XPK Start: Thu Apr 23 18:29:29 UTC 2026 2026-04-23 18:29:34.036942: 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 00:00:1776968974.049904 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776968974.053660 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776968974.065486 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776968974.065506 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776968974.065509 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776968974.065510 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 18:29:53.176882: 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 18:29:53.706000 139999630419776 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 18:30:02,746:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 18:30:02.746217 139999630419776 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 18:30:02,748:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-z60b6-slice-job-0-0.mt-15-ocdbt-true-sv-z60b6:8482 I0423 18:30:02.748577 139999630419776 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-z60b6-slice-job-0-0.mt-15-ocdbt-true-sv-z60b6:8482 I0423 18:30:04.028093 139999630419776 max_utils.py:284] Jax distributed system initialized! I0423 18:30:10.891058 139999630419776 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 18:30:10.891161 139999630419776 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 18:30:10.891201 139999630419776 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 18:30:10.891235 139999630419776 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 18:30:11.514829 139999630419776 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 18:30:11.630391 139999630419776 checkpointing.py:688] Setting up checkpoint logger... I0423 18:30:11.630514 139999630419776 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 18:30:11.630559 139999630419776 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 18:30:11.630779 139999630419776 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 0x7f53b2571130>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 18:30:15.457381 139999630419776 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 18:30:15.457633 139999630419776 checkpoint_manager.py:709] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f3f4024b410>}, handler_registry=None I0423 18:30:15.457874 139999630419776 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f3f4024b410>` 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 18:30:15.457935 139999630419776 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f3f402454f0>` 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 18:30:15.457972 139999630419776 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 0x7f3f4024b410>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f3f4024b410>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f3f402454f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f3f402454f0>}). I0423 18:30:15.458307 139999630419776 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 18:30:15.458385 139999630419776 async_checkpointer.py:177] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f3ef86e0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 18:30:16.145622 139999630419776 checkpoint_manager.py:1818] Found 0 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_15_ocdbt_true/checkpoints I0423 18:30:16.159706 139999630419776 checkpoint_manager.py:929] [process=3][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_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f3ef8654ec0> I0423 18:30:16.159830 139999630419776 checkpointing.py:302] Checkpoint manager created! I0423 18:30:17.091683 139999630419776 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 18:30:17.091802 139999630419776 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 18:30:18.530198 139999630419776 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0423 18:30:18.530318 139999630419776 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:30:18.530373 139999630419776 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:30:18.530418 139999630419776 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0423 18:30:18.530456 139999630419776 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:30:18.530498 139999630419776 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:30:18.530535 139999630419776 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:30:18.530572 139999630419776 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0423 18:30:18.530608 139999630419776 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:30:18.530644 139999630419776 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:30:18.530677 139999630419776 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0423 18:30:18.530709 139999630419776 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0423 18:30:18.926289 139999630419776 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:30:18.926443 139999630419776 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:30:18.932025 139999630419776 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:30:18.932121 139999630419776 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:30:19.090731 139999630419776 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:30:19.090865 139999630419776 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:30:19.249915 139999630419776 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:30:19.250068 139999630419776 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:30:19.415240 139999630419776 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 18:30:19.415378 139999630419776 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:30:19.569097 139999630419776 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 18:30:19.569258 139999630419776 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:30:19.718125 139999630419776 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 18:30:19.718256 139999630419776 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:30:19.878492 139999630419776 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 18:30:19.878633 139999630419776 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:30:19.903029 139999630419776 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 18:30:19.903129 139999630419776 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 18:30:30.860205 139999630419776 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 18:30:30.862799 139999630419776 metric_logger.py:289] number parameters: 1.104 billion I0423 18:30:42.030216 139999630419776 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 18:30:42.182678 139999630419776 checkpointing.py:798] Waited 0.1524486541748047 seconds for step 0 to finish before starting checkpointing. I0423 18:30:42.185060 139999630419776 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:30:42.186825 139999630419776 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 0 I0423 18:30:42.189561 139999630419776 async_checkpointer.py:452] [process=3] 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_15_ocdbt_true/checkpoints/0. I0423 18:30:43.008164 139999630419776 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 18:30:43.009285 139999630419776 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:30:43.009345 139999630419776 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 18:30:43.257392 139999630419776 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250374s I0423 18:30:43.257575 139999630419776 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.881 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5354931354522705 s) (per-host) I0423 18:30:43.257627 139999630419776 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.535558s (batch_requests_ready=0.278455s, total_serialization_initiated=0.257030s, others=0.000073s) I0423 18:30:43.257721 139999630419776 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.539600s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.539581) I0423 18:30:43.258709 139876414519040 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0423 18:30:43.258881 139999630419776 async_checkpointer.py:561] Finished blocking save. Time taken: 1.071988s. 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_15_ocdbt_true/checkpoints/0. I0423 18:30:43.265779 139999630419776 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:30:43.266043 139871549110016 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:30:43.266173 139999630419776 standard_logger.py:34] {'step': 0, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776969042.1850417, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776969042.1868622, 'checkpointer_blocking_duration_secs': 1.0721404552459717, 'get_old_steps_start_time': 1776969043.2590253, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776969042.1831117, 'checkpoint_manager_blocking_duration_secs': 1.0830235481262207} I0423 18:30:43.266278 139999630419776 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 18:30:43.266353 139999630419776 max_utils.py:750] Memstats: After params initialized: I0423 18:30:43.266403 139999630419776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0423 18:30:43.266434 139999630419776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0423 18:30:43.266461 139999630419776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0423 18:30:43.266486 139999630419776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0423 18:30:43.537362 139999630419776 metric_logger.py:185] completed step: 0, seconds: 11.166, TFLOP/s/device: 1.217, Tokens/s/device: 183.407, total_weights: 65536, loss: 10.874 I0423 18:30:43.719512 139999630419776 metric_logger.py:185] completed step: 1, seconds: 1.505, TFLOP/s/device: 9.026, Tokens/s/device: 1360.527, total_weights: 65536, loss: 10.874 I0423 18:30:44.148764 139999630419776 metric_logger.py:185] completed step: 2, seconds: 0.036, TFLOP/s/device: 372.322, Tokens/s/device: 56120.352, total_weights: 65536, loss: 10.560 I0423 18:30:44.296274 139999630419776 metric_logger.py:185] completed step: 3, seconds: 0.420, TFLOP/s/device: 32.338, Tokens/s/device: 4874.276, total_weights: 65536, loss: 9.979 I0423 18:30:44.302275 139999630419776 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0423 18:30:44.591190 139999630419776 checkpointing.py:798] Waited 0.28888654708862305 seconds for step 5 to finish before starting checkpointing. I0423 18:30:44.594187 139999630419776 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:30:45.915385 2602 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 18:30:47.880088 139876901033728 array_metadata_store.py:203] [process=3][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_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_3 I0423 18:31:21.157700 139876414519040 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 41.097 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.43557906150818 s) (per-host) I0423 18:31:21.157820 139876414519040 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.898998s. I0423 18:31:28.596141 139876414519040 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 45.337303s. I0423 18:31:28.596446 139871549110016 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:31:28.596582 139871549110016 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:31:28.596632 139871549110016 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:31:28.598080 139871549110016 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 18:31:28.598272 139999630419776 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0423 18:31:28.598418 139999630419776 checkpoint_manager.py:1458] Waiting for previous save to complete took 44.004234 seconds. If this number is high, consider checkpointing less frequently. I0423 18:31:28.600435 139999630419776 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 5 I0423 18:31:28.604529 139999630419776 async_checkpointer.py:452] [process=3] 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_15_ocdbt_true/checkpoints/5. I0423 18:31:29.575485 139999630419776 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:31:29.575638 139999630419776 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 18:31:29.620191 139999630419776 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046678s I0423 18:31:29.620351 139999630419776 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.033 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30649256706237793 s) (per-host) I0423 18:31:29.620401 139999630419776 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306552s (batch_requests_ready=0.254662s, total_serialization_initiated=0.051823s, others=0.000068s) I0423 18:31:29.620499 139999630419776 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311335s (all_items=0.000014s, per_item={'items': '0.00001431'}, temp_paths=0.311320) I0423 18:31:29.621331 139871582680832 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0423 18:31:29.621485 139999630419776 async_checkpointer.py:561] Finished blocking save. Time taken: 1.020983s. 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_15_ocdbt_true/checkpoints/5. I0423 18:31:30.073201 139999630419776 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:31:30.073581 139871549110016 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:31:30.073755 139999630419776 standard_logger.py:34] {'step': 5, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776969044.5941584, 'wait_for_prev_duration_secs': 44.004234313964844, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776969088.6004736, 'checkpointer_blocking_duration_secs': 1.0211162567138672, 'get_old_steps_start_time': 1776969089.6216085, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776969044.591536, 'checkpoint_manager_blocking_duration_secs': 45.482181787490845} I0423 18:31:30.073878 139999630419776 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0423 18:31:30.074684 139999630419776 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.430, Tokens/s/device: 12575.450, total_weights: 65536, loss: 9.457 I0423 18:31:30.082064 139999630419776 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.321, Tokens/s/device: 13915.690, total_weights: 65536, loss: 8.955 I0423 18:31:50.084029 139999630419776 metric_logger.py:185] completed step: 6, seconds: 45.778, TFLOP/s/device: 0.297, Tokens/s/device: 44.737, total_weights: 65536, loss: 8.465 I0423 18:31:50.231255 139999630419776 metric_logger.py:185] completed step: 7, seconds: 19.856, TFLOP/s/device: 0.684, Tokens/s/device: 103.141, total_weights: 65536, loss: 7.998 I0423 18:31:50.378658 139999630419776 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.401, Tokens/s/device: 13324.832, total_weights: 65536, loss: 7.567 I0423 18:31:50.527081 139999630419776 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 18:31:50.527706 139999630419776 checkpointing.py:798] Waited 0.0006291866302490234 seconds for step 9 to finish before starting checkpointing. I0423 18:31:50.530325 139999630419776 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:31:53.580209 139876901033728 array_metadata_store.py:203] [process=3][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_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_3 I0423 18:32:30.067552 139871582680832 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 26.000 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.753657579422 s) (per-host) I0423 18:32:30.067660 139871582680832 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.446219s. I0423 18:32:40.117743 139871582680832 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 70.496286s. I0423 18:32:40.118061 139871549110016 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:32:40.118180 139871549110016 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:32:40.118227 139871549110016 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:32:40.119816 139871549110016 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0423 18:32:40.119991 139999630419776 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0423 18:32:40.120110 139999630419776 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.589797 seconds. If this number is high, consider checkpointing less frequently. I0423 18:32:40.121705 139999630419776 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 9 I0423 18:32:40.125036 139999630419776 async_checkpointer.py:452] [process=3] 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_15_ocdbt_true/checkpoints/9. I0423 18:32:41.402241 139999630419776 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:32:41.402340 139999630419776 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 18:32:41.446646 139999630419776 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046399s I0423 18:32:41.446810 139999630419776 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.025 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30696797370910645 s) (per-host) I0423 18:32:41.446857 139999630419776 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307026s (batch_requests_ready=0.255871s, total_serialization_initiated=0.051092s, others=0.000063s) I0423 18:32:41.446962 139999630419776 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311171s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.311159) I0423 18:32:41.447841 139876414519040 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0423 18:32:41.447980 139999630419776 async_checkpointer.py:561] Finished blocking save. Time taken: 1.326206s. 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_15_ocdbt_true/checkpoints/9. I0423 18:32:41.889354 139999630419776 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:32:41.889742 139871549110016 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:32:41.889909 139999630419776 standard_logger.py:34] {'step': 9, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776969110.530297, 'wait_for_prev_duration_secs': 49.589797019958496, 'time_between_consecutive_saves_sec': 21.932176113128662, 'checkpointer_blocking_start_time': 1776969160.1217444, 'checkpointer_blocking_duration_secs': 1.3264095783233643, 'get_old_steps_start_time': 1776969161.4481661, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1776969110.5279396, 'checkpoint_manager_blocking_duration_secs': 51.36193656921387} I0423 18:32:41.890032 139999630419776 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 18:32:41.890075 139999630419776 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:32:46.221172 139876901033728 array_metadata_store.py:203] [process=3][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_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_3 I0423 18:33:23.258321 139876414519040 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 37.504 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.1184401512146 s) (per-host) I0423 18:33:23.258437 139876414519040 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.810378s. I0423 18:33:29.263159 139876414519040 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 47.815084s. I0423 18:33:29.263472 139871549110016 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:33:29.263607 139871549110016 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:33:29.263652 139871549110016 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:33:29.265145 139871549110016 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 18:33:29.265314 139999630419776 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0423 18:33:29.265488 139999630419776 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:33:29.266048 139999630419776 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.542, Tokens/s/device: 13948.958, 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: Thu Apr 23 18:33:40 UTC 2026 EXIT_CODE=0