XPK Start: Fri Apr 24 17:44:52 UTC 2026 2026-04-24 17:44:56.425760: 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 00:00:1777052696.438927 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777052696.442975 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777052696.454465 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052696.454487 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052696.454489 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052696.454491 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 17:45:15.882573: 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) I0424 17:45:16.416310 133898153457472 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 17:45:25,457:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 17:45:25.457285 133898153457472 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 17:45:25,459:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-72l5u-slice-job-0-0.mt-16-zarr3-true-rs-72l5u:8482 I0424 17:45:25.459824 133898153457472 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-72l5u-slice-job-0-0.mt-16-zarr3-true-rs-72l5u:8482 I0424 17:45:26.297911 133898153457472 max_utils.py:284] Jax distributed system initialized! I0424 17:45:33.264818 133898153457472 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 17:45:33.264925 133898153457472 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 17:45:33.264967 133898153457472 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 I0424 17:45:33.265003 133898153457472 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 17:45:33.889515 133898153457472 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 17:45:34.007545 133898153457472 checkpointing.py:688] Setting up checkpoint logger... I0424 17:45:34.007661 133898153457472 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 17:45:34.007706 133898153457472 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 17:45:34.007914 133898153457472 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 0x79c7047793a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 17:45:35.274791 133898153457472 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 17:45:35.275119 133898153457472 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 0x79b1f058ca10>}, handler_registry=None I0424 17:45:35.275356 133898153457472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79b1f058ca10>` 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`. I0424 17:45:35.275405 133898153457472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79b2a81b7fe0>` 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`. I0424 17:45:35.275441 133898153457472 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 0x79b1f058ca10>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79b1f058ca10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79b2a81b7fe0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79b2a81b7fe0>}). I0424 17:45:35.275771 133898153457472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 17:45:35.275849 133898153457472 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79b280278f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 17:45:37.075451 133777418348288 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777052570566555037, 'commit_timestamp_nsecs': 1777052615510237889, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9/_CHECKPOINT_METADATA I0424 17:45:37.322682 133777418348288 checkpoint_manager.py:1877] Missing metrics for step 9 I0424 17:45:37.619195 133898153457472 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints I0424 17:45:37.755425 133898153457472 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79b2a81b7350> I0424 17:45:37.755598 133898153457472 checkpointing.py:302] Checkpoint manager created! I0424 17:45:38.693748 133898153457472 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 17:45:38.693871 133898153457472 checkpointing.py:582] restoring from this run's directory step 9 I0424 17:45:39.193935 133898153457472 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9. I0424 17:45:40.613957 1419 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 17:45:46.621218 133898153457472 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 262.815 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.012875556945801 s) (per-host) I0424 17:45:46.639112 133898153457472 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 245.629 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.4430811405181885 s) (per-host) I0424 17:45:48.231185 133898153457472 checkpointer.py:318] Finished restoring checkpoint in 9.28 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9. I0424 17:45:48.231348 133898153457472 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1777052738.9492033, 'checkpointer_duration_secs': 9.282111883163452, 'checkpoint_manager_start_time': 1777052738.6945918, 'checkpoint_manager_duration_secs': 9.536725282669067} fsdp: 32 I0424 17:45:48.315561 133898153457472 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 17:45:48.315679 133898153457472 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:45:48.315731 133898153457472 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:45:48.315775 133898153457472 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 17:45:48.315812 133898153457472 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:45:48.315848 133898153457472 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:45:48.315886 133898153457472 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:45:48.315924 133898153457472 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 17:45:48.315960 133898153457472 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:45:48.315995 133898153457472 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:45:48.316027 133898153457472 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 17:45:48.316060 133898153457472 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 17:45:48.719048 133898153457472 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:45:48.719212 133898153457472 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:45:48.724650 133898153457472 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:45:48.724746 133898153457472 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:45:48.884311 133898153457472 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:45:48.884447 133898153457472 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:45:49.041381 133898153457472 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:45:49.041516 133898153457472 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:45:49.205811 133898153457472 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:45:49.205947 133898153457472 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:45:49.360403 133898153457472 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:45:49.360533 133898153457472 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:45:49.514713 133898153457472 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:45:49.514844 133898153457472 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:45:49.681352 133898153457472 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 17:45:49.681482 133898153457472 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:45:49.704430 133898153457472 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 17:45:49.704528 133898153457472 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 17:46:00.915032 133898153457472 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. I0424 17:46:00.917519 133898153457472 metric_logger.py:289] number parameters: 1.104 billion I0424 17:46:11.949489 133898153457472 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0424 17:46:12.101153 133898153457472 checkpointing.py:798] Waited 0.15160322189331055 seconds for step 10 to finish before starting checkpointing. I0424 17:46:12.103642 133898153457472 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:46:12.105577 133898153457472 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0424 17:46:12.109365 133898153457472 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/10. I0424 17:46:12.694407 133898153457472 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 17:46:12.695551 133898153457472 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:46:12.695609 133898153457472 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 I0424 17:46:12.949967 133898153457472 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256752s I0424 17:46:12.950152 133898153457472 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.916 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5289208889007568 s) (per-host) I0424 17:46:12.950209 133898153457472 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.528989s (batch_requests_ready=0.267252s, total_serialization_initiated=0.261659s, others=0.000078s) I0424 17:46:12.950310 133898153457472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.533026s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.533008) I0424 17:46:12.951192 133777376384768 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:46:12.951321 133898153457472 async_checkpointer.py:561] Finished blocking save. Time taken: 0.845673s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/10. I0424 17:46:12.967390 133898153457472 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:46:12.967713 133773534426880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:46:12.967873 133898153457472 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052772.1036234, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052772.105616, 'checkpointer_blocking_duration_secs': 0.8458802700042725, 'get_old_steps_start_time': 1777052772.9515195, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1777052772.101541, 'checkpoint_manager_blocking_duration_secs': 0.8662941455841064} I0424 17:46:12.967983 133898153457472 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0424 17:46:12.968057 133898153457472 max_utils.py:750] Memstats: After params initialized: I0424 17:46:12.968120 133898153457472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 17:46:12.968156 133898153457472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 17:46:12.968183 133898153457472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 17:46:12.968207 133898153457472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 17:46:13.244133 133898153457472 metric_logger.py:185] completed step: 10, seconds: 11.031, TFLOP/s/device: 1.232, Tokens/s/device: 185.661, total_weights: 65536, loss: 6.827 I0424 17:46:13.420152 133898153457472 metric_logger.py:185] completed step: 11, seconds: 1.293, TFLOP/s/device: 10.509, Tokens/s/device: 1583.969, total_weights: 65536, loss: 6.522 I0424 17:46:13.567524 133898153457472 metric_logger.py:185] completed step: 12, seconds: 0.020, TFLOP/s/device: 665.318, Tokens/s/device: 100284.007, total_weights: 65536, loss: 6.260 I0424 17:46:13.715043 133898153457472 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 83.057, Tokens/s/device: 12519.256, total_weights: 65536, loss: 6.039 I0424 17:46:13.722534 133898153457472 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0424 17:46:14.010113 133898153457472 checkpointing.py:798] Waited 0.28752923011779785 seconds for step 15 to finish before starting checkpointing. I0424 17:46:14.012724 133898153457472 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:46:17.434544 133773626681088 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0424 17:46:54.788291 133777376384768 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.284 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.367024421691895 s) (per-host) I0424 17:46:54.788416 133777376384768 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.837016s. I0424 17:47:01.762507 133777376384768 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.811092s. I0424 17:47:01.762793 133773534426880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:47:01.762917 133773534426880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:47:01.762966 133773534426880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:47:01.764626 133773534426880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0424 17:47:01.764829 133898153457472 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. W0424 17:47:01.764956 133898153457472 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.752235 seconds. If this number is high, consider checkpointing less frequently. I0424 17:47:01.766990 133898153457472 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0424 17:47:01.771240 133898153457472 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/15. I0424 17:47:02.321242 133898153457472 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:47:02.321352 133898153457472 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 I0424 17:47:02.370678 133898153457472 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051489s I0424 17:47:02.370845 133898153457472 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.897 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31501150131225586 s) (per-host) I0424 17:47:02.370898 133898153457472 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.315072s (batch_requests_ready=0.258832s, total_serialization_initiated=0.056172s, others=0.000069s) I0424 17:47:02.370994 133898153457472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.319865s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.319852) I0424 17:47:02.371850 133773526034176 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:47:02.372007 133898153457472 async_checkpointer.py:561] Finished blocking save. Time taken: 0.604942s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/15. I0424 17:47:02.390009 133898153457472 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:47:02.390301 133773534426880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:47:02.390445 133898153457472 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052774.0126967, 'wait_for_prev_duration_secs': 47.752235412597656, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052821.7670326, 'checkpointer_blocking_duration_secs': 0.6051013469696045, 'get_old_steps_start_time': 1777052822.3721592, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1777052774.0105126, 'checkpoint_manager_blocking_duration_secs': 48.37989830970764} I0424 17:47:02.390557 133898153457472 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0424 17:47:02.391339 133898153457472 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.443, Tokens/s/device: 13933.963, total_weights: 65536, loss: 5.858 I0424 17:47:02.399161 133898153457472 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.283, Tokens/s/device: 13759.095, total_weights: 65536, loss: 5.713 I0424 17:47:20.523144 133898153457472 metric_logger.py:185] completed step: 16, seconds: 48.675, TFLOP/s/device: 0.279, Tokens/s/device: 42.075, total_weights: 65536, loss: 5.599 I0424 17:47:20.670845 133898153457472 metric_logger.py:185] completed step: 17, seconds: 17.979, TFLOP/s/device: 0.756, Tokens/s/device: 113.910, total_weights: 65536, loss: 5.511 I0424 17:47:20.818493 133898153457472 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.632, Tokens/s/device: 13359.513, total_weights: 65536, loss: 5.442 I0424 17:47:20.966888 133898153457472 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0424 17:47:20.967638 133898153457472 checkpointing.py:798] Waited 0.0007603168487548828 seconds for step 19 to finish before starting checkpointing. I0424 17:47:20.969561 133898153457472 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:47:25.534435 133773626681088 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0424 17:48:02.791264 133773526034176 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.008 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.735393047332764 s) (per-host) I0424 17:48:02.791387 133773526034176 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.419426s. I0424 17:48:11.710057 133773526034176 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 69.338081s. I0424 17:48:11.710299 133773534426880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:48:11.710359 133773534426880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:48:11.710413 133773534426880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:48:11.711925 133773534426880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0424 17:48:11.712133 133898153457472 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. W0424 17:48:11.712265 133898153457472 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.742714 seconds. If this number is high, consider checkpointing less frequently. I0424 17:48:11.714966 133898153457472 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0424 17:48:11.718720 133898153457472 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/19. I0424 17:48:12.266056 133898153457472 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:48:12.266174 133898153457472 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 I0424 17:48:12.316149 133898153457472 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.052094s I0424 17:48:12.316316 133898153457472 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.933 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31269145011901855 s) (per-host) I0424 17:48:12.316366 133898153457472 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.312756s (batch_requests_ready=0.255957s, total_serialization_initiated=0.056730s, others=0.000069s) I0424 17:48:12.316457 133898153457472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.317113s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.317100) I0424 17:48:12.317296 133776302642944 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:48:12.317458 133898153457472 async_checkpointer.py:561] Finished blocking save. Time taken: 0.602425s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/19. I0424 17:48:12.569328 133898153457472 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:48:12.569710 133773534426880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:48:12.569857 133898153457472 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052840.9695354, 'wait_for_prev_duration_secs': 50.742713928222656, 'time_between_consecutive_saves_sec': 19.204869985580444, 'checkpointer_blocking_start_time': 1777052891.715006, 'checkpointer_blocking_duration_secs': 0.6025617122650146, 'get_old_steps_start_time': 1777052892.31758, 'get_old_steps_duration_secs': 1.9311904907226562e-05, 'checkpoint_manager_blocking_start_time': 1777052840.9678638, 'checkpoint_manager_blocking_duration_secs': 51.6019606590271} I0424 17:48:12.570027 133898153457472 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0424 17:48:12.570072 133898153457472 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:48:17.051251 133777376384768 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0424 17:48:53.765576 133776302642944 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.824 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.761932373046875 s) (per-host) I0424 17:48:53.765676 133776302642944 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.448263s. I0424 17:49:00.869794 133776302642944 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.552361s. I0424 17:49:00.870064 133773534426880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:49:00.870213 133773534426880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:49:00.870260 133773534426880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:49:00.872870 133773534426880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0424 17:49:00.873039 133898153457472 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. I0424 17:49:00.873230 133898153457472 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:49:00.873905 133898153457472 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.292, Tokens/s/device: 13911.248, 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: Fri Apr 24 17:49:13 UTC 2026 EXIT_CODE=0