XPK Start: Sun Apr 19 06:10:28 UTC 2026 2026-04-19 06:10:32.905356: 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 00:00:1776579032.918253 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776579032.921942 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776579032.933160 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579032.933178 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579032.933181 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579032.933182 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 06:10:52.181196: 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) I0419 06:10:52.703863 133792770201408 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 06:11:01,744:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 06:11:01.744434 133792770201408 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 06:11:01,746:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-rs-va8ng-slice-job-0-0.mt-15-ocdbt-false-rs-va8ng:8482 I0419 06:11:01.746918 133792770201408 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-va8ng-slice-job-0-0.mt-15-ocdbt-false-rs-va8ng:8482 I0419 06:11:03.153283 133792770201408 max_utils.py:284] Jax distributed system initialized! I0419 06:11:10.195429 133792770201408 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 06:11:10.195537 133792770201408 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 06:11:10.195578 133792770201408 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 I0419 06:11:10.195613 133792770201408 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 06:11:10.818165 133792770201408 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 06:11:10.933978 133792770201408 checkpointing.py:688] Setting up checkpoint logger... I0419 06:11:10.934098 133792770201408 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0419 06:11:10.934144 133792770201408 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 06:11:10.934386 133792770201408 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x79aef253fbf0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 06:11:12.214974 133792770201408 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 06:11:12.215408 133792770201408 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 0x79ad493e2270>}, handler_registry=None I0419 06:11:12.215658 133792770201408 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79ad493e2270>` 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`. I0419 06:11:12.215710 133792770201408 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x799b07cf6090>` 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`. I0419 06:11:12.215747 133792770201408 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 0x79ad493e2270>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79ad493e2270>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x799b07cf6090>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x799b07cf6090>}). I0419 06:11:12.216080 133792770201408 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 06:11:12.216159 133792770201408 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x799b0700cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 06:11:15.575451 133672158078720 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776578938220551934, 'commit_timestamp_nsecs': 1776578953514851848, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/9/_CHECKPOINT_METADATA I0419 06:11:15.821875 133672158078720 checkpoint_manager.py:1877] Missing metrics for step 9 I0419 06:11:17.391019 133792770201408 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints I0419 06:11:17.873845 133792770201408 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x799b06024800> I0419 06:11:17.874016 133792770201408 checkpointing.py:302] Checkpoint manager created! I0419 06:11:18.807066 133792770201408 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 06:11:18.807183 133792770201408 checkpointing.py:582] restoring from this run's directory step 9 I0419 06:11:19.309810 133792770201408 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/9. I0419 06:11:25.062295 1454 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 06:11:35.708901 133792770201408 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 125.780 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.563842296600342 s) (per-host) I0419 06:11:35.727860 133792770201408 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 114.341 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.841149806976318 s) (per-host) I0419 06:11:36.103091 133792770201408 checkpointer.py:318] Finished restoring checkpoint in 17.05 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/9. I0419 06:11:36.103286 133792770201408 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776579079.0569644, 'checkpointer_duration_secs': 17.046252965927124, 'checkpoint_manager_start_time': 1776579078.8078728, 'checkpoint_manager_duration_secs': 17.29534673690796} fsdp: 32 I0419 06:11:36.183272 133792770201408 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 06:11:36.183392 133792770201408 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:11:36.183445 133792770201408 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:11:36.183492 133792770201408 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 06:11:36.183537 133792770201408 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:11:36.183575 133792770201408 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:11:36.183613 133792770201408 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:11:36.183650 133792770201408 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 06:11:36.183686 133792770201408 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:11:36.183719 133792770201408 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:11:36.183753 133792770201408 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 06:11:36.183786 133792770201408 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 06:11:36.581539 133792770201408 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:11:36.581688 133792770201408 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:11:36.587058 133792770201408 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:11:36.587151 133792770201408 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:11:36.733797 133792770201408 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:11:36.733927 133792770201408 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:11:36.877640 133792770201408 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:11:36.877773 133792770201408 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:11:37.031377 133792770201408 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:11:37.031517 133792770201408 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:11:37.170201 133792770201408 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:11:37.170351 133792770201408 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:11:37.314011 133792770201408 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:11:37.314143 133792770201408 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:11:37.464158 133792770201408 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 06:11:37.464352 133792770201408 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:11:37.488107 133792770201408 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 06:11:37.488201 133792770201408 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 06:11:48.738543 133792770201408 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. I0419 06:11:48.741143 133792770201408 metric_logger.py:289] number parameters: 1.104 billion I0419 06:11:59.700146 133792770201408 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 06:11:59.857386 133792770201408 checkpointing.py:798] Waited 0.1572256088256836 seconds for step 10 to finish before starting checkpointing. I0419 06:11:59.859747 133792770201408 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:11:59.861688 133792770201408 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0419 06:11:59.864433 133792770201408 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/10. I0419 06:12:00.438541 133792770201408 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 06:12:00.439696 133792770201408 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:12:00.439755 133792770201408 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 I0419 06:12:00.685459 133792770201408 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.248055s I0419 06:12:00.685630 133792770201408 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.093 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4987804889678955 s) (per-host) I0419 06:12:00.685683 133792770201408 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.498844s (batch_requests_ready=0.246019s, total_serialization_initiated=0.252753s, others=0.000072s) I0419 06:12:00.685778 133792770201408 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.503005s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.502985) I0419 06:12:00.686645 133668878137088 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:12:00.686816 133792770201408 async_checkpointer.py:561] Finished blocking save. Time taken: 0.825054s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/10. I0419 06:12:01.202530 133792770201408 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:12:01.202883 133666732934912 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:12:01.203048 133792770201408 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579119.8597298, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776579119.861727, 'checkpointer_blocking_duration_secs': 0.8251972198486328, 'get_old_steps_start_time': 1776579120.6869447, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776579119.8577642, 'checkpoint_manager_blocking_duration_secs': 1.3452439308166504} I0419 06:12:01.203161 133792770201408 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 06:12:01.203270 133792770201408 max_utils.py:750] Memstats: After params initialized: I0419 06:12:01.203336 133792770201408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 06:12:01.203371 133792770201408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 06:12:01.203419 133792770201408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 06:12:01.203446 133792770201408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 06:12:01.479558 133792770201408 metric_logger.py:185] completed step: 10, seconds: 10.958, TFLOP/s/device: 1.240, Tokens/s/device: 186.896, total_weights: 65536, loss: 6.827 I0419 06:12:01.663089 133792770201408 metric_logger.py:185] completed step: 11, seconds: 1.778, TFLOP/s/device: 7.643, Tokens/s/device: 1152.039, total_weights: 65536, loss: 6.522 I0419 06:12:01.811070 133792770201408 metric_logger.py:185] completed step: 12, seconds: 0.028, TFLOP/s/device: 484.977, Tokens/s/device: 73101.085, total_weights: 65536, loss: 6.260 I0419 06:12:01.958621 133792770201408 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 83.102, Tokens/s/device: 12525.994, total_weights: 65536, loss: 6.039 I0419 06:12:01.965098 133792770201408 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 06:12:02.253741 133792770201408 checkpointing.py:798] Waited 0.28861331939697266 seconds for step 15 to finish before starting checkpointing. I0419 06:12:02.257119 133792770201408 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:12:05.802350 133666749720320 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6 I0419 06:12:12.037416 133668878137088 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 133.294 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.850499629974365 s) (per-host) I0419 06:12:12.037534 133668878137088 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.350781s. I0419 06:12:18.933845 133668878137088 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 18.247076s. I0419 06:12:18.934132 133666732934912 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:12:18.934292 133666732934912 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:12:18.934340 133666732934912 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:12:18.936033 133666732934912 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 06:12:18.936230 133792770201408 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. W0419 06:12:18.936383 133792770201408 checkpoint_manager.py:1458] Waiting for previous save to complete took 16.679281 seconds. If this number is high, consider checkpointing less frequently. I0419 06:12:18.938261 133792770201408 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0419 06:12:18.942387 133792770201408 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/15. I0419 06:12:19.479270 133792770201408 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:12:19.479373 133792770201408 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 I0419 06:12:19.517450 133792770201408 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040346s I0419 06:12:19.517623 133792770201408 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.187 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29741573333740234 s) (per-host) I0419 06:12:19.517674 133792770201408 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.297476s (batch_requests_ready=0.252542s, total_serialization_initiated=0.044865s, others=0.000069s) I0419 06:12:19.517765 133792770201408 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.302341s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.302327) I0419 06:12:19.518614 133664581621504 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:12:19.518776 133792770201408 async_checkpointer.py:561] Finished blocking save. Time taken: 0.580441s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/15. I0419 06:12:20.242144 133792770201408 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:12:20.242563 133666732934912 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:12:20.242735 133792770201408 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579122.257079, 'wait_for_prev_duration_secs': 16.67928147315979, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776579138.9383008, 'checkpointer_blocking_duration_secs': 0.5805802345275879, 'get_old_steps_start_time': 1776579139.5189, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776579122.254098, 'checkpoint_manager_blocking_duration_secs': 17.98859977722168} I0419 06:12:20.242846 133792770201408 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 06:12:20.243750 133792770201408 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.160, Tokens/s/device: 13891.338, total_weights: 65536, loss: 5.858 I0419 06:12:20.251601 133792770201408 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.801, Tokens/s/device: 13837.183, total_weights: 65536, loss: 5.713 I0419 06:12:38.652228 133792770201408 metric_logger.py:185] completed step: 16, seconds: 18.285, TFLOP/s/device: 0.743, Tokens/s/device: 112.004, total_weights: 65536, loss: 5.599 I0419 06:12:38.799676 133792770201408 metric_logger.py:185] completed step: 17, seconds: 18.255, TFLOP/s/device: 0.744, Tokens/s/device: 112.188, total_weights: 65536, loss: 5.511 I0419 06:12:38.947215 133792770201408 metric_logger.py:185] completed step: 18, seconds: 0.154, TFLOP/s/device: 88.224, Tokens/s/device: 13298.010, total_weights: 65536, loss: 5.442 I0419 06:12:39.095992 133792770201408 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 06:12:39.096644 133792770201408 checkpointing.py:798] Waited 0.0006635189056396484 seconds for step 19 to finish before starting checkpointing. I0419 06:12:39.098551 133792770201408 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:12:42.957109 133666749720320 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6 I0419 06:12:48.361428 133664581621504 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 54.205 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 29.141162633895874 s) (per-host) I0419 06:12:48.361549 133664581621504 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.842824s. I0419 06:12:56.203504 133664581621504 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 36.684763s. I0419 06:12:56.203729 133666732934912 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:12:56.203802 133666732934912 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:12:56.203852 133666732934912 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:12:56.205694 133666732934912 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 06:12:56.205896 133792770201408 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. W0419 06:12:56.206020 133792770201408 checkpoint_manager.py:1458] Waiting for previous save to complete took 17.107486 seconds. If this number is high, consider checkpointing less frequently. I0419 06:12:56.207551 133792770201408 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0419 06:12:56.211046 133792770201408 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/19. I0419 06:12:57.626633 133792770201408 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:12:57.626731 133792770201408 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 I0419 06:12:57.670179 133792770201408 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045647s I0419 06:12:57.670342 133792770201408 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.088 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7387316226959229 s) (per-host) I0419 06:12:57.670395 133792770201408 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.738794s (batch_requests_ready=0.688552s, total_serialization_initiated=0.050175s, others=0.000067s) I0419 06:12:57.670485 133792770201408 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.742911s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.742900) I0419 06:12:57.671323 133664590014208 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:12:57.671476 133792770201408 async_checkpointer.py:561] Finished blocking save. Time taken: 1.463855s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/19. I0419 06:12:57.676054 133792770201408 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:12:57.676291 133666732934912 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:12:57.676409 133792770201408 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579159.0985172, 'wait_for_prev_duration_secs': 17.107486486434937, 'time_between_consecutive_saves_sec': 20.162445545196533, 'checkpointer_blocking_start_time': 1776579176.2075899, 'checkpointer_blocking_duration_secs': 1.4639954566955566, 'get_old_steps_start_time': 1776579177.671599, 'get_old_steps_duration_secs': 2.0503997802734375e-05, 'checkpoint_manager_blocking_start_time': 1776579159.0968657, 'checkpoint_manager_blocking_duration_secs': 18.579512357711792} I0419 06:12:57.676518 133792770201408 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 06:12:57.676560 133792770201408 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:13:02.862068 133666749720320 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6 I0419 06:13:07.794540 133664590014208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 145.412 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.862881422042847 s) (per-host) I0419 06:13:07.794656 133664590014208 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.123226s. I0419 06:13:12.980826 133664590014208 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 15.309380s. I0419 06:13:12.981049 133666732934912 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:13:12.981110 133666732934912 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:13:12.981154 133666732934912 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:13:12.982741 133666732934912 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 06:13:12.982928 133792770201408 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. I0419 06:13:12.983102 133792770201408 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:13:12.983839 133792770201408 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.562, Tokens/s/device: 13951.904, 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: Sun Apr 19 06:13:24 UTC 2026 EXIT_CODE=0