XPK Start: Sun Apr 19 09:23:03 UTC 2026 2026-04-19 09:23:07.368188: 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 00:00:1776590587.380673 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776590587.384388 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776590587.395824 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776590587.395841 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776590587.395843 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776590587.395845 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 09:23:26.672138: 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 09:23:27.192016 138243719374656 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 09:23:36,232:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 09:23:36.232855 138243719374656 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 09:23:36,235:jax._src.distributed:157: Connecting to JAX distributed service on mt-08-checkpoint-async--j48m6-slice-job-0-0.mt-08-checkpoint-async--j48m6:8482 I0419 09:23:36.235253 138243719374656 distributed.py:157] Connecting to JAX distributed service on mt-08-checkpoint-async--j48m6-slice-job-0-0.mt-08-checkpoint-async--j48m6:8482 I0419 09:23:37.288272 138243719374656 max_utils.py:284] Jax distributed system initialized! I0419 09:23:44.173602 138243719374656 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 09:23:44.173710 138243719374656 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 09:23:44.173752 138243719374656 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 09:23:44.173786 138243719374656 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 09:23:44.789891 138243719374656 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 09:23:44.790189 138243719374656 checkpointing.py:677] Setting up checkpoint logger... I0419 09:23:44.790246 138243719374656 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 09:23:44.790292 138243719374656 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 09:23:44.790643 138243719374656 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 0x7dbae874ad50>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 09:23:46.470751 138243719374656 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0419 09:23:46.471126 138243719374656 checkpoint_manager.py:709] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7da7e5355880>}, handler_registry=None I0419 09:23:46.471364 138243719374656 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7da7e5355880>` 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 09:23:46.471412 138243719374656 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7da7e4daf9e0>` 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 09:23:46.471448 138243719374656 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 0x7da7e5355880>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7da7e5355880>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7da7e4daf9e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7da7e4daf9e0>}). I0419 09:23:46.474113 138243719374656 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 09:23:46.474200 138243719374656 async_checkpointer.py:177] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dbad33f32e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 09:23:48.061025 138122918618880 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776590411636726312, 'commit_timestamp_nsecs': 1776590458771483027, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/5/_CHECKPOINT_METADATA I0419 09:23:48.313977 138122918618880 checkpoint_manager.py:1877] Missing metrics for step 5 I0419 09:23:48.745969 138243719374656 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints I0419 09:23:49.450001 138243719374656 checkpoint_manager.py:929] [process=4][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/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7da7e4daf1a0> I0419 09:23:49.450197 138243719374656 checkpointing.py:301] Checkpoint manager created! I0419 09:23:50.330229 138243719374656 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 09:23:50.330332 138243719374656 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 09:23:50.706761 138243719374656 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 09:23:50.706853 138243719374656 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 09:23:50.727647 138243719374656 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 09:23:50.727749 138243719374656 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 09:23:50.751484 138243719374656 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 09:23:50.751552 138243719374656 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 09:23:50.767566 138243719374656 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 09:23:50.767631 138243719374656 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 09:23:50.783574 138243719374656 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 09:23:50.783638 138243719374656 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 09:23:50.808274 138243719374656 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 09:23:50.808349 138243719374656 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 09:23:50.827825 138243719374656 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 09:23:50.827889 138243719374656 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 09:23:51.027784 138243719374656 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0419 09:23:51.027901 138243719374656 checkpointing.py:581] restoring from this run's directory step 9 I0419 09:23:51.532840 138243719374656 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/9. I0419 09:23:52.916052 1424 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 09:23:59.312052 138243719374656 jax_array_handlers.py:837] [process=4] /jax/orbax/read/worker/io/requested throughput: 246.875 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.401106119155884 s) (per-host) I0419 09:23:59.325809 138243719374656 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/read/gbytes_per_sec: 232.746 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.799663066864014 s) (per-host) I0419 09:24:00.018925 138243719374656 checkpointer.py:318] Finished restoring checkpoint in 8.74 seconds from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/9. I0419 09:24:00.019120 138243719374656 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776590631.277555, 'checkpointer_duration_secs': 8.741533279418945, 'checkpoint_manager_start_time': 1776590631.028162, 'checkpoint_manager_duration_secs': 8.990928649902344} fsdp: 32 I0419 09:24:00.423431 138243719374656 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0419 09:24:00.423566 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 09:24:00.423621 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 09:24:00.423678 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0419 09:24:00.423739 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0419 09:24:00.423781 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0419 09:24:00.423835 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 09:24:00.423888 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0419 09:24:00.423930 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0419 09:24:00.423982 138243719374656 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 09:24:00.424033 138243719374656 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0419 09:24:00.424080 138243719374656 maxtext_utils.py:1501] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0419 09:24:00.923414 138243719374656 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 09:24:00.923514 138243719374656 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 09:24:00.938495 138243719374656 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 09:24:00.938560 138243719374656 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 09:24:11.708683 138243719374656 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 09:24:11.709618 138243719374656 metric_logger.py:289] number parameters: 1.104 billion I0419 09:24:22.850650 138243719374656 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0419 09:24:23.003355 138243719374656 checkpointing.py:776] Waited 0.15268993377685547 seconds for step 10 to finish before starting checkpointing. I0419 09:24:23.005902 138243719374656 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 09:24:23.007657 138243719374656 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 10 I0419 09:24:23.010414 138243719374656 async_checkpointer.py:452] [process=4] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/10. I0419 09:24:23.648052 138243719374656 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 09:24:23.649052 138243719374656 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 09:24:23.649108 138243719374656 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 09:24:23.900794 138243719374656 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252849s I0419 09:24:23.900977 138243719374656 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.795 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5518383979797363 s) (per-host) I0419 09:24:23.901038 138243719374656 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.551909s (batch_requests_ready=0.297026s, total_serialization_initiated=0.254791s, others=0.000091s) I0419 09:24:23.901130 138243719374656 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.555815s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.555796) I0419 09:24:23.901979 138116920710912 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0419 09:24:23.902145 138243719374656 async_checkpointer.py:561] Finished blocking save. Time taken: 0.894419s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/10. I0419 09:24:24.315399 138243719374656 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 09:24:24.315790 138121249289984 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0419 09:24:24.315967 138243719374656 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776590663.0058827, 'wait_for_prev_duration_secs': 8.034706115722656e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776590663.007695, 'checkpointer_blocking_duration_secs': 0.8945558071136475, 'get_old_steps_start_time': 1776590663.902271, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776590663.0039535, 'checkpoint_manager_blocking_duration_secs': 1.311964988708496} I0419 09:24:24.316085 138243719374656 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0419 09:24:24.316139 138243719374656 max_utils.py:750] Memstats: After params initialized: I0419 09:24:24.316196 138243719374656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0419 09:24:24.316231 138243719374656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0419 09:24:24.316259 138243719374656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0419 09:24:24.316285 138243719374656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0419 09:24:24.584337 138243719374656 metric_logger.py:185] completed step: 10, seconds: 11.141, TFLOP/s/device: 1.220, Tokens/s/device: 183.827, total_weights: 65536, loss: 6.833 I0419 09:24:24.750277 138243719374656 metric_logger.py:185] completed step: 11, seconds: 1.732, TFLOP/s/device: 7.843, Tokens/s/device: 1182.193, total_weights: 65536, loss: 6.528 I0419 09:24:24.897657 138243719374656 metric_logger.py:185] completed step: 12, seconds: 0.019, TFLOP/s/device: 731.395, Tokens/s/device: 110243.850, total_weights: 65536, loss: 6.265 I0419 09:24:25.045242 138243719374656 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.726, Tokens/s/device: 13373.820, total_weights: 65536, loss: 6.045 I0419 09:24:25.049841 138243719374656 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0419 09:24:25.339971 138243719374656 checkpointing.py:776] Waited 0.2900662422180176 seconds for step 15 to finish before starting checkpointing. I0419 09:24:25.342866 138243719374656 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 09:24:28.386103 138116912318208 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/10/items/array_metadatas/process_4 I0419 09:25:05.312620 138116920710912 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 37.642 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.96344971656799 s) (per-host) I0419 09:25:05.312734 138116920710912 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.410644s. I0419 09:25:13.160382 138116920710912 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 49.258276s. I0419 09:25:13.160706 138121249289984 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 09:25:13.160837 138121249289984 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0419 09:25:13.160886 138121249289984 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 09:25:13.163216 138121249289984 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 09:25:13.163393 138243719374656 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0419 09:25:13.163528 138243719374656 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.820670 seconds. If this number is high, consider checkpointing less frequently. I0419 09:25:13.165432 138243719374656 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 15 I0419 09:25:13.169496 138243719374656 async_checkpointer.py:452] [process=4] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/15. I0419 09:25:14.152626 138243719374656 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 09:25:14.152729 138243719374656 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 09:25:14.196014 138243719374656 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044722s I0419 09:25:14.196183 138243719374656 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.117 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7286972999572754 s) (per-host) I0419 09:25:14.196233 138243719374656 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.728757s (batch_requests_ready=0.682060s, total_serialization_initiated=0.046630s, others=0.000068s) I0419 09:25:14.196320 138243719374656 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.734411s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=0.734397) I0419 09:25:14.197144 138121735804672 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0419 09:25:14.197305 138243719374656 async_checkpointer.py:561] Finished blocking save. Time taken: 1.031801s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/15. I0419 09:25:14.211456 138243719374656 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 09:25:14.211753 138121249289984 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0419 09:25:14.211921 138243719374656 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776590665.342836, 'wait_for_prev_duration_secs': 47.820669651031494, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776590713.1654701, 'checkpointer_blocking_duration_secs': 1.0319430828094482, 'get_old_steps_start_time': 1776590714.1974332, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776590665.340524, 'checkpoint_manager_blocking_duration_secs': 48.871360540390015} I0419 09:25:14.212039 138243719374656 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0419 09:25:14.212631 138243719374656 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.015, Tokens/s/device: 13869.513, total_weights: 65536, loss: 5.864 I0419 09:25:14.218420 138243719374656 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.221, Tokens/s/device: 13900.578, total_weights: 65536, loss: 5.719 I0419 09:25:32.046121 138243719374656 metric_logger.py:185] completed step: 16, seconds: 49.167, TFLOP/s/device: 0.276, Tokens/s/device: 41.654, total_weights: 65536, loss: 5.605 I0419 09:25:32.193661 138243719374656 metric_logger.py:185] completed step: 17, seconds: 17.682, TFLOP/s/device: 0.768, Tokens/s/device: 115.826, total_weights: 65536, loss: 5.517 I0419 09:25:32.341065 138243719374656 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.302, Tokens/s/device: 13460.578, total_weights: 65536, loss: 5.448 I0419 09:25:32.488122 138243719374656 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0419 09:25:32.488769 138243719374656 checkpointing.py:776] Waited 0.0006673336029052734 seconds for step 19 to finish before starting checkpointing. I0419 09:25:32.490750 138243719374656 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 09:25:36.247154 138122314639104 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/15/items/array_metadatas/process_4 I0419 09:26:13.661399 138121735804672 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 26.242 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.19387125968933 s) (per-host) I0419 09:26:13.661518 138121735804672 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.464262s. I0419 09:26:21.360173 138121735804672 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 67.162902s. I0419 09:26:21.360474 138121249289984 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 09:26:21.360605 138121249289984 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0419 09:26:21.360653 138121249289984 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 09:26:21.363168 138121249289984 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 09:26:21.363344 138243719374656 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0419 09:26:21.363470 138243719374656 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.872734 seconds. If this number is high, consider checkpointing less frequently. I0419 09:26:21.367280 138243719374656 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 19 I0419 09:26:21.370862 138243719374656 async_checkpointer.py:452] [process=4] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/19. I0419 09:26:21.908875 138243719374656 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 09:26:21.908989 138243719374656 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 09:26:21.952530 138243719374656 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044585s I0419 09:26:21.952690 138243719374656 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.069 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.304304838180542 s) (per-host) I0419 09:26:21.952742 138243719374656 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304364s (batch_requests_ready=0.257682s, total_serialization_initiated=0.046618s, others=0.000064s) I0419 09:26:21.952831 138243719374656 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308374s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.308362) I0419 09:26:21.953653 138122306246400 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0419 09:26:21.953811 138243719374656 async_checkpointer.py:561] Finished blocking save. Time taken: 0.586456s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/19. I0419 09:26:22.301339 138243719374656 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 09:26:22.301703 138121249289984 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0419 09:26:22.301874 138243719374656 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776590732.49072, 'wait_for_prev_duration_secs': 48.87273406982422, 'time_between_consecutive_saves_sec': 19.32746195793152, 'checkpointer_blocking_start_time': 1776590781.3673236, 'checkpointer_blocking_duration_secs': 0.5865957736968994, 'get_old_steps_start_time': 1776590781.9539323, 'get_old_steps_duration_secs': 4.363059997558594e-05, 'checkpoint_manager_blocking_start_time': 1776590732.4890106, 'checkpoint_manager_blocking_duration_secs': 49.812827825546265} I0419 09:26:22.302024 138243719374656 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0419 09:26:22.302081 138243719374656 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 09:26:27.174168 138122314639104 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_08_checkpoint_async_true/checkpoints/19/items/array_metadatas/process_4 I0419 09:27:03.447143 138122306246400 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 37.791 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.79871416091919 s) (per-host) I0419 09:27:03.447267 138122306246400 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.493502s. I0419 09:27:11.774212 138122306246400 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 49.820432s. I0419 09:27:11.774497 138121249289984 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 09:27:11.774617 138121249289984 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0419 09:27:11.774660 138121249289984 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 09:27:11.776458 138121249289984 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 09:27:11.776613 138243719374656 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0419 09:27:11.776752 138243719374656 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 09:27:11.777436 138243719374656 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.209, Tokens/s/device: 13898.692, total_weights: 65536, loss: 5.391 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 09:27:24 UTC 2026 EXIT_CODE=0