XPK Start: Sun Apr 19 10:11:06 UTC 2026
2026-04-19 10:11:10.920299: 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 00:00:1776593470.932961 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776593470.936790 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776593470.948026 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776593470.948044 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776593470.948047 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776593470.948049 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-19 10:11:30.266620: 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 10:11:30.790947 134447627163456 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-19 10:11:39,831:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0419 10:11:39.831392 134447627163456 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-19 10:11:39,836:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-rs-ecj9g-slice-job-0-0.mt-15-ocdbt-false-rs-ecj9g:8482
I0419 10:11:39.836179 134447627163456 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-ecj9g-slice-job-0-0.mt-15-ocdbt-false-rs-ecj9g:8482
I0419 10:11:40.558874 134447627163456 max_utils.py:284] Jax distributed system initialized!
I0419 10:11:47.563616 134447627163456 max_utils.py:800] System Information: Jax Version: 0.8.1
I0419 10:11:47.563725 134447627163456 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0419 10:11:47.563768 134447627163456 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 10:11:47.563801 134447627163456 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 10:11:48.188023 134447627163456 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 10:11:48.188318 134447627163456 checkpointing.py:677] Setting up checkpoint logger...
I0419 10:11:48.188375 134447627163456 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True
I0419 10:11:48.188431 134447627163456 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0419 10:11:48.188812 134447627163456 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7a4708379010>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 10:11:49.556190 134447627163456 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 10:11:49.556440 134447627163456 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a32dc229bb0>}, handler_registry=None
I0419 10:11:49.556691 134447627163456 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a32dc229bb0>` 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 10:11:49.556743 134447627163456 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a32dc0cb0b0>` 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 10:11:49.556781 134447627163456 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 0x7a32dc229bb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a32dc229bb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a32dc0cb0b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a32dc0cb0b0>}).
I0419 10:11:49.559406 134447627163456 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0419 10:11:49.559488 134447627163456 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a46f5fe32e0> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0419 10:11:52.205429 134327669069568 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776593319710690256, 'commit_timestamp_nsecs': 1776593334069495851, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0419 10:11:52.458118 134327115446016 checkpoint_manager.py:1877] Missing metrics for step 5
I0419 10:11:52.941736 134447627163456 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_15_ocdbt_false/checkpoints
I0419 10:11:53.091417 134447627163456 checkpoint_manager.py:929] [process=7][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_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a32dc22ac60>
I0419 10:11:53.091629 134447627163456 checkpointing.py:301] Checkpoint manager created!
I0419 10:11:53.981360 134447627163456 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 10:11:53.981475 134447627163456 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 10:11:54.358968 134447627163456 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 10:11:54.359066 134447627163456 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 10:11:54.375144 134447627163456 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 10:11:54.375204 134447627163456 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 10:11:54.399142 134447627163456 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 10:11:54.399216 134447627163456 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:11:54.415416 134447627163456 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 10:11:54.415479 134447627163456 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:11:54.431675 134447627163456 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 10:11:54.431737 134447627163456 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:11:54.455929 134447627163456 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 10:11:54.456002 134447627163456 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 10:11:54.481038 134447627163456 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 10:11:54.481144 134447627163456 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 10:11:54.681427 134447627163456 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 10:11:54.681553 134447627163456 checkpointing.py:581] restoring from this run's directory step 9
I0419 10:11:55.176078 134447627163456 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_15_ocdbt_false/checkpoints/9.
I0419 10:11:57.943118 1446 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 10:12:07.886295 134447627163456 jax_array_handlers.py:837] [process=7] /jax/orbax/read/worker/io/requested throughput: 141.351 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.179749250411987 s) (per-host)
I0419 10:12:07.894105 134447627163456 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/read/gbytes_per_sec: 136.624 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.583555459976196 s) (per-host)
I0419 10:12:08.500971 134447627163456 checkpointer.py:318] Finished restoring checkpoint in 13.57 seconds from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_15_ocdbt_false/checkpoints/9.
I0419 10:12:08.501133 134447627163456 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_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776593514.9281614, 'checkpointer_duration_secs': 13.572938442230225, 'checkpoint_manager_start_time': 1776593514.6818016, 'checkpoint_manager_duration_secs': 13.819300413131714}
fsdp: 32
I0419 10:12:08.907247 134447627163456 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: PartitionSpec('norm',)
Physical: (None,)
I0419 10:12:08.907375 134447627163456 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 10:12:08.907430 134447627163456 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 10:12:08.907489 134447627163456 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 10:12:08.907561 134447627163456 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 10:12:08.907605 134447627163456 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 10:12:08.907671 134447627163456 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 10:12:08.907813 134447627163456 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 10:12:08.907883 134447627163456 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 10:12:08.907930 134447627163456 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 10:12:08.907988 134447627163456 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: PartitionSpec('embed', 'vocab')
Physical: ('fsdp', None)
I0419 10:12:08.908041 134447627163456 maxtext_utils.py:1501] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: PartitionSpec('vocab', 'embed')
Physical: (None, 'fsdp')
I0419 10:12:09.414342 134447627163456 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 10:12:09.414444 134447627163456 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 10:12:09.429547 134447627163456 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 10:12:09.429612 134447627163456 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 10:12:20.177965 134447627163456 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 10:12:20.178913 134447627163456 metric_logger.py:289] number parameters: 1.104 billion
I0419 10:12:31.164430 134447627163456 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0419 10:12:31.701626 134447627163456 checkpointing.py:776] Waited 0.5371809005737305 seconds for step 10 to finish before starting checkpointing.
I0419 10:12:31.704321 134447627163456 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 10:12:31.706099 134447627163456 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 10
I0419 10:12:31.708741 134447627163456 async_checkpointer.py:452] [process=7] 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_15_ocdbt_false/checkpoints/10.
I0419 10:12:32.744842 134447627163456 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 10:12:32.745835 134447627163456 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0419 10:12:32.745893 134447627163456 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 10:12:32.996773 134447627163456 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252025s
I0419 10:12:32.996941 134447627163456 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.871 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5372974872589111 s) (per-host)
I0419 10:12:32.996993 134447627163456 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.537360s (batch_requests_ready=0.282809s, total_serialization_initiated=0.254481s, others=0.000070s)
I0419 10:12:32.997083 134447627163456 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.541514s (all_items=0.000040s, per_item={'items': '0.00004029'}, temp_paths=0.541474)
I0419 10:12:32.998018 134324942784256 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started.
I0419 10:12:32.998194 134447627163456 async_checkpointer.py:561] Finished blocking save. Time taken: 1.292025s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_15_ocdbt_false/checkpoints/10.
I0419 10:12:33.414568 134447627163456 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 10:12:33.414970 134322283607808 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0419 10:12:33.415135 134447627163456 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776593551.7043045, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776593551.7061386, 'checkpointer_blocking_duration_secs': 1.292158842086792, 'get_old_steps_start_time': 1776593552.998314, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776593551.7022784, 'checkpoint_manager_blocking_duration_secs': 1.7128188610076904}
I0419 10:12:33.415245 134447627163456 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0419 10:12:33.415300 134447627163456 max_utils.py:750]
Memstats: After params initialized:
I0419 10:12:33.415357 134447627163456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0))
I0419 10:12:33.415392 134447627163456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0))
I0419 10:12:33.415421 134447627163456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0))
I0419 10:12:33.415446 134447627163456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0))
I0419 10:12:33.677511 134447627163456 metric_logger.py:185] completed step: 10, seconds: 10.985, TFLOP/s/device: 1.237, Tokens/s/device: 186.429, total_weights: 65536, loss: 6.833
I0419 10:12:33.863341 134447627163456 metric_logger.py:185] completed step: 11, seconds: 2.512, TFLOP/s/device: 5.410, Tokens/s/device: 815.382, total_weights: 65536, loss: 6.528
I0419 10:12:34.010769 134447627163456 metric_logger.py:185] completed step: 12, seconds: 0.038, TFLOP/s/device: 354.709, Tokens/s/device: 53465.605, total_weights: 65536, loss: 6.265
I0419 10:12:34.158215 134447627163456 metric_logger.py:185] completed step: 13, seconds: 0.154, TFLOP/s/device: 88.254, Tokens/s/device: 13302.588, total_weights: 65536, loss: 6.045
I0419 10:12:34.163270 134447627163456 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0419 10:12:34.453043 134447627163456 checkpointing.py:776] Waited 0.28973841667175293 seconds for step 15 to finish before starting checkpointing.
I0419 10:12:34.455957 134447627163456 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 10:12:38.158086 134323332171520 array_metadata_store.py:203] [process=7][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_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_7
I0419 10:12:43.629025 134324942784256 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 141.423 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.169343948364258 s) (per-host)
I0419 10:12:43.629146 134324942784256 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.631005s.
I0419 10:12:49.012809 134324942784256 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 16.014653s.
I0419 10:12:49.013081 134322283607808 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 10:12:49.013198 134322283607808 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0419 10:12:49.013247 134322283607808 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0419 10:12:49.015854 134322283607808 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0419 10:12:49.016047 134447627163456 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0419 10:12:49.016177 134447627163456 checkpoint_manager.py:1458] Waiting for previous save to complete took 14.560227 seconds. If this number is high, consider checkpointing less frequently.
I0419 10:12:49.017952 134447627163456 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 15
I0419 10:12:49.022120 134447627163456 async_checkpointer.py:452] [process=7] 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_15_ocdbt_false/checkpoints/15.
I0419 10:12:49.980635 134447627163456 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0419 10:12:49.980743 134447627163456 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 10:12:50.022797 134447627163456 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043391s
I0419 10:12:50.022972 134447627163456 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.196 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29686522483825684 s) (per-host)
I0419 10:12:50.023026 134447627163456 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.296928s (batch_requests_ready=0.251605s, total_serialization_initiated=0.045251s, others=0.000072s)
I0419 10:12:50.023116 134447627163456 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.302023s (all_items=0.000013s, per_item={'items': '0.00001264'}, temp_paths=0.302010)
I0419 10:12:50.024060 134320102553344 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started.
I0419 10:12:50.024234 134447627163456 async_checkpointer.py:561] Finished blocking save. Time taken: 1.006209s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_15_ocdbt_false/checkpoints/15.
I0419 10:12:50.453892 134447627163456 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 10:12:50.454277 134322283607808 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0419 10:12:50.454451 134447627163456 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776593554.4559276, 'wait_for_prev_duration_secs': 14.560226678848267, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776593569.0179918, 'checkpointer_blocking_duration_secs': 1.0063481330871582, 'get_old_steps_start_time': 1776593570.0243587, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776593554.4537246, 'checkpoint_manager_blocking_duration_secs': 16.000689268112183}
I0419 10:12:50.454588 134447627163456 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0419 10:12:50.455212 134447627163456 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.185, Tokens/s/device: 13895.108, total_weights: 65536, loss: 5.864
I0419 10:12:50.461189 134447627163456 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.187, Tokens/s/device: 13895.485, total_weights: 65536, loss: 5.719
I0419 10:13:07.999742 134447627163456 metric_logger.py:185] completed step: 16, seconds: 16.297, TFLOP/s/device: 0.834, Tokens/s/device: 125.669, total_weights: 65536, loss: 5.605
I0419 10:13:08.147137 134447627163456 metric_logger.py:185] completed step: 17, seconds: 17.392, TFLOP/s/device: 0.781, Tokens/s/device: 117.754, total_weights: 65536, loss: 5.517
I0419 10:13:08.294571 134447627163456 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.195, Tokens/s/device: 13444.407, total_weights: 65536, loss: 5.448
I0419 10:13:08.441637 134447627163456 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0419 10:13:08.442317 134447627163456 checkpointing.py:776] Waited 0.0007002353668212891 seconds for step 19 to finish before starting checkpointing.
I0419 10:13:08.444349 134447627163456 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 10:13:12.783214 134323332171520 array_metadata_store.py:203] [process=7][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_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_7
I0419 10:13:18.320124 134320102553344 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 55.242 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 28.59396529197693 s) (per-host)
I0419 10:13:18.320245 134320102553344 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.296074s.
I0419 10:13:22.534465 134320102553344 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 32.510279s.
I0419 10:13:22.534810 134322283607808 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 10:13:22.534941 134322283607808 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0419 10:13:22.534999 134322283607808 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0419 10:13:22.536424 134322283607808 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0419 10:13:22.536603 134447627163456 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0419 10:13:22.536749 134447627163456 checkpoint_manager.py:1458] Waiting for previous save to complete took 14.092411 seconds. If this number is high, consider checkpointing less frequently.
I0419 10:13:22.538181 134447627163456 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 19
I0419 10:13:22.541683 134447627163456 async_checkpointer.py:452] [process=7] 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_15_ocdbt_false/checkpoints/19.
I0419 10:13:23.080258 134447627163456 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0419 10:13:23.080355 134447627163456 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 10:13:23.122915 134447627163456 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043652s
I0419 10:13:23.123075 134447627163456 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.152 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2994375228881836 s) (per-host)
I0419 10:13:23.123122 134447627163456 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299494s (batch_requests_ready=0.253717s, total_serialization_initiated=0.045715s, others=0.000061s)
I0419 10:13:23.123211 134447627163456 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303551s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.303540)
I0419 10:13:23.124071 134320127731456 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started.
I0419 10:13:23.124227 134447627163456 async_checkpointer.py:561] Finished blocking save. Time taken: 0.585976s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260419_085454/linen_xpk_test_pipeline_scan_nnx_20260419_085454_15_ocdbt_false/checkpoints/19.
I0419 10:13:23.557617 134447627163456 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 10:13:23.558001 134322283607808 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0419 10:13:23.558170 134447627163456 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776593588.4443214, 'wait_for_prev_duration_secs': 14.092411041259766, 'time_between_consecutive_saves_sec': 19.428427696228027, 'checkpointer_blocking_start_time': 1776593602.53822, 'checkpointer_blocking_duration_secs': 0.5861144065856934, 'get_old_steps_start_time': 1776593603.1243463, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776593588.4425726, 'checkpoint_manager_blocking_duration_secs': 15.115565538406372}
I0419 10:13:23.558279 134447627163456 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0419 10:13:23.558322 134447627163456 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 10:13:27.690356 134323332171520 array_metadata_store.py:203] [process=7][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_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_7
I0419 10:13:33.219629 134320127731456 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 151.944 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.395947456359863 s) (per-host)
I0419 10:13:33.219747 134320127731456 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.095567s.
I0419 10:13:38.215565 134320127731456 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 15.091347s.
I0419 10:13:38.215875 134322283607808 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 10:13:38.215986 134322283607808 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0419 10:13:38.216032 134322283607808 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0419 10:13:38.217339 134322283607808 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0419 10:13:38.217557 134447627163456 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0419 10:13:38.217704 134447627163456 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 10:13:38.218426 134447627163456 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.159, Tokens/s/device: 13891.244, 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 10:13:47 UTC 2026
EXIT_CODE=0