host) I0420 20:25:05.222186 133508389783360 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282573s (batch_requests_ready=0.251957s, total_serialization_initiated=0.029927s, others=0.000689s) I0420 20:25:05.222259 133508389783360 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286396s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.286379) I0420 20:25:05.223066 133382826620672 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 20:25:05.223228 133508389783360 async_checkpointer.py:561] Finished blocking save. Time taken: 0.576681s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19. I0420 20:25:05.269384 133508389783360 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:25:05.269697 133382784657152 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:25:05.269858 133508389783360 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776716656.9226284, 'wait_for_prev_duration_secs': 47.72235155105591, 'time_between_consecutive_saves_sec': 18.829495668411255, 'checkpointer_blocking_start_time': 1776716704.646518, 'checkpointer_blocking_duration_secs': 0.5768218040466309, 'get_old_steps_start_time': 1776716705.2233567, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776716656.9208033, 'checkpoint_manager_blocking_duration_secs': 48.349021673202515} I0420 20:25:05.270008 133508389783360 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 20:25:05.270052 133508389783360 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:25:07.028836 133383313135360 atomicity.py:137] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19/items I0420 20:25:10.247505 133385485797120 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_0 I0420 20:25:49.457031 133383321528064 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.306380s (commit=40.195678s, array_metadata_write=1.110701s) I0420 20:25:49.458189 133382826620672 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 35.482 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.51853156089783 s) (per-host) I0420 20:25:49.458302 133382826620672 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.235114s. I0420 20:25:51.246321 133382826620672 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000218s. I0420 20:25:51.327293 133382826620672 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 20:25:52.074548 133382826620672 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.679968s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19/items I0420 20:25:52.077180 133382826620672 atomicity.py:608] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19/items I0420 20:25:52.529156 133382826620672 atomicity.py:608] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19 I0420 20:25:53.871891 133382826620672 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19`. I0420 20:25:53.872691 133382826620672 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 49.226148s. directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19 I0420 20:25:53.875061 133382826620672 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 48.651871s. I0420 20:25:53.875266 133382784657152 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:25:53.875377 133382784657152 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:25:53.875432 133382784657152 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:25:53.877372 133382784657152 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 20:25:53.877571 133508389783360 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0420 20:25:53.877657 133508389783360 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:25:53.878287 133508389783360 metric_logger.py:185] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.863, Tokens/s/device: 13846.539, total_weights: 65536, loss: 5.391 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 20:26:07 UTC 2026 EXIT_CODE=0