XPK Start: Wed Apr 22 18:33:35 UTC 2026 2026-04-22 18:33:39.324319: 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 00:00:1776882819.337392 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776882819.341409 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776882819.352785 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882819.352804 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882819.352806 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882819.352808 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 18:33:58.682155: 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) I0422 18:33:59.213821 132056917698368 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 18:34:08,255:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 18:34:08.255424 132056917698368 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 18:34:08,257:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-av1wx-slice-job-0-0.mt-15-ocdbt-true-rs-av1wx:8482 I0422 18:34:08.257794 132056917698368 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-av1wx-slice-job-0-0.mt-15-ocdbt-true-rs-av1wx:8482 I0422 18:34:09.205788 132056917698368 max_utils.py:284] Jax distributed system initialized! I0422 18:34:16.398143 132056917698368 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:34:16.398254 132056917698368 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:34:16.398296 132056917698368 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 I0422 18:34:16.398331 132056917698368 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:34:17.028253 132056917698368 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:34:17.146162 132056917698368 checkpointing.py:688] Setting up checkpoint logger... I0422 18:34:17.146283 132056917698368 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 18:34:17.146327 132056917698368 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 18:34:17.146543 132056917698368 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 0x781a6bd6d580>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:34:18.420660 132056917698368 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 18:34:18.420979 132056917698368 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 0x7806084c58b0>}, handler_registry=None I0422 18:34:18.421286 132056917698368 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7806084c58b0>` 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`. I0422 18:34:18.421337 132056917698368 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7805d85525d0>` 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`. I0422 18:34:18.421373 132056917698368 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 0x7806084c58b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7806084c58b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7805d85525d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7805d85525d0>}). I0422 18:34:18.421701 132056917698368 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 18:34:18.421780 132056917698368 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7805d86d0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:34:20.831465 131936185984768 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776882643575821110, 'commit_timestamp_nsecs': 1776882689403945469, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5/_CHECKPOINT_METADATA I0422 18:34:21.092172 131936185984768 checkpoint_manager.py:1877] Missing metrics for step 5 I0422 18:34:21.564365 132056917698368 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints I0422 18:34:22.069037 132056917698368 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7805d82f3d70> I0422 18:34:22.069270 132056917698368 checkpointing.py:302] Checkpoint manager created! I0422 18:34:23.006893 132056917698368 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:34:23.007014 132056917698368 checkpointing.py:582] restoring from this run's directory step 9 I0422 18:34:23.508020 132056917698368 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9. I0422 18:34:25.345743 1428 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 18:34:32.506153 132056917698368 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 220.522 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.166080474853516 s) (per-host) I0422 18:34:32.514633 132056917698368 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 208.730 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.582052707672119 s) (per-host) I0422 18:34:33.858384 132056917698368 checkpointer.py:318] Finished restoring checkpoint in 10.60 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9. I0422 18:34:33.858543 132056917698368 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776882863.2610753, 'checkpointer_duration_secs': 10.597436666488647, 'checkpoint_manager_start_time': 1776882863.007655, 'checkpoint_manager_duration_secs': 10.850859642028809} fsdp: 32 I0422 18:34:33.943112 132056917698368 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 18:34:33.943233 132056917698368 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:34:33.943286 132056917698368 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:34:33.943330 132056917698368 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 18:34:33.943367 132056917698368 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:34:33.943402 132056917698368 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:34:33.943439 132056917698368 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:34:33.943477 132056917698368 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 18:34:33.943513 132056917698368 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:34:33.943553 132056917698368 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:34:33.943589 132056917698368 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 18:34:33.943624 132056917698368 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 18:34:34.352204 132056917698368 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:34:34.352363 132056917698368 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:34:34.357853 132056917698368 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:34:34.357951 132056917698368 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:34:34.518581 132056917698368 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:34:34.518717 132056917698368 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:34:34.677972 132056917698368 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:34:34.678131 132056917698368 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:34:34.844370 132056917698368 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:34:34.844506 132056917698368 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:34:34.998009 132056917698368 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:34:34.998161 132056917698368 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:34:35.152647 132056917698368 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:34:35.152781 132056917698368 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:34:35.314953 132056917698368 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:34:35.315088 132056917698368 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:34:35.338499 132056917698368 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:34:35.338600 132056917698368 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 18:34:46.191285 132056917698368 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. I0422 18:34:46.193733 132056917698368 metric_logger.py:289] number parameters: 1.104 billion I0422 18:34:57.268133 132056917698368 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0422 18:34:57.517053 132056917698368 checkpointing.py:798] Waited 0.24891901016235352 seconds for step 10 to finish before starting checkpointing. I0422 18:34:57.519220 132056917698368 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:34:57.521285 132056917698368 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0422 18:34:57.524275 132056917698368 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/10. I0422 18:34:58.120012 132056917698368 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 18:34:58.121168 132056917698368 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:34:58.121224 132056917698368 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 I0422 18:34:58.363897 132056917698368 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245088s I0422 18:34:58.364065 132056917698368 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.958 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5214447975158691 s) (per-host) I0422 18:34:58.364136 132056917698368 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.521526s (batch_requests_ready=0.271603s, total_serialization_initiated=0.249834s, others=0.000088s) I0422 18:34:58.364233 132056917698368 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.525633s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.525614) I0422 18:34:58.365153 131935582004992 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:34:58.365287 132056917698368 async_checkpointer.py:561] Finished blocking save. Time taken: 0.843932s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/10. I0422 18:34:58.791286 132056917698368 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:34:58.791684 131930783737600 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:34:58.791847 132056917698368 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882897.5191998, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776882897.5213218, 'checkpointer_blocking_duration_secs': 0.8441429138183594, 'get_old_steps_start_time': 1776882898.3654823, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776882897.5174422, 'checkpoint_manager_blocking_duration_secs': 1.2743685245513916} I0422 18:34:58.791959 132056917698368 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0422 18:34:58.792031 132056917698368 max_utils.py:750] Memstats: After params initialized: I0422 18:34:58.792088 132056917698368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 18:34:58.792140 132056917698368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 18:34:58.792168 132056917698368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 18:34:58.792192 132056917698368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 18:34:59.064254 132056917698368 metric_logger.py:185] completed step: 10, seconds: 11.073, TFLOP/s/device: 1.227, Tokens/s/device: 184.949, total_weights: 65536, loss: 6.827 I0422 18:34:59.239039 132056917698368 metric_logger.py:185] completed step: 11, seconds: 1.794, TFLOP/s/device: 7.572, Tokens/s/device: 1141.292, total_weights: 65536, loss: 6.522 I0422 18:34:59.386868 132056917698368 metric_logger.py:185] completed step: 12, seconds: 0.019, TFLOP/s/device: 703.960, Tokens/s/device: 106108.492, total_weights: 65536, loss: 6.260 I0422 18:34:59.534440 132056917698368 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.141, Tokens/s/device: 12531.972, total_weights: 65536, loss: 6.039 I0422 18:34:59.541701 132056917698368 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0422 18:34:59.829479 132056917698368 checkpointing.py:798] Waited 0.2877476215362549 seconds for step 15 to finish before starting checkpointing. I0422 18:34:59.832016 132056917698368 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:35:03.751028 131931341567744 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6 I0422 18:35:39.658319 131935582004992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.775 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.81565070152283 s) (per-host) I0422 18:35:39.658442 131935582004992 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.293074s. I0422 18:35:48.070569 131935582004992 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.705186s. I0422 18:35:48.070855 131930783737600 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:35:48.070974 131930783737600 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:35:48.071020 131930783737600 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:35:48.072518 131930783737600 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0422 18:35:48.072706 132056917698368 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. W0422 18:35:48.072843 132056917698368 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.240832 seconds. If this number is high, consider checkpointing less frequently. I0422 18:35:48.074776 132056917698368 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0422 18:35:48.079020 132056917698368 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/15. I0422 18:35:48.666227 132056917698368 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:35:48.666327 132056917698368 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 I0422 18:35:48.710227 132056917698368 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046129s I0422 18:35:48.710396 132056917698368 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.082 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.303525447845459 s) (per-host) I0422 18:35:48.710450 132056917698368 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.303589s (batch_requests_ready=0.252745s, total_serialization_initiated=0.050773s, others=0.000072s) I0422 18:35:48.710545 132056917698368 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308475s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.308461) I0422 18:35:48.711440 131934533441280 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:35:48.711617 132056917698368 async_checkpointer.py:561] Finished blocking save. Time taken: 0.636769s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/15. I0422 18:35:49.154863 132056917698368 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:35:49.155259 131930783737600 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:35:49.155435 132056917698368 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882899.8319874, 'wait_for_prev_duration_secs': 48.24083161354065, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776882948.0748155, 'checkpointer_blocking_duration_secs': 0.6369073390960693, 'get_old_steps_start_time': 1776882948.711742, 'get_old_steps_duration_secs': 2.384185791015625e-05, 'checkpoint_manager_blocking_start_time': 1776882899.8298295, 'checkpoint_manager_blocking_duration_secs': 49.325570583343506} I0422 18:35:49.155562 132056917698368 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0422 18:35:49.156387 132056917698368 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.942, Tokens/s/device: 13858.438, total_weights: 65536, loss: 5.858 I0422 18:35:49.163923 132056917698368 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.593, Tokens/s/device: 13805.935, total_weights: 65536, loss: 5.713 I0422 18:36:07.369586 132056917698368 metric_logger.py:185] completed step: 16, seconds: 49.621, TFLOP/s/device: 0.274, Tokens/s/device: 41.273, total_weights: 65536, loss: 5.599 I0422 18:36:07.517029 132056917698368 metric_logger.py:185] completed step: 17, seconds: 18.060, TFLOP/s/device: 0.752, Tokens/s/device: 113.399, total_weights: 65536, loss: 5.511 I0422 18:36:07.664684 132056917698368 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.567, Tokens/s/device: 13349.760, total_weights: 65536, loss: 5.442 I0422 18:36:07.813284 132056917698368 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0422 18:36:07.813916 132056917698368 checkpointing.py:798] Waited 0.0006439685821533203 seconds for step 19 to finish before starting checkpointing. I0422 18:36:07.816489 132056917698368 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:36:10.747122 131931341567744 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6 I0422 18:36:47.850630 131934533441280 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.573 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.443721294403076 s) (per-host) I0422 18:36:47.850750 131934533441280 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.139180s. I0422 18:36:56.856740 131934533441280 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.145156s. I0422 18:36:56.857033 131930783737600 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:36:56.857179 131930783737600 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:36:56.857228 131930783737600 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:36:56.859114 131930783737600 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0422 18:36:56.859310 132056917698368 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. W0422 18:36:56.859438 132056917698368 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.042958 seconds. If this number is high, consider checkpointing less frequently. I0422 18:36:56.861132 132056917698368 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0422 18:36:56.864580 132056917698368 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/19. I0422 18:36:58.293680 132056917698368 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:36:58.293776 132056917698368 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 I0422 18:36:58.338365 132056917698368 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047120s I0422 18:36:58.338530 132056917698368 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.074 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7438344955444336 s) (per-host) I0422 18:36:58.338579 132056917698368 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.743893s (batch_requests_ready=0.692071s, total_serialization_initiated=0.051758s, others=0.000064s) I0422 18:36:58.338672 132056917698368 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.748441s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.748430) I0422 18:36:58.339554 131934499870464 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:36:58.339715 132056917698368 async_checkpointer.py:561] Finished blocking save. Time taken: 1.478515s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/19. I0422 18:36:58.341664 132056917698368 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:36:58.341891 131930783737600 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:36:58.342041 132056917698368 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882967.8164601, 'wait_for_prev_duration_secs': 49.04295778274536, 'time_between_consecutive_saves_sec': 19.74390411376953, 'checkpointer_blocking_start_time': 1776883016.8611689, 'checkpointer_blocking_duration_secs': 1.4786524772644043, 'get_old_steps_start_time': 1776883018.3398361, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1776882967.814143, 'checkpoint_manager_blocking_duration_secs': 50.527868032455444} I0422 18:36:58.342166 132056917698368 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0422 18:36:58.342209 132056917698368 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:37:03.439640 131935582004992 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6 I0422 18:37:40.135891 131934499870464 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.131 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.54115319252014 s) (per-host) I0422 18:37:40.136007 131934499870464 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.796340s. I0422 18:37:49.159254 131934499870464 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.819571s. I0422 18:37:49.159549 131930783737600 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:37:49.159672 131930783737600 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:37:49.159715 131930783737600 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:37:49.161163 131930783737600 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0422 18:37:49.161350 132056917698368 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. I0422 18:37:49.161517 132056917698368 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:37:49.162180 132056917698368 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.352, Tokens/s/device: 13920.231, 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: Wed Apr 22 18:38:00 UTC 2026 EXIT_CODE=0