XPK Start: Thu Apr 23 18:52:49 UTC 2026 2026-04-23 18:52:53.964469: 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 00:00:1776970373.978037 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776970373.981808 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776970373.993647 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970373.993664 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970373.993667 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970373.993669 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 18:53:13.337234: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0423 18:53:13.870834 138873215493952 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 18:53:22,912:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 18:53:22.912644 138873215493952 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 18:53:22,915:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-7bvbb-slice-job-0-0.mt-16-zarr3-true-rs-7bvbb:8482 I0423 18:53:22.915012 138873215493952 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-7bvbb-slice-job-0-0.mt-16-zarr3-true-rs-7bvbb:8482 I0423 18:53:23.758831 138873215493952 max_utils.py:284] Jax distributed system initialized! I0423 18:53:30.672415 138873215493952 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 18:53:30.672527 138873215493952 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 18:53:30.672568 138873215493952 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0423 18:53:30.672603 138873215493952 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 18:53:31.290961 138873215493952 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 18:53:31.406205 138873215493952 checkpointing.py:688] Setting up checkpoint logger... I0423 18:53:31.406325 138873215493952 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 18:53:31.406370 138873215493952 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 18:53:31.406590 138873215493952 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 0x7e4d79571370>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 18:53:32.688442 138873215493952 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 18:53:32.688766 138873215493952 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e38e4591c10>}, handler_registry=None I0423 18:53:32.689003 138873215493952 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e38e4591c10>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0423 18:53:32.689051 138873215493952 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e391445f500>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0423 18:53:32.689085 138873215493952 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 0x7e38e4591c10>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e38e4591c10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e391445f500>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e391445f500>}). I0423 18:53:32.689450 138873215493952 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 18:53:32.689529 138873215493952 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e38e4554f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 18:53:34.628475 138752961779456 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776970198098438179, 'commit_timestamp_nsecs': 1776970244500454433, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/5/_CHECKPOINT_METADATA I0423 18:53:34.885563 138752408155904 checkpoint_manager.py:1877] Missing metrics for step 9 I0423 18:53:35.045528 138873215493952 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints I0423 18:53:35.286643 138873215493952 checkpoint_manager.py:929] [process=6][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/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e38e45d9070> I0423 18:53:35.286820 138873215493952 checkpointing.py:302] Checkpoint manager created! I0423 18:53:36.227801 138873215493952 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 18:53:36.227939 138873215493952 checkpointing.py:582] restoring from this run's directory step 9 I0423 18:53:37.601483 138873215493952 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/9. I0423 18:53:39.009689 1420 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 18:53:46.029707 138873215493952 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 224.924 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.025821924209595 s) (per-host) I0423 18:53:46.038222 138873215493952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 212.929 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.432562589645386 s) (per-host) I0423 18:53:46.170657 138873215493952 checkpointer.py:318] Finished restoring checkpoint in 8.81 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/9. I0423 18:53:46.170820 138873215493952 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776970417.3585389, 'checkpointer_duration_secs': 8.812249183654785, 'checkpoint_manager_start_time': 1776970416.2285953, 'checkpoint_manager_duration_secs': 9.942195177078247} fsdp: 32 I0423 18:53:46.252800 138873215493952 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0423 18:53:46.252918 138873215493952 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:53:46.252970 138873215493952 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:53:46.253012 138873215493952 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0423 18:53:46.253048 138873215493952 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:53:46.253082 138873215493952 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:53:46.253152 138873215493952 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:53:46.253198 138873215493952 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0423 18:53:46.253237 138873215493952 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:53:46.253276 138873215493952 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:53:46.253313 138873215493952 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0423 18:53:46.253347 138873215493952 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0423 18:53:46.651353 138873215493952 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:53:46.651501 138873215493952 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:53:46.657134 138873215493952 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:53:46.657231 138873215493952 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:53:46.814114 138873215493952 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:53:46.814249 138873215493952 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:53:46.971494 138873215493952 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:53:46.971639 138873215493952 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:53:47.137671 138873215493952 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:53:47.137813 138873215493952 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:53:47.293957 138873215493952 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:53:47.294110 138873215493952 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:53:47.448199 138873215493952 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:53:47.448334 138873215493952 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:53:47.610940 138873215493952 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 18:53:47.611076 138873215493952 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:53:47.633982 138873215493952 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 18:53:47.634079 138873215493952 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 18:53:58.999735 138873215493952 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. I0423 18:53:59.002152 138873215493952 metric_logger.py:289] number parameters: 1.104 billion I0423 18:54:10.247749 138873215493952 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0423 18:54:10.399941 138873215493952 checkpointing.py:798] Waited 0.15218186378479004 seconds for step 10 to finish before starting checkpointing. I0423 18:54:10.402342 138873215493952 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:54:10.404348 138873215493952 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0423 18:54:10.407023 138873215493952 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/10. I0423 18:54:10.987884 138873215493952 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 18:54:10.989029 138873215493952 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:54:10.989105 138873215493952 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 I0423 18:54:11.235541 138873215493952 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.248796s I0423 18:54:11.235716 138873215493952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.011 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5123593807220459 s) (per-host) I0423 18:54:11.235767 138873215493952 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.512420s (batch_requests_ready=0.258878s, total_serialization_initiated=0.253473s, others=0.000069s) I0423 18:54:11.235860 138873215493952 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.516480s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.516460) I0423 18:54:11.236731 138751317628672 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:54:11.236900 138873215493952 async_checkpointer.py:561] Finished blocking save. Time taken: 0.832484s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/10. I0423 18:54:11.278047 138873215493952 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:54:11.278378 138747576317696 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:54:11.278523 138873215493952 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970450.4023232, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776970450.4043858, 'checkpointer_blocking_duration_secs': 0.8326225280761719, 'get_old_steps_start_time': 1776970451.2370274, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776970450.400335, 'checkpoint_manager_blocking_duration_secs': 0.8781523704528809} I0423 18:54:11.278634 138873215493952 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0423 18:54:11.278709 138873215493952 max_utils.py:750] Memstats: After params initialized: I0423 18:54:11.278763 138873215493952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0423 18:54:11.278798 138873215493952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0423 18:54:11.278825 138873215493952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0423 18:54:11.278850 138873215493952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0423 18:54:11.553796 138873215493952 metric_logger.py:185] completed step: 10, seconds: 11.245, TFLOP/s/device: 1.208, Tokens/s/device: 182.132, total_weights: 65536, loss: 6.827 I0423 18:54:11.721794 138873215493952 metric_logger.py:185] completed step: 11, seconds: 1.304, TFLOP/s/device: 10.418, Tokens/s/device: 1570.273, total_weights: 65536, loss: 6.522 I0423 18:54:11.869025 138873215493952 metric_logger.py:185] completed step: 12, seconds: 0.013, TFLOP/s/device: 1068.591, Tokens/s/device: 161069.603, total_weights: 65536, loss: 6.260 I0423 18:54:12.017127 138873215493952 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.268, Tokens/s/device: 12551.019, total_weights: 65536, loss: 6.039 I0423 18:54:12.024534 138873215493952 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0423 18:54:12.311964 138873215493952 checkpointing.py:798] Waited 0.2874019145965576 seconds for step 15 to finish before starting checkpointing. I0423 18:54:12.314848 138873215493952 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:54:16.774783 138748104795904 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0423 18:54:53.941748 138751317628672 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.549 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.218353033065796 s) (per-host) I0423 18:54:53.941864 138751317628672 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.705023s. I0423 18:55:00.969748 138751317628672 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.732892s. I0423 18:55:00.969999 138747576317696 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:55:00.970062 138747576317696 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:55:00.970158 138747576317696 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:55:00.971741 138747576317696 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0423 18:55:00.971929 138873215493952 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0423 18:55:00.972066 138873215493952 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.657227 seconds. If this number is high, consider checkpointing less frequently. I0423 18:55:00.974223 138873215493952 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0423 18:55:00.978456 138873215493952 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/15. I0423 18:55:01.696556 138873215493952 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:55:01.696660 138873215493952 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 I0423 18:55:01.740992 138873215493952 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046433s I0423 18:55:01.741177 138873215493952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.034 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30640459060668945 s) (per-host) I0423 18:55:01.741233 138873215493952 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306471s (batch_requests_ready=0.255399s, total_serialization_initiated=0.050997s, others=0.000075s) I0423 18:55:01.741333 138873215493952 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311429s (all_items=0.000014s, per_item={'items': '0.00001431'}, temp_paths=0.311415) I0423 18:55:01.742176 138748608096000 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:55:01.742354 138873215493952 async_checkpointer.py:561] Finished blocking save. Time taken: 0.768045s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/15. I0423 18:55:01.752497 138873215493952 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:55:01.752812 138747576317696 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:55:01.752971 138873215493952 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970452.3148162, 'wait_for_prev_duration_secs': 48.6572265625, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776970500.9742699, 'checkpointer_blocking_duration_secs': 0.7681946754455566, 'get_old_steps_start_time': 1776970501.742485, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776970452.3123708, 'checkpoint_manager_blocking_duration_secs': 49.44056701660156} I0423 18:55:01.753084 138873215493952 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0423 18:55:01.753825 138873215493952 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.283, Tokens/s/device: 13909.831, total_weights: 65536, loss: 5.858 I0423 18:55:01.761379 138873215493952 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.133, Tokens/s/device: 13736.485, total_weights: 65536, loss: 5.713 I0423 18:55:19.791441 138873215493952 metric_logger.py:185] completed step: 16, seconds: 49.735, TFLOP/s/device: 0.273, Tokens/s/device: 41.178, total_weights: 65536, loss: 5.599 I0423 18:55:19.938690 138873215493952 metric_logger.py:185] completed step: 17, seconds: 17.885, TFLOP/s/device: 0.760, Tokens/s/device: 114.510, total_weights: 65536, loss: 5.511 I0423 18:55:20.086250 138873215493952 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.562, Tokens/s/device: 13348.977, total_weights: 65536, loss: 5.442 I0423 18:55:20.234861 138873215493952 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0423 18:55:20.235595 138873215493952 checkpointing.py:798] Waited 0.0007421970367431641 seconds for step 19 to finish before starting checkpointing. I0423 18:55:20.237582 138873215493952 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:55:24.646379 138748104795904 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0423 18:56:00.962110 138748608096000 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.536 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.52728796005249 s) (per-host) I0423 18:56:00.962212 138748608096000 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.219914s. I0423 18:56:08.830289 138748608096000 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 67.087975s. I0423 18:56:08.830596 138747576317696 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:56:08.830734 138747576317696 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:56:08.830784 138747576317696 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:56:08.832931 138747576317696 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0423 18:56:08.833124 138873215493952 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0423 18:56:08.833257 138873215493952 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.595702 seconds. If this number is high, consider checkpointing less frequently. I0423 18:56:08.834920 138873215493952 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0423 18:56:08.838304 138873215493952 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/19. I0423 18:56:09.840234 138873215493952 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:56:09.840333 138873215493952 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 I0423 18:56:09.881166 138873215493952 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043071s I0423 18:56:09.881328 138873215493952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.079 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.303729772567749 s) (per-host) I0423 18:56:09.881376 138873215493952 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.303786s (batch_requests_ready=0.256130s, total_serialization_initiated=0.047595s, others=0.000062s) I0423 18:56:09.881468 138873215493952 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.307753s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.307737) I0423 18:56:09.882295 138751317628672 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:56:09.882424 138873215493952 async_checkpointer.py:561] Finished blocking save. Time taken: 1.047438s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/19. I0423 18:56:09.896996 138873215493952 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:56:09.897304 138747576317696 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:56:09.897473 138873215493952 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970520.2375402, 'wait_for_prev_duration_secs': 48.595701932907104, 'time_between_consecutive_saves_sec': 19.26576018333435, 'checkpointer_blocking_start_time': 1776970568.8349612, 'checkpointer_blocking_duration_secs': 1.0476329326629639, 'get_old_steps_start_time': 1776970569.8826072, 'get_old_steps_duration_secs': 2.0503997802734375e-05, 'checkpoint_manager_blocking_start_time': 1776970520.2358382, 'checkpoint_manager_blocking_duration_secs': 49.66160464286804} I0423 18:56:09.897584 138873215493952 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0423 18:56:09.897626 138873215493952 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:56:14.928569 138752357799680 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0423 18:56:52.392429 138751317628672 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.894 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.814793825149536 s) (per-host) I0423 18:56:52.392534 138751317628672 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.510032s. I0423 18:57:06.003364 138751317628672 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 56.120847s. I0423 18:57:06.003632 138747576317696 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:57:06.003756 138747576317696 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:57:06.003798 138747576317696 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:57:06.005124 138747576317696 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0423 18:57:06.005309 138873215493952 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0423 18:57:06.005501 138873215493952 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:57:06.006159 138873215493952 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.518, Tokens/s/device: 13945.349, total_weights: 65536, loss: 5.385 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 18:57:15 UTC 2026 EXIT_CODE=0