XPK Start: Fri Apr 24 17:39:38 UTC 2026 2026-04-24 17:39:42.610237: 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 00:00:1777052382.623271 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777052382.627015 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777052382.638694 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052382.638716 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052382.638718 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052382.638720 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 17:40:01.885410: 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 17:40:02.418011 139766024124224 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 17:40:11,460:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 17:40:11.460395 139766024124224 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 17:40:11,462:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-sv-uhe0h-slice-job-0-0.mt-16-zarr3-true-sv-uhe0h:8482 I0424 17:40:11.462739 139766024124224 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-sv-uhe0h-slice-job-0-0.mt-16-zarr3-true-sv-uhe0h:8482 I0424 17:40:12.290686 139766024124224 max_utils.py:284] Jax distributed system initialized! I0424 17:40:19.285493 139766024124224 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 17:40:19.285605 139766024124224 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 17:40:19.285648 139766024124224 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 17:40:19.285683 139766024124224 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 17:40:19.909459 139766024124224 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 17:40:20.026234 139766024124224 checkpointing.py:688] Setting up checkpoint logger... I0424 17:40:20.026356 139766024124224 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 17:40:20.026402 139766024124224 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 17:40:20.026610 139766024124224 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 0x7f1d41d79580>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 17:40:22.900234 139766024124224 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 17:40:22.900477 139766024124224 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 0x7f08ac57ed50>}, handler_registry=None I0424 17:40:22.900722 139766024124224 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f08ac57ed50>` 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 17:40:22.900771 139766024124224 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f08ac3a4260>` 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 17:40:22.900807 139766024124224 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 0x7f08ac57ed50>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f08ac57ed50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f08ac3a4260>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f08ac3a4260>}). I0424 17:40:22.901187 139766024124224 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 17:40:22.901268 139766024124224 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f08ac31cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 17:40:23.698688 139766024124224 checkpoint_manager.py:1818] Found 0 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 17:40:23.714747 139766024124224 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 0x7f08ac57f140> I0424 17:40:23.714880 139766024124224 checkpointing.py:302] Checkpoint manager created! I0424 17:40:24.652879 139766024124224 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 17:40:24.652997 139766024124224 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 17:40:26.106272 139766024124224 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 17:40:26.106395 139766024124224 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:40:26.106450 139766024124224 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:40:26.106495 139766024124224 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 17:40:26.106538 139766024124224 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:40:26.106574 139766024124224 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:40:26.106610 139766024124224 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:40:26.106647 139766024124224 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 17:40:26.106682 139766024124224 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:40:26.106716 139766024124224 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:40:26.106749 139766024124224 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 17:40:26.106782 139766024124224 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 17:40:26.504271 139766024124224 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:40:26.504439 139766024124224 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:40:26.510054 139766024124224 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:40:26.510187 139766024124224 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:40:26.677001 139766024124224 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:40:26.677170 139766024124224 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:40:26.839529 139766024124224 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:40:26.839666 139766024124224 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:40:27.012270 139766024124224 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 17:40:27.012412 139766024124224 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:40:27.168494 139766024124224 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 17:40:27.168629 139766024124224 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:40:27.326226 139766024124224 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 17:40:27.326364 139766024124224 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:40:27.495915 139766024124224 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 17:40:27.496058 139766024124224 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:40:27.520434 139766024124224 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 17:40:27.520538 139766024124224 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 17:40:38.506423 139766024124224 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 17:40:38.509581 139766024124224 metric_logger.py:289] number parameters: 1.104 billion I0424 17:40:49.788338 139766024124224 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 17:40:49.944000 139766024124224 checkpointing.py:798] Waited 0.15564870834350586 seconds for step 0 to finish before starting checkpointing. I0424 17:40:49.946432 139766024124224 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:40:49.948183 139766024124224 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 17:40:49.951230 139766024124224 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/0. I0424 17:40:50.533042 139766024124224 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 17:40:50.534122 139766024124224 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:40:50.534183 139766024124224 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 17:40:50.784082 139766024124224 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252160s I0424 17:40:50.784277 139766024124224 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.800 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5509340763092041 s) (per-host) I0424 17:40:50.784330 139766024124224 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.550995s (batch_requests_ready=0.292212s, total_serialization_initiated=0.258714s, others=0.000070s) I0424 17:40:50.784424 139766024124224 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.554971s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.554952) I0424 17:40:50.785471 139640409396992 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:40:50.785646 139766024124224 async_checkpointer.py:561] Finished blocking save. Time taken: 0.837392s. 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/0. I0424 17:40:50.810375 139766024124224 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:40:50.810640 139638832355072 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:40:50.810789 139766024124224 standard_logger.py:34] {'step': 0, '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': 1777052449.9464128, 'wait_for_prev_duration_secs': 6.747245788574219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052449.9482207, 'checkpointer_blocking_duration_secs': 0.8375334739685059, 'get_old_steps_start_time': 1777052450.7857752, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1777052449.9444485, 'checkpoint_manager_blocking_duration_secs': 0.866290807723999} I0424 17:40:50.810902 139766024124224 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 17:40:50.810973 139766024124224 max_utils.py:750] Memstats: After params initialized: I0424 17:40:50.811022 139766024124224 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 17:40:50.811053 139766024124224 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 17:40:50.811079 139766024124224 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 17:40:50.811123 139766024124224 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 17:40:51.094396 139766024124224 metric_logger.py:185] completed step: 0, seconds: 11.278, TFLOP/s/device: 1.205, Tokens/s/device: 181.598, total_weights: 65536, loss: 10.874 I0424 17:40:51.252164 139766024124224 metric_logger.py:185] completed step: 1, seconds: 1.304, TFLOP/s/device: 10.419, Tokens/s/device: 1570.481, total_weights: 65536, loss: 10.874 I0424 17:40:51.682569 139766024124224 metric_logger.py:185] completed step: 2, seconds: 0.012, TFLOP/s/device: 1114.613, Tokens/s/device: 168006.563, total_weights: 65536, loss: 10.560 I0424 17:40:51.829965 139766024124224 metric_logger.py:185] completed step: 3, seconds: 0.422, TFLOP/s/device: 32.233, Tokens/s/device: 4858.561, total_weights: 65536, loss: 9.979 I0424 17:40:51.836080 139766024124224 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 17:40:52.124964 139766024124224 checkpointing.py:798] Waited 0.28885412216186523 seconds for step 5 to finish before starting checkpointing. I0424 17:40:52.127692 139766024124224 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:40:53.789586 2622 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 17:40:55.806663 139640417789696 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/0/items/array_metadatas/process_6 I0424 17:41:28.290952 139640409396992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 41.505 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.05756640434265 s) (per-host) I0424 17:41:28.291078 139640409396992 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.505489s. I0424 17:41:37.323892 139640409396992 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 46.538288s. I0424 17:41:37.324192 139638832355072 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:41:37.324316 139638832355072 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:41:37.324367 139638832355072 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:41:37.325977 139638832355072 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 17:41:37.326185 139766024124224 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0424 17:41:37.326331 139766024124224 checkpoint_manager.py:1458] Waiting for previous save to complete took 45.198645 seconds. If this number is high, consider checkpointing less frequently. I0424 17:41:37.328271 139766024124224 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0424 17:41:37.332286 139766024124224 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/5. I0424 17:41:37.885055 139766024124224 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:41:37.885224 139766024124224 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 17:41:37.931455 139766024124224 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048503s I0424 17:41:37.931626 139766024124224 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.100 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3024613857269287 s) (per-host) I0424 17:41:37.931677 139766024124224 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.302521s (batch_requests_ready=0.249189s, total_serialization_initiated=0.053264s, others=0.000068s) I0424 17:41:37.931770 139766024124224 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.307340s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.307324) I0424 17:41:37.932657 139638840747776 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:41:37.932787 139766024124224 async_checkpointer.py:561] Finished blocking save. Time taken: 0.604442s. 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/5. I0424 17:41:37.978748 139766024124224 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:41:37.979069 139638832355072 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:41:37.979244 139766024124224 standard_logger.py:34] {'step': 5, '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': 1777052452.12766, 'wait_for_prev_duration_secs': 45.19864535331726, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052497.3283114, 'checkpointer_blocking_duration_secs': 0.604649543762207, 'get_old_steps_start_time': 1777052497.9329805, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1777052452.1255236, 'checkpoint_manager_blocking_duration_secs': 45.8536856174469} I0424 17:41:37.979362 139766024124224 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 17:41:37.980157 139766024124224 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.458, Tokens/s/device: 12579.698, total_weights: 65536, loss: 9.457 I0424 17:41:37.987746 139766024124224 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.379, Tokens/s/device: 13924.300, total_weights: 65536, loss: 8.955 I0424 17:41:57.905525 139766024124224 metric_logger.py:185] completed step: 6, seconds: 46.150, TFLOP/s/device: 0.294, Tokens/s/device: 44.377, total_weights: 65536, loss: 8.465 I0424 17:41:58.052904 139766024124224 metric_logger.py:185] completed step: 7, seconds: 19.773, TFLOP/s/device: 0.687, Tokens/s/device: 103.578, total_weights: 65536, loss: 7.998 I0424 17:41:58.200480 139766024124224 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.582, Tokens/s/device: 13352.110, total_weights: 65536, loss: 7.567 I0424 17:41:58.348944 139766024124224 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 17:41:58.349591 139766024124224 checkpointing.py:798] Waited 0.0006580352783203125 seconds for step 9 to finish before starting checkpointing. I0424 17:41:58.351489 139766024124224 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:42:02.433475 139640417789696 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/5/items/array_metadatas/process_6 I0424 17:42:39.078215 139638840747776 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.706 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.448991537094116 s) (per-host) I0424 17:42:39.078336 139638840747776 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.145468s. I0424 17:42:48.525403 139638840747776 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 70.592522s. I0424 17:42:48.525665 139638832355072 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:42:48.525730 139638832355072 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:42:48.525783 139638832355072 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:42:48.527495 139638832355072 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 17:42:48.527670 139766024124224 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0424 17:42:48.527802 139766024124224 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.176323 seconds. If this number is high, consider checkpointing less frequently. I0424 17:42:48.529434 139766024124224 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0424 17:42:48.532871 139766024124224 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/9. I0424 17:42:49.090041 139766024124224 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:42:49.090161 139766024124224 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 17:42:49.133832 139766024124224 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045796s I0424 17:42:49.134000 139766024124224 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.855 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3177449703216553 s) (per-host) I0424 17:42:49.134049 139766024124224 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.317805s (batch_requests_ready=0.267502s, total_serialization_initiated=0.050237s, others=0.000066s) I0424 17:42:49.134159 139766024124224 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.322381s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.322368) I0424 17:42:49.135065 139640409396992 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:42:49.135237 139766024124224 async_checkpointer.py:561] Finished blocking save. Time taken: 0.605732s. 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/9. I0424 17:42:49.146747 139766024124224 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:42:49.146982 139638832355072 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:42:49.147149 139766024124224 standard_logger.py:34] {'step': 9, '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': 1777052518.3514616, 'wait_for_prev_duration_secs': 50.17632341384888, 'time_between_consecutive_saves_sec': 21.025446891784668, 'checkpointer_blocking_start_time': 1777052568.5294733, 'checkpointer_blocking_duration_secs': 0.605879545211792, 'get_old_steps_start_time': 1777052569.135366, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1777052518.3498158, 'checkpoint_manager_blocking_duration_secs': 50.79730200767517} I0424 17:42:49.147261 139766024124224 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 17:42:49.147304 139766024124224 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:42:53.897121 139640417789696 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/9/items/array_metadatas/process_6 I0424 17:43:31.081301 139640409396992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.374 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.26502585411072 s) (per-host) I0424 17:43:31.081403 139640409396992 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.946216s. I0424 17:43:36.647391 139640409396992 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.512188s. I0424 17:43:36.647681 139638832355072 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:43:36.647811 139638832355072 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:43:36.647853 139638832355072 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:43:36.649389 139638832355072 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 17:43:36.649568 139766024124224 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0424 17:43:36.649731 139766024124224 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:43:36.650379 139766024124224 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.483, Tokens/s/device: 13940.033, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 24 17:43:49 UTC 2026 EXIT_CODE=0