XPK Start: Sat Apr 25 17:30:37 UTC 2026 2026-04-25 17:30:41.413328: 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 00:00:1777138241.426161 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777138241.429895 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777138241.441602 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138241.441621 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138241.441623 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138241.441625 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:31:00.777867: 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) I0425 17:31:01.313194 138490824501056 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 17:31:10,353:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 17:31:10.353269 138490824501056 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 17:31:10,355:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-95uq0-slice-job-0-0.mt-15-ocdbt-true-rs-95uq0:8482 I0425 17:31:10.355711 138490824501056 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-95uq0-slice-job-0-0.mt-15-ocdbt-true-rs-95uq0:8482 I0425 17:31:11.309629 138490824501056 max_utils.py:284] Jax distributed system initialized! I0425 17:31:18.334602 138490824501056 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 17:31:18.334709 138490824501056 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 17:31:18.334749 138490824501056 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 I0425 17:31:18.334783 138490824501056 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 17:31:18.953725 138490824501056 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 17:31:19.068392 138490824501056 checkpointing.py:688] Setting up checkpoint logger... I0425 17:31:19.068510 138490824501056 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 17:31:19.068555 138490824501056 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 17:31:19.068759 138490824501056 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 0x7df46b971370>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 17:31:20.356790 138490824501056 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 17:31:20.357134 138490824501056 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 0x7df3230906e0>}, handler_registry=None I0425 17:31:20.357377 138490824501056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7df3230906e0>` 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`. I0425 17:31:20.357428 138490824501056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ddf80188800>` 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`. I0425 17:31:20.357464 138490824501056 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 0x7df3230906e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7df3230906e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ddf80188800>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ddf80188800>}). I0425 17:31:20.357807 138490824501056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 17:31:20.357884 138490824501056 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ddf8015cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 17:31:22.047266 138370978125568 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777138113907229295, 'commit_timestamp_nsecs': 1777138160773448503, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9/_CHECKPOINT_METADATA I0425 17:31:22.348864 138370994910976 checkpoint_manager.py:1877] Missing metrics for step 0 I0425 17:31:22.373841 138490824501056 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints I0425 17:31:23.026003 138490824501056 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ddff013ae70> I0425 17:31:23.026199 138490824501056 checkpointing.py:302] Checkpoint manager created! I0425 17:31:23.964378 138490824501056 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 17:31:23.964498 138490824501056 checkpointing.py:582] restoring from this run's directory step 9 I0425 17:31:24.463085 138490824501056 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9. I0425 17:31:25.985854 1411 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 17:31:32.275151 138490824501056 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 251.038 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.294969797134399 s) (per-host) I0425 17:31:32.283438 138490824501056 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 235.628 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.716552019119263 s) (per-host) I0425 17:31:34.376154 138490824501056 checkpointer.py:318] Finished restoring checkpoint in 10.16 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9. I0425 17:31:34.376321 138490824501056 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1777138284.2120643, 'checkpointer_duration_secs': 10.164225339889526, 'checkpoint_manager_start_time': 1777138283.9651139, 'checkpoint_manager_duration_secs': 10.411177635192871} fsdp: 32 I0425 17:31:34.456877 138490824501056 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 17:31:34.456994 138490824501056 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:31:34.457046 138490824501056 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:31:34.457089 138490824501056 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 17:31:34.457142 138490824501056 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:31:34.457178 138490824501056 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:31:34.457215 138490824501056 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:31:34.457253 138490824501056 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 17:31:34.457296 138490824501056 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:31:34.457331 138490824501056 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:31:34.457365 138490824501056 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 17:31:34.457398 138490824501056 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 17:31:34.856088 138490824501056 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:31:34.856260 138490824501056 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:31:34.861759 138490824501056 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:31:34.861856 138490824501056 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:31:35.016850 138490824501056 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:31:35.016984 138490824501056 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:31:35.169511 138490824501056 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:31:35.169647 138490824501056 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:31:35.332062 138490824501056 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:31:35.332221 138490824501056 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:31:35.484330 138490824501056 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:31:35.484459 138490824501056 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:31:35.634931 138490824501056 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:31:35.635061 138490824501056 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:31:35.797193 138490824501056 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 17:31:35.797330 138490824501056 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:31:35.820917 138490824501056 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 17:31:35.821011 138490824501056 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 17:31:46.932103 138490824501056 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. I0425 17:31:46.934530 138490824501056 metric_logger.py:289] number parameters: 1.104 billion I0425 17:31:58.479140 138490824501056 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0425 17:31:58.630072 138490824501056 checkpointing.py:798] Waited 0.15092062950134277 seconds for step 10 to finish before starting checkpointing. I0425 17:31:58.632399 138490824501056 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:31:58.634349 138490824501056 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0425 17:31:58.636936 138490824501056 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/10. I0425 17:31:59.240537 138490824501056 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 17:31:59.241703 138490824501056 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:31:59.241759 138490824501056 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 I0425 17:31:59.495004 138490824501056 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.255590s I0425 17:31:59.495194 138490824501056 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.891 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5335173606872559 s) (per-host) I0425 17:31:59.495249 138490824501056 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.533584s (batch_requests_ready=0.273269s, total_serialization_initiated=0.260240s, others=0.000075s) I0425 17:31:59.495340 138490824501056 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.537650s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.537632) I0425 17:31:59.496256 138369443014400 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:31:59.496432 138490824501056 async_checkpointer.py:561] Finished blocking save. Time taken: 0.862015s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/10. I0425 17:31:59.771578 138490824501056 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:31:59.771995 138365726881536 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:31:59.772186 138490824501056 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138318.6323793, 'wait_for_prev_duration_secs': 6.651878356933594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138318.6343846, 'checkpointer_blocking_duration_secs': 0.862156867980957, 'get_old_steps_start_time': 1777138319.4965599, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1777138318.6304712, 'checkpoint_manager_blocking_duration_secs': 1.1416759490966797} I0425 17:31:59.772304 138490824501056 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0425 17:31:59.772381 138490824501056 max_utils.py:750] Memstats: After params initialized: I0425 17:31:59.772438 138490824501056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 17:31:59.772472 138490824501056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 17:31:59.772499 138490824501056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 17:31:59.772523 138490824501056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 17:32:00.047350 138490824501056 metric_logger.py:185] completed step: 10, seconds: 11.544, TFLOP/s/device: 1.177, Tokens/s/device: 177.415, total_weights: 65536, loss: 6.827 I0425 17:32:00.228600 138490824501056 metric_logger.py:185] completed step: 11, seconds: 1.567, TFLOP/s/device: 8.674, Tokens/s/device: 1307.365, total_weights: 65536, loss: 6.522 I0425 17:32:00.376191 138490824501056 metric_logger.py:185] completed step: 12, seconds: 0.026, TFLOP/s/device: 516.758, Tokens/s/device: 77891.454, total_weights: 65536, loss: 6.260 I0425 17:32:00.523540 138490824501056 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.412, Tokens/s/device: 12572.748, total_weights: 65536, loss: 6.039 I0425 17:32:00.530915 138490824501056 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0425 17:32:00.818456 138490824501056 checkpointing.py:798] Waited 0.28751587867736816 seconds for step 15 to finish before starting checkpointing. I0425 17:32:00.821127 138490824501056 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:32:04.949326 138366255359744 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6 I0425 17:32:41.595009 138369443014400 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.051 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.63329553604126 s) (per-host) I0425 17:32:41.595149 138369443014400 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.098778s. I0425 17:32:50.485733 138369443014400 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.989347s. I0425 17:32:50.486011 138365726881536 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:32:50.486162 138365726881536 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:32:50.486209 138365726881536 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:32:50.487670 138365726881536 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 17:32:50.487839 138490824501056 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. W0425 17:32:50.487974 138490824501056 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.666878 seconds. If this number is high, consider checkpointing less frequently. I0425 17:32:50.489828 138490824501056 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0425 17:32:50.493805 138490824501056 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/15. I0425 17:32:51.040163 138490824501056 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:32:51.040261 138490824501056 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 I0425 17:32:51.080469 138490824501056 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042230s I0425 17:32:51.080617 138490824501056 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.119 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30132269859313965 s) (per-host) I0425 17:32:51.080664 138490824501056 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.301379s (batch_requests_ready=0.254593s, total_serialization_initiated=0.046722s, others=0.000064s) I0425 17:32:51.080753 138490824501056 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.306041s (all_items=0.000014s, per_item={'items': '0.00001359'}, temp_paths=0.306027) I0425 17:32:51.081619 138367832401664 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:32:51.081775 138490824501056 async_checkpointer.py:561] Finished blocking save. Time taken: 0.591876s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/15. I0425 17:32:51.098591 138490824501056 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:32:51.098876 138365726881536 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:32:51.099039 138490824501056 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138320.821073, 'wait_for_prev_duration_secs': 49.66687774658203, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138370.4898686, 'checkpointer_blocking_duration_secs': 0.5920116901397705, 'get_old_steps_start_time': 1777138371.0818992, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1777138320.8188283, 'checkpoint_manager_blocking_duration_secs': 50.28017807006836} I0425 17:32:51.099160 138490824501056 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0425 17:32:51.099967 138490824501056 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.370, Tokens/s/device: 13923.070, total_weights: 65536, loss: 5.858 I0425 17:32:51.107535 138490824501056 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.352, Tokens/s/device: 13769.641, total_weights: 65536, loss: 5.713 I0425 17:33:09.729505 138490824501056 metric_logger.py:185] completed step: 16, seconds: 50.575, TFLOP/s/device: 0.269, Tokens/s/device: 40.494, total_weights: 65536, loss: 5.599 I0425 17:33:09.876965 138490824501056 metric_logger.py:185] completed step: 17, seconds: 18.476, TFLOP/s/device: 0.735, Tokens/s/device: 110.844, total_weights: 65536, loss: 5.511 I0425 17:33:10.024422 138490824501056 metric_logger.py:185] completed step: 18, seconds: 0.154, TFLOP/s/device: 88.367, Tokens/s/device: 13319.632, total_weights: 65536, loss: 5.442 I0425 17:33:10.172803 138490824501056 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0425 17:33:10.173471 138490824501056 checkpointing.py:798] Waited 0.0006799697875976562 seconds for step 19 to finish before starting checkpointing. I0425 17:33:10.175459 138490824501056 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:33:14.741669 138366255359744 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6 I0425 17:33:54.829913 138367832401664 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 24.662 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 64.05056476593018 s) (per-host) I0425 17:33:54.830033 138367832401664 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 63.748305s. I0425 17:34:04.709292 138367832401664 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 73.627548s. I0425 17:34:04.709587 138365726881536 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:34:04.709713 138365726881536 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:34:04.709762 138365726881536 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:34:04.712206 138365726881536 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 17:34:04.712426 138490824501056 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. W0425 17:34:04.712548 138490824501056 checkpoint_manager.py:1458] Waiting for previous save to complete took 54.537104 seconds. If this number is high, consider checkpointing less frequently. I0425 17:34:04.714247 138490824501056 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0425 17:34:04.717944 138490824501056 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/19. I0425 17:34:05.547263 138490824501056 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:34:05.547365 138490824501056 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 I0425 17:34:05.592556 138490824501056 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047637s I0425 17:34:05.592726 138490824501056 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.922 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5279219150543213 s) (per-host) I0425 17:34:05.592775 138490824501056 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.527980s (batch_requests_ready=0.475800s, total_serialization_initiated=0.052117s, others=0.000063s) I0425 17:34:05.592863 138490824501056 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.531946s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.531934) I0425 17:34:05.594533 138367824008960 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:34:05.594696 138490824501056 async_checkpointer.py:561] Finished blocking save. Time taken: 0.880378s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/19. I0425 17:34:05.596626 138490824501056 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:34:05.596830 138365726881536 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:34:05.596949 138490824501056 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138390.1754282, 'wait_for_prev_duration_secs': 54.53710412979126, 'time_between_consecutive_saves_sec': 19.687719583511353, 'checkpointer_blocking_start_time': 1777138444.7142863, 'checkpointer_blocking_duration_secs': 0.8805234432220459, 'get_old_steps_start_time': 1777138445.594824, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1777138390.1736996, 'checkpoint_manager_blocking_duration_secs': 55.42321968078613} I0425 17:34:05.597051 138490824501056 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0425 17:34:05.597091 138490824501056 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:34:10.685666 138369443014400 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6 I0425 17:34:46.328855 138367824008960 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.280 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.26400923728943 s) (per-host) I0425 17:34:46.328969 138367824008960 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.734328s. I0425 17:34:54.513692 138367824008960 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.919037s. I0425 17:34:54.513970 138365726881536 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:34:54.514091 138365726881536 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:34:54.514159 138365726881536 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:34:54.515638 138365726881536 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 17:34:54.515808 138490824501056 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. I0425 17:34:54.515986 138490824501056 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:34:54.516686 138490824501056 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.548, Tokens/s/device: 13949.813, 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: Sat Apr 25 17:35:07 UTC 2026 EXIT_CODE=0