XPK Start: Sun Apr 19 10:20:03 UTC 2026
2026-04-19 10:20:08.062487: 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 00:00:1776594008.075439 11 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776594008.079237 11 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776594008.090755 11 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776594008.090773 11 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776594008.090775 11 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776594008.090777 11 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-19 10:20:27.444327: 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 10:20:27.973559 133107469997888 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-19 10:20:37,014:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0419 10:20:37.014760 133107469997888 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-19 10:20:37,017:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-y6c1i-slice-job-0-0.mt-16-zarr3-true-rs-y6c1i:8482
I0419 10:20:37.017070 133107469997888 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-y6c1i-slice-job-0-0.mt-16-zarr3-true-rs-y6c1i:8482
I0419 10:20:38.122697 133107469997888 max_utils.py:284] Jax distributed system initialized!
I0419 10:20:43.961999 133107469997888 max_utils.py:800] System Information: Jax Version: 0.8.1
I0419 10:20:43.962106 133107469997888 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0419 10:20:43.962147 133107469997888 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 10:20:43.962186 133107469997888 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 10:20:44.583702 133107469997888 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 10:20:44.584023 133107469997888 checkpointing.py:677] Setting up checkpoint logger...
I0419 10:20:44.584084 133107469997888 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 10:20:44.584130 133107469997888 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0419 10:20:44.584490 133107469997888 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 0x790f07978f20>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 10:20:45.833576 133107469997888 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 10:20:45.833815 133107469997888 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 0x78fcc2936720>}, handler_registry=None
I0419 10:20:45.834063 133107469997888 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78fcc2936720>` 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 10:20:45.834112 133107469997888 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78fcc19837d0>` 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 10:20:45.834150 133107469997888 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 0x78fcc2936720>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78fcc2936720>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78fcc19837d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78fcc19837d0>}).
I0419 10:20:45.836804 133107469997888 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0419 10:20:45.836888 133107469997888 async_checkpointer.py:177] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x790eedde32e0> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0419 10:20:48.162191 132989954545408 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776593830291599305, 'commit_timestamp_nsecs': 1776593876673103544, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_true/checkpoints/5/_CHECKPOINT_METADATA
I0419 10:20:48.409631 132989946152704 checkpoint_manager.py:1877] Missing metrics for step 9
I0419 10:20:48.827011 133107469997888 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_16_zarr3_true/checkpoints
I0419 10:20:49.265179 133107469997888 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_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78fcc19805f0>
I0419 10:20:49.265359 133107469997888 checkpointing.py:301] Checkpoint manager created!
I0419 10:20:50.150181 133107469997888 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 10:20:50.150291 133107469997888 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 10:20:50.520843 133107469997888 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 10:20:50.520950 133107469997888 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 10:20:50.536623 133107469997888 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 10:20:50.536679 133107469997888 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 10:20:50.560053 133107469997888 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 10:20:50.560122 133107469997888 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:20:50.576009 133107469997888 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 10:20:50.576073 133107469997888 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:20:50.591818 133107469997888 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 10:20:50.591889 133107469997888 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:20:50.615609 133107469997888 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 10:20:50.615678 133107469997888 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:20:50.634989 133107469997888 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 10:20:50.635052 133107469997888 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 10:20:50.824857 133107469997888 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 10:20:50.824986 133107469997888 checkpointing.py:581] restoring from this run's directory step 9
I0419 10:20:51.317282 133107469997888 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_16_zarr3_true/checkpoints/9.
I0419 10:20:53.145535 1368 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 10:20:59.674518 133107469997888 jax_array_handlers.py:837] [process=4] /jax/orbax/read/worker/io/requested throughput: 241.850 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.53409481048584 s) (per-host)
I0419 10:20:59.686059 133107469997888 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/read/gbytes_per_sec: 227.814 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.9468748569488525 s) (per-host)
I0419 10:20:59.865722 133107469997888 checkpointer.py:318] Finished restoring checkpoint in 8.80 seconds from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_true/checkpoints/9.
I0419 10:20:59.865882 133107469997888 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_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776594051.0685425, 'checkpointer_duration_secs': 8.797306299209595, 'checkpoint_manager_start_time': 1776594050.8252308, 'checkpoint_manager_duration_secs': 9.040620803833008}
fsdp: 32
I0419 10:21:00.261850 133107469997888 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: PartitionSpec('norm',)
Physical: (None,)
I0419 10:21:00.262003 133107469997888 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 10:21:00.262062 133107469997888 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 10:21:00.262123 133107469997888 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 10:21:00.262177 133107469997888 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 10:21:00.262217 133107469997888 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 10:21:00.262272 133107469997888 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 10:21:00.262327 133107469997888 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 10:21:00.262369 133107469997888 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 10:21:00.262407 133107469997888 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 10:21:00.262456 133107469997888 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: PartitionSpec('embed', 'vocab')
Physical: ('fsdp', None)
I0419 10:21:00.262503 133107469997888 maxtext_utils.py:1501] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: PartitionSpec('vocab', 'embed')
Physical: (None, 'fsdp')
I0419 10:21:00.759437 133107469997888 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 10:21:00.759530 133107469997888 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 10:21:00.774388 133107469997888 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 10:21:00.774445 133107469997888 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 10:21:11.518138 133107469997888 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 10:21:11.518986 133107469997888 metric_logger.py:289] number parameters: 1.104 billion
I0419 10:21:22.542412 133107469997888 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0419 10:21:22.753637 133107469997888 checkpointing.py:776] Waited 0.21120476722717285 seconds for step 10 to finish before starting checkpointing.
I0419 10:21:22.756483 133107469997888 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 10:21:22.758511 133107469997888 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 10
I0419 10:21:22.761674 133107469997888 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_16_zarr3_true/checkpoints/10.
I0419 10:21:23.767624 133107469997888 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 10:21:23.768672 133107469997888 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0419 10:21:23.768735 133107469997888 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 10:21:24.022372 133107469997888 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254839s
I0419 10:21:24.022548 133107469997888 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.020 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5107274055480957 s) (per-host)
I0419 10:21:24.022600 133107469997888 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.510791s (batch_requests_ready=0.253870s, total_serialization_initiated=0.256848s, others=0.000072s)
I0419 10:21:24.022691 133107469997888 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.514841s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.514822)
I0419 10:21:24.023624 132989166024448 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started.
I0419 10:21:24.023803 133107469997888 async_checkpointer.py:561] Finished blocking save. Time taken: 1.265218s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_true/checkpoints/10.
I0419 10:21:24.056876 133107469997888 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 10:21:24.057212 132984770377472 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 10:21:24.057373 133107469997888 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776594082.7564647, 'wait_for_prev_duration_secs': 6.29425048828125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776594082.7585514, 'checkpointer_blocking_duration_secs': 1.2653608322143555, 'get_old_steps_start_time': 1776594084.0239305, 'get_old_steps_duration_secs': 4.315376281738281e-05, 'checkpoint_manager_blocking_start_time': 1776594082.754238, 'checkpoint_manager_blocking_duration_secs': 1.303098201751709}
I0419 10:21:24.057475 133107469997888 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0419 10:21:24.057527 133107469997888 max_utils.py:750]
Memstats: After params initialized:
I0419 10:21:24.057579 133107469997888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0))
I0419 10:21:24.057611 133107469997888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0))
I0419 10:21:24.057638 133107469997888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0))
I0419 10:21:24.057662 133107469997888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0))
I0419 10:21:24.318829 133107469997888 metric_logger.py:185] completed step: 10, seconds: 11.023, TFLOP/s/device: 1.233, Tokens/s/device: 185.788, total_weights: 65536, loss: 6.833
I0419 10:21:24.491705 133107469997888 metric_logger.py:185] completed step: 11, seconds: 1.775, TFLOP/s/device: 7.654, Tokens/s/device: 1153.765, total_weights: 65536, loss: 6.528
I0419 10:21:24.639526 133107469997888 metric_logger.py:185] completed step: 12, seconds: 0.026, TFLOP/s/device: 528.209, Tokens/s/device: 79617.463, total_weights: 65536, loss: 6.265
I0419 10:21:24.786862 133107469997888 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.748, Tokens/s/device: 13377.140, total_weights: 65536, loss: 6.045
I0419 10:21:24.791519 133107469997888 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0419 10:21:25.081544 133107469997888 checkpointing.py:776] Waited 0.2899894714355469 seconds for step 15 to finish before starting checkpointing.
I0419 10:21:25.084345 133107469997888 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 10:21:28.477991 132984787162880 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_16_zarr3_true/checkpoints/10/items/array_metadatas/process_4
I0419 10:22:05.709612 132989166024448 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 37.433 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.197754859924316 s) (per-host)
I0419 10:22:05.709732 132989166024448 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.685990s.
I0419 10:22:14.006231 132989166024448 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 49.982470s.
I0419 10:22:14.006517 132984770377472 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 10:22:14.006637 132984770377472 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 10:22:14.006686 132984770377472 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0419 10:22:14.007955 132984770377472 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0419 10:22:14.008132 133107469997888 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 10:22:14.008269 133107469997888 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.923928 seconds. If this number is high, consider checkpointing less frequently.
I0419 10:22:14.010308 133107469997888 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 15
I0419 10:22:14.014769 133107469997888 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_16_zarr3_true/checkpoints/15.
I0419 10:22:14.979648 133107469997888 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0419 10:22:14.979748 133107469997888 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 10:22:15.029842 133107469997888 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051498s
I0419 10:22:15.030027 133107469997888 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.979 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3098008632659912 s) (per-host)
I0419 10:22:15.030084 133107469997888 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.309868s (batch_requests_ready=0.256294s, total_serialization_initiated=0.053499s, others=0.000075s)
I0419 10:22:15.030181 133107469997888 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.314803s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.314790)
I0419 10:22:15.031053 132983176550144 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started.
I0419 10:22:15.031218 133107469997888 async_checkpointer.py:561] Finished blocking save. Time taken: 1.020834s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_true/checkpoints/15.
I0419 10:22:15.033260 133107469997888 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 10:22:15.033610 132984770377472 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 10:22:15.033776 133107469997888 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776594085.0843167, 'wait_for_prev_duration_secs': 48.923927783966064, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776594134.0103478, 'checkpointer_blocking_duration_secs': 1.0209815502166748, 'get_old_steps_start_time': 1776594135.0313492, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776594085.0822434, 'checkpoint_manager_blocking_duration_secs': 49.95149564743042}
I0419 10:22:15.033897 133107469997888 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0419 10:22:15.034489 133107469997888 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.858, Tokens/s/device: 13845.883, total_weights: 65536, loss: 5.864
I0419 10:22:15.040083 133107469997888 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.301, Tokens/s/device: 13912.571, total_weights: 65536, loss: 5.719
I0419 10:22:33.010371 133107469997888 metric_logger.py:185] completed step: 16, seconds: 50.247, TFLOP/s/device: 0.270, Tokens/s/device: 40.758, total_weights: 65536, loss: 5.605
I0419 10:22:33.157671 133107469997888 metric_logger.py:185] completed step: 17, seconds: 17.824, TFLOP/s/device: 0.762, Tokens/s/device: 114.903, total_weights: 65536, loss: 5.517
I0419 10:22:33.305204 133107469997888 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.341, Tokens/s/device: 13466.419, total_weights: 65536, loss: 5.448
I0419 10:22:33.452276 133107469997888 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0419 10:22:33.452976 133107469997888 checkpointing.py:776] Waited 0.0006875991821289062 seconds for step 19 to finish before starting checkpointing.
I0419 10:22:33.455056 133107469997888 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 10:22:37.453081 132984787162880 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_16_zarr3_true/checkpoints/15/items/array_metadatas/process_4
I0419 10:23:14.707829 132983176550144 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 26.332 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.98757195472717 s) (per-host)
I0419 10:23:14.707932 132983176550144 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.676766s.
I0419 10:23:24.329556 132983176550144 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 69.298373s.
I0419 10:23:24.329845 132984770377472 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 10:23:24.329990 132984770377472 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 10:23:24.330056 132984770377472 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0419 10:23:24.331640 132984770377472 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0419 10:23:24.331825 133107469997888 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 10:23:24.332009 133107469997888 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.876966 seconds. If this number is high, consider checkpointing less frequently.
I0419 10:23:24.333585 133107469997888 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 19
I0419 10:23:24.337337 133107469997888 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_16_zarr3_true/checkpoints/19.
I0419 10:23:24.892756 133107469997888 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0419 10:23:24.892853 133107469997888 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 10:23:24.938053 133107469997888 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046234s
I0419 10:23:24.938210 133107469997888 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.070 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30426716804504395 s) (per-host)
I0419 10:23:24.938260 133107469997888 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304326s (batch_requests_ready=0.256032s, total_serialization_initiated=0.048231s, others=0.000063s)
I0419 10:23:24.938342 133107469997888 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308499s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.308487)
I0419 10:23:24.939217 132989157631744 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started.
I0419 10:23:24.939373 133107469997888 async_checkpointer.py:561] Finished blocking save. Time taken: 0.605721s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_16_zarr3_true/checkpoints/19.
I0419 10:23:25.375648 133107469997888 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 10:23:25.376028 132984770377472 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 10:23:25.376172 133107469997888 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776594153.4550269, 'wait_for_prev_duration_secs': 50.87696552276611, 'time_between_consecutive_saves_sec': 19.447033882141113, 'checkpointer_blocking_start_time': 1776594204.3336246, 'checkpointer_blocking_duration_secs': 0.6058511734008789, 'get_old_steps_start_time': 1776594204.9394898, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1776594153.4532018, 'checkpoint_manager_blocking_duration_secs': 51.92293882369995}
I0419 10:23:25.376357 133107469997888 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0419 10:23:25.376411 133107469997888 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 10:23:29.574955 132989166024448 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_16_zarr3_true/checkpoints/19/items/array_metadatas/process_4
I0419 10:24:06.497134 132989157631744 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 37.732 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.8631489276886 s) (per-host)
I0419 10:24:06.497253 132989157631744 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.557925s.
I0419 10:24:14.516244 132989157631744 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 49.576900s.
I0419 10:24:14.516582 132984770377472 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 10:24:14.516644 132984770377472 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 10:24:14.516683 132984770377472 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0419 10:24:14.518242 132984770377472 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0419 10:24:14.518398 133107469997888 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 10:24:14.518547 133107469997888 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 10:24:14.519211 133107469997888 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.224, Tokens/s/device: 13900.956, 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 10:24:25 UTC 2026
EXIT_CODE=0