XPK Start: Sat Apr 18 19:50:46 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-18 19:51:10.747563: 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)
I0418 19:51:10.927767 139117071529792 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 19:51:19.968739 139117071529792 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 19:51:19.970961 139117071529792 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-qr9e6-slice-job-0-0.mt-16-zarr3-false-sv-qr9e6:8482
I0418 19:51:21.696129 139117071529792 max_utils.py:284] Jax distributed system initialized!
I0418 19:51:27.653770 139117071529792 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 19:51:27.653888 139117071529792 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 19:51:27.653928 139117071529792 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0418 19:51:27.653962 139117071529792 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 19:51:28.344476 139117071529792 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 19:51:28.344750 139117071529792 checkpointing.py:677] Setting up checkpoint logger...
I0418 19:51:28.344803 139117071529792 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0418 19:51:28.344862 139117071529792 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 19:51:28.345193 139117071529792 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e8628d8dfd0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 19:51:31.166200 139117071529792 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 19:51:31.166433 139117071529792 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e720434f170>}, handler_registry=None
I0418 19:51:31.166666 139117071529792 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e720434f170>` 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`.
I0418 19:51:31.166712 139117071529792 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e7204352870>` 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`.
I0418 19:51:31.166748 139117071529792 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 0x7e720434f170>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e720434f170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e7204352870>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e7204352870>}).
I0418 19:51:31.167075 139117071529792 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 19:51:31.167145 139117071529792 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e70e465fd80> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 19:51:33.201721 139117071529792 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints
I0418 19:51:34.475622 139117071529792 checkpoint_manager.py:929] [process=0][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_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e72043500e0>
I0418 19:51:34.475793 139117071529792 checkpointing.py:301] Checkpoint manager created!
I0418 19:51:35.387178 139117071529792 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 19:51:35.387288 139117071529792 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 19:51:35.768685 139117071529792 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 19:51:35.768776 139117071529792 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 19:51:35.785243 139117071529792 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 19:51:35.785299 139117071529792 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 19:51:35.809202 139117071529792 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 19:51:35.809267 139117071529792 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 19:51:35.825659 139117071529792 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 19:51:35.825721 139117071529792 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 19:51:35.846991 139117071529792 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 19:51:35.847086 139117071529792 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 19:51:35.872331 139117071529792 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 19:51:35.872404 139117071529792 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 19:51:35.894300 139117071529792 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 19:51:35.894363 139117071529792 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 19:51:36.113126 139117071529792 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 19:51:36.113258 139117071529792 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 19:51:37.534700 139117071529792 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 19:51:37.534915 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0418 19:51:37.535005 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0418 19:51:37.535082 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0418 19:51:37.535151 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0418 19:51:37.535191 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0418 19:51:37.535244 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/key/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0418 19:51:37.535297 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/out/kernel
Shape: float32[16,16,128,2048]
Logical: P('heads', 'layers', 'kv', 'embed')
Physical: (None, None, None, 'fsdp')
I0418 19:51:37.535337 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/query/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'q_heads', 'kv')
Physical: ('fsdp', None, None, None)
I0418 19:51:37.535372 139117071529792 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/value/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0418 19:51:37.535419 139117071529792 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 19:51:37.535468 139117071529792 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 19:51:38.032956 139117071529792 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 19:51:38.033051 139117071529792 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 19:51:38.048598 139117071529792 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 19:51:38.048659 139117071529792 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 19:51:48.982849 139117071529792 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.
I0418 19:51:48.990690 139117071529792 metric_logger.py:301] number parameters: 1.104 billion
I0418 19:52:00.619645 139117071529792 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 19:52:00.773612 139117071529792 checkpointing.py:776] Waited 0.153947114944458 seconds for step 0 to finish before starting checkpointing.
I0418 19:52:00.775986 139117071529792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 19:52:00.777597 139117071529792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 19:52:00.779017 139117071529792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0.
I0418 19:52:01.597563 139117071529792 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 19:52:01.611669 139117071529792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 19:52:01.611736 139117071529792 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
I0418 19:52:01.849591 138989209179904 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0
I0418 19:52:01.880700 139117071529792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.269676s
I0418 19:52:01.881594 139117071529792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.586 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27615833282470703 s) (per-host)
I0418 19:52:01.881656 139117071529792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.276237s (batch_requests_ready=0.002187s, total_serialization_initiated=0.273262s, others=0.000788s)
I0418 19:52:01.881730 139117071529792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.280159s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.280142)
I0418 19:52:01.882474 139117071529792 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.10 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0.
I0418 19:52:01.882883 138987567126272 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 20:12:01.882843
I0418 19:52:01.902731 139117071529792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 19:52:01.903067 138989217572608 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 19:52:01.903230 139117071529792 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776541920.775968, 'wait_for_prev_duration_secs': 6.0558319091796875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776541920.777635, 'checkpointer_blocking_duration_secs': 1.105393648147583, 'get_old_steps_start_time': 1776541921.8830516, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776541920.7741756, 'checkpoint_manager_blocking_duration_secs': 1.129014015197754}
I0418 19:52:01.903333 139117071529792 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 19:52:01.903384 139117071529792 max_utils.py:750]
Memstats: After params initialized:
I0418 19:52:01.903437 139117071529792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 19:52:01.903471 139117071529792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 19:52:01.903499 139117071529792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 19:52:01.903525 139117071529792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 19:52:02.254756 139117071529792 metric_logger.py:196] completed step: 0, seconds: 11.326, TFLOP/s/device: 1.200, Tokens/s/device: 180.820, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 19:52:02.257641 139117071529792 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/tensorboard/'
I0418 19:52:02.659852 139117071529792 metric_logger.py:196] completed step: 1, seconds: 1.633, TFLOP/s/device: 8.318, Tokens/s/device: 1253.842, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 19:52:02.806022 139117071529792 metric_logger.py:196] completed step: 2, seconds: 0.406, TFLOP/s/device: 33.490, Tokens/s/device: 5047.928, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0418 19:52:02.953518 139117071529792 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1462.397, Tokens/s/device: 220428.372, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0418 19:52:02.962650 139117071529792 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0418 19:52:03.247521 139117071529792 checkpointing.py:776] Waited 0.28473520278930664 seconds for step 5 to finish before starting checkpointing.
I0418 19:52:03.250464 139117071529792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 19:52:04.500638 138989192394496 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/items
I0418 19:52:04.639586 138985444820736 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776541924238048659, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776541924238048659, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 19:52:05.622274 2865 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 19:52:07.798691 138987583911680 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/items/array_metadatas/process_0
I0418 19:52:47.320662 138987575518976 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.710060s (commit=40.398626s, array_metadata_write=1.311434s)
I0418 19:52:47.322445 138987567126272 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.552 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.71697402000427 s) (per-host)
I0418 19:52:47.322572 138987567126272 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.439570s.
I0418 19:52:48.065578 138987567126272 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776541924238048659, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 19:52:48.793526 138985444820736 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776541924238048659, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 19:52:49.654630 138987567126272 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000392s.
I0418 19:52:50.755311 138987567126272 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.570055s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/items
I0418 19:52:50.757458 138987567126272 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0/items
I0418 19:52:51.278134 138987567126272 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0
I0418 19:52:53.501884 138987567126272 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0`.
I0418 19:52:53.503014 138987567126272 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.73 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/0
I0418 19:52:53.504489 138987567126272 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.621489s.
I0418 19:52:53.504696 138989217572608 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 19:52:53.504801 138989217572608 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 19:52:53.504919 138989217572608 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 19:52:53.506779 138989217572608 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 19:52:53.507063 139117071529792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0418 19:52:53.507246 139117071529792 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.256787 seconds. If this number is high, consider checkpointing less frequently.
I0418 19:52:53.509151 139117071529792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5
I0418 19:52:53.511576 139117071529792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5.
I0418 19:52:54.258105 139117071529792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 19:52:54.258239 139117071529792 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
I0418 19:52:54.302822 139117071529792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045710s
I0418 19:52:54.303839 139117071529792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 28.257 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05459094047546387 s) (per-host)
I0418 19:52:54.303959 139117071529792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.054753s (batch_requests_ready=0.004024s, total_serialization_initiated=0.049803s, others=0.000926s)
I0418 19:52:54.304101 139117071529792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.059702s (all_items=0.000024s, per_item={'items': '0.00002408'}, temp_paths=0.059678)
I0418 19:52:54.305151 139117071529792 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.80 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5.
I0418 19:52:54.305648 138989217572608 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 20:12:54.305611
I0418 19:52:54.308127 139117071529792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 19:52:54.308508 138988161595136 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 19:52:54.308739 139117071529792 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776541923.2504172, 'wait_for_prev_duration_secs': 50.256787061691284, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776541973.5092173, 'checkpointer_blocking_duration_secs': 0.7966616153717041, 'get_old_steps_start_time': 1776541974.3059058, 'get_old_steps_duration_secs': 4.291534423828125e-05, 'checkpoint_manager_blocking_start_time': 1776541923.248186, 'checkpoint_manager_blocking_duration_secs': 51.0604989528656}
I0418 19:52:54.308902 139117071529792 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0418 19:52:54.921036 138984886560512 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5
I0418 19:52:56.638727 138987604883200 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5/items
I0418 19:52:59.492737 138989768644352 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5/items/array_metadatas/process_0
I0418 19:53:17.275035 139117071529792 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.380, Tokens/s/device: 13924.490, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0418 19:53:17.284075 139117071529792 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.168, Tokens/s/device: 13892.563, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0418 19:53:17.429785 139117071529792 metric_logger.py:196] completed step: 6, seconds: 74.320, TFLOP/s/device: 0.183, Tokens/s/device: 27.557, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0418 19:53:17.577035 139117071529792 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1939.079, Tokens/s/device: 292279.149, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0418 19:53:17.724940 139117071529792 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.455, Tokens/s/device: 13935.860, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0418 19:53:17.871530 139117071529792 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 19:53:17.872315 139117071529792 checkpointing.py:776] Waited 0.0008039474487304688 seconds for step 9 to finish before starting checkpointing.
I0418 19:53:17.874327 139117071529792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 19:53:40.836673 138989209179904 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.075196s (commit=41.050632s, array_metadata_write=2.024564s)
I0418 19:53:40.837832 138989217572608 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.905 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.588555335998535 s) (per-host)
I0418 19:53:40.837985 138989217572608 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.532187s.
I0418 19:53:43.139179 138989217572608 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000242s.
I0418 19:53:44.413492 138989217572608 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.291419s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5/items
I0418 19:53:44.414995 138989217572608 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5/items
I0418 19:53:44.947100 138989217572608 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5
I0418 19:53:46.320713 138989217572608 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5`.
I0418 19:53:46.321572 138989217572608 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.81 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/5
I0418 19:53:46.323053 138989217572608 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.017255s.
I0418 19:53:46.323171 138988161595136 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 19:53:46.323239 138988161595136 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 19:53:46.323318 138988161595136 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0418 19:53:46.326894 138988161595136 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0418 19:53:46.327010 139117071529792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5.
W0418 19:53:46.327069 139117071529792 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.452755 seconds. If this number is high, consider checkpointing less frequently.
I0418 19:53:46.328801 139117071529792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 19:53:46.330964 139117071529792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9.
I0418 19:53:47.045884 139117071529792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 19:53:47.045980 139117071529792 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
I0418 19:53:47.071301 139117071529792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026309s
I0418 19:53:47.072134 139117071529792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 50.318 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03065657615661621 s) (per-host)
I0418 19:53:47.072203 139117071529792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030748s (batch_requests_ready=0.001784s, total_serialization_initiated=0.028181s, others=0.000784s)
I0418 19:53:47.072294 139117071529792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034794s (all_items=0.000014s, per_item={'items': '0.00001431'}, temp_paths=0.034780)
I0418 19:53:47.073011 139117071529792 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9.
I0418 19:53:47.073291 138989209179904 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 20:13:47.073253
I0418 19:53:47.081270 139117071529792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 19:53:47.081511 138988161595136 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 19:53:47.081639 139117071529792 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776541997.8742933, 'wait_for_prev_duration_secs': 28.452754974365234, 'time_between_consecutive_saves_sec': 24.367435455322266, 'checkpointer_blocking_start_time': 1776542026.3288612, 'checkpointer_blocking_duration_secs': 0.7445716857910156, 'get_old_steps_start_time': 1776542027.0734508, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776541997.8725517, 'checkpoint_manager_blocking_duration_secs': 29.209054708480835}
I0418 19:53:47.081755 139117071529792 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 19:53:47.081798 139117071529792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 19:53:47.707293 138989217572608 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9
I0418 19:53:50.738734 138988144809728 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9/items
I0418 19:53:55.705337 138989768644352 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9/items/array_metadatas/process_0
I0418 19:54:34.172961 138989235336960 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.463018s (commit=40.267415s, array_metadata_write=1.195603s)
I0418 19:54:34.174078 138989209179904 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.514 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.13257384300232 s) (per-host)
I0418 19:54:34.174201 138989209179904 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.100798s.
I0418 19:54:36.091356 138989209179904 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000240s.
I0418 19:54:36.967411 138989209179904 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.489587s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9/items
I0418 19:54:36.970887 138989209179904 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9/items
I0418 19:54:37.657775 138989209179904 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9
I0418 19:54:39.470232 138989209179904 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9`.
I0418 19:54:39.471063 138989209179904 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.14 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_16_zarr3_false/checkpoints/9
I0418 19:54:39.472653 138989209179904 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.399248s.
I0418 19:54:39.472862 138988161595136 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 19:54:39.472984 138988161595136 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 19:54:39.473050 138988161595136 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 19:54:39.475717 138988161595136 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 19:54:39.475906 139117071529792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0418 19:54:39.476050 139117071529792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 19:54:39.477111 139117071529792 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.379, Tokens/s/device: 13924.395, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sat Apr 18 19:54:51 UTC 2026
EXIT_CODE=0