XPK Start: Sat Apr 18 13:22:04 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 13:22:28.249456: 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 13:22:28.779864 133579444102976 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 13:22:37.821823 133579444102976 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 13:22:37.824126 133579444102976 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--lxlw1-slice-job-0-0.mt-08-checkpoint-async--lxlw1:8482
I0418 13:22:39.516245 133579444102976 max_utils.py:284] Jax distributed system initialized!
I0418 13:22:45.799290 133579444102976 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 13:22:45.799392 133579444102976 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 13:22:45.799432 133579444102976 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 13:22:45.799467 133579444102976 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 13:22:46.482312 133579444102976 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 13:22:46.482578 133579444102976 checkpointing.py:677] Setting up checkpoint logger...
I0418 13:22:46.482631 133579444102976 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 13:22:46.482674 133579444102976 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 13:22:46.483026 133579444102976 base_pytree_checkpoint_handler.py:441] 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 0x797cabb48110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 13:22:49.400939 133579444102976 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 13:22:49.401179 133579444102976 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 0x79679c727470>}, handler_registry=None
I0418 13:22:49.401442 133579444102976 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79679c727470>` 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 13:22:49.401494 133579444102976 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79679c726930>` 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 13:22:49.401530 133579444102976 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 0x79679c727470>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79679c727470>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79679c726930>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79679c726930>}).
I0418 13:22:49.401906 133579444102976 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 13:22:49.402002 133579444102976 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79679c6cb9c0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 13:22:51.837120 133579444102976 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints
I0418 13:22:52.261760 133579444102976 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_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79679c726ed0>
I0418 13:22:52.261976 133579444102976 checkpointing.py:301] Checkpoint manager created!
I0418 13:22:53.162135 133579444102976 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 13:22:53.162244 133579444102976 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 13:22:53.537018 133579444102976 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 13:22:53.537107 133579444102976 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 13:22:53.553258 133579444102976 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 13:22:53.553316 133579444102976 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 13:22:53.577366 133579444102976 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 13:22:53.577431 133579444102976 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 13:22:53.593751 133579444102976 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 13:22:53.593812 133579444102976 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 13:22:53.610116 133579444102976 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 13:22:53.610174 133579444102976 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 13:22:53.634675 133579444102976 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 13:22:53.634746 133579444102976 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 13:22:53.655688 133579444102976 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 13:22:53.655750 133579444102976 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 13:22:53.870280 133579444102976 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 13:22:53.870385 133579444102976 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 13:22:55.300100 133579444102976 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 13:22:55.300234 133579444102976 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 13:22:55.300287 133579444102976 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 13:22:55.300343 133579444102976 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 13:22:55.300395 133579444102976 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 13:22:55.300433 133579444102976 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 13:22:55.300484 133579444102976 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 13:22:55.300533 133579444102976 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 13:22:55.300575 133579444102976 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 13:22:55.300609 133579444102976 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 13:22:55.300655 133579444102976 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 13:22:55.300701 133579444102976 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 13:22:55.786420 133579444102976 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 13:22:55.786514 133579444102976 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 13:22:55.803307 133579444102976 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 13:22:55.803374 133579444102976 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 13:23:06.813741 133579444102976 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 13:23:06.821616 133579444102976 metric_logger.py:301] number parameters: 1.104 billion
I0418 13:23:18.357060 133579444102976 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 13:23:18.510509 133579444102976 checkpointing.py:776] Waited 0.1534280776977539 seconds for step 0 to finish before starting checkpointing.
I0418 13:23:18.513103 133579444102976 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 13:23:18.514769 133579444102976 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 13:23:18.516181 133579444102976 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0.
I0418 13:23:19.264911 133579444102976 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 13:23:19.279176 133579444102976 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 13:23:19.279243 133579444102976 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 13:23:19.547393 133450582902528 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0
I0418 13:23:19.549714 133579444102976 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271263s
I0418 13:23:19.550587 133579444102976 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.551 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.277911901473999 s) (per-host)
I0418 13:23:19.550661 133579444102976 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278006s (batch_requests_ready=0.002372s, total_serialization_initiated=0.274857s, others=0.000777s)
I0418 13:23:19.550743 133579444102976 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.281875s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.281857)
I0418 13:23:19.551518 133579444102976 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0.
I0418 13:23:19.551872 133450591295232 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:43:19.551818
I0418 13:23:19.566046 133579444102976 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 13:23:19.566319 133448995960576 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 13:23:19.566452 133579444102976 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776518598.5130847, 'wait_for_prev_duration_secs': 6.0558319091796875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776518598.514807, 'checkpointer_blocking_duration_secs': 1.037212610244751, 'get_old_steps_start_time': 1776518599.5520437, 'get_old_steps_duration_secs': 3.4332275390625e-05, 'checkpoint_manager_blocking_start_time': 1776518598.5110502, 'checkpoint_manager_blocking_duration_secs': 1.0553638935089111}
I0418 13:23:19.566552 133579444102976 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 13:23:19.566602 133579444102976 max_utils.py:750]
Memstats: After params initialized:
I0418 13:23:19.566651 133579444102976 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 13:23:19.566685 133579444102976 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 13:23:19.566710 133579444102976 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 13:23:19.566733 133579444102976 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 13:23:19.905622 133579444102976 metric_logger.py:196] completed step: 0, seconds: 11.234, TFLOP/s/device: 1.209, Tokens/s/device: 182.307, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 13:23:19.909007 133579444102976 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/tensorboard/'
I0418 13:23:20.337255 133579444102976 metric_logger.py:196] completed step: 1, seconds: 1.547, TFLOP/s/device: 8.785, Tokens/s/device: 1324.172, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 13:23:20.483360 133579444102976 metric_logger.py:196] completed step: 2, seconds: 0.432, TFLOP/s/device: 31.420, Tokens/s/device: 4736.027, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0418 13:23:20.630658 133579444102976 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1438.553, Tokens/s/device: 216834.304, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0418 13:23:20.639996 133579444102976 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0418 13:23:20.924654 133579444102976 checkpointing.py:776] Waited 0.2845025062561035 seconds for step 5 to finish before starting checkpointing.
I0418 13:23:20.927701 133579444102976 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 13:23:22.176653 133450566117120 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/items
I0418 13:23:22.811578 133448423614208 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776518601917348229, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776518601917348229, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA
I0418 13:23:24.208633 2846 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 13:23:25.987097 133449485133568 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_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/items/array_metadatas/process_0
I0418 13:24:02.249755 133449476740864 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.053984s (commit=36.872213s, array_metadata_write=1.181772s)
I0418 13:24:02.251254 133450591295232 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 36.753 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.97853112220764 s) (per-host)
I0418 13:24:02.251422 133450591295232 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.699434s.
I0418 13:24:03.957935 133450591295232 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776518601917348229, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA
I0418 13:24:05.170188 133448423614208 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776518601917348229, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA
I0418 13:24:05.544537 133450591295232 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000392s.
I0418 13:24:05.764785 133450591295232 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 13:24:06.709940 133450591295232 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.619495s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/items
I0418 13:24:06.712255 133450591295232 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/items
I0418 13:24:07.239258 133450591295232 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0
I0418 13:24:09.065747 133450591295232 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0`.
I0418 13:24:09.066887 133450591295232 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.55 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0
I0418 13:24:09.068525 133450591295232 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.516533s.
I0418 13:24:09.068814 133448995960576 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 13:24:09.068998 133448995960576 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 13:24:09.069120 133448995960576 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 13:24:09.071158 133448995960576 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 13:24:09.071441 133579444102976 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 13:24:09.071613 133579444102976 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.143919 seconds. If this number is high, consider checkpointing less frequently.
I0418 13:24:09.073496 133579444102976 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5
I0418 13:24:09.075993 133579444102976 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5.
I0418 13:24:10.201626 133579444102976 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 13:24:10.201757 133579444102976 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 13:24:10.239620 133579444102976 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039000s
I0418 13:24:10.240656 133579444102976 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 32.159 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.047966957092285156 s) (per-host)
I0418 13:24:10.240761 133579444102976 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.048096s (batch_requests_ready=0.004316s, total_serialization_initiated=0.042832s, others=0.000948s)
I0418 13:24:10.240909 133579444102976 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.052933s (all_items=0.000025s, per_item={'items': '0.00002503'}, temp_paths=0.052908)
I0418 13:24:10.241930 133579444102976 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5.
I0418 13:24:10.242316 133448995960576 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:44:10.242274
I0418 13:24:10.244813 133579444102976 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 13:24:10.245172 133448970782464 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 13:24:10.245394 133579444102976 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776518600.9276538, 'wait_for_prev_duration_secs': 48.14391875267029, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776518649.0735593, 'checkpointer_blocking_duration_secs': 1.1689393520355225, 'get_old_steps_start_time': 1776518650.2425265, 'get_old_steps_duration_secs': 4.124641418457031e-05, 'checkpoint_manager_blocking_start_time': 1776518600.9253957, 'checkpoint_manager_blocking_duration_secs': 49.31994581222534}
I0418 13:24:10.245538 133579444102976 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0418 13:24:10.437214 133450591295232 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5
I0418 13:24:12.622688 133450052916992 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5/items
I0418 13:24:15.927627 133449489856256 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_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5/items/array_metadatas/process_0
I0418 13:24:33.767322 133579444102976 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.392, Tokens/s/device: 13926.383, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0418 13:24:33.776364 133579444102976 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.287, Tokens/s/device: 13910.492, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0418 13:24:33.921942 133579444102976 metric_logger.py:196] completed step: 6, seconds: 73.135, TFLOP/s/device: 0.186, Tokens/s/device: 28.003, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0418 13:24:34.069574 133579444102976 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1952.174, Tokens/s/device: 294252.874, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0418 13:24:34.216631 133579444102976 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.521, Tokens/s/device: 13945.729, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0418 13:24:34.363708 133579444102976 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 13:24:34.364483 133579444102976 checkpointing.py:776] Waited 0.0007920265197753906 seconds for step 9 to finish before starting checkpointing.
I0418 13:24:34.366577 133579444102976 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 13:24:56.756812 133449481463552 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.037577s (commit=41.397577s, array_metadata_write=1.640000s)
I0418 13:24:56.757995 133448995960576 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.922 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.56528687477112 s) (per-host)
I0418 13:24:56.758128 133448995960576 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.515664s.
I0418 13:24:59.734796 133448995960576 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 13:24:59.919437 133448995960576 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000243s.
I0418 13:25:00.674510 133448995960576 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.191455s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5/items
I0418 13:25:00.676330 133448995960576 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5/items
I0418 13:25:01.205523 133448995960576 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5
I0418 13:25:02.605010 133448995960576 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5`.
I0418 13:25:02.605831 133448995960576 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.53 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/5
I0418 13:25:02.607396 133448995960576 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.364931s.
I0418 13:25:02.607612 133448970782464 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 13:25:02.607729 133448970782464 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 13:25:02.607795 133448970782464 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0418 13:25:02.610424 133448970782464 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0418 13:25:02.610582 133579444102976 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 13:25:02.610672 133579444102976 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.244105 seconds. If this number is high, consider checkpointing less frequently.
I0418 13:25:02.613562 133579444102976 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 13:25:02.615630 133579444102976 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9.
I0418 13:25:03.329531 133579444102976 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 13:25:03.329625 133579444102976 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 13:25:03.356075 133579444102976 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027469s
I0418 13:25:03.356898 133579444102976 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.216 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03199315071105957 s) (per-host)
I0418 13:25:03.356975 133579444102976 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032090s (batch_requests_ready=0.001913s, total_serialization_initiated=0.029435s, others=0.000743s)
I0418 13:25:03.357059 133579444102976 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036194s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.036177)
I0418 13:25:03.357751 133579444102976 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9.
I0418 13:25:03.358100 133448985462528 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:45:03.358076
I0418 13:25:03.366596 133579444102976 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 13:25:03.366863 133448970782464 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 13:25:03.366990 133579444102976 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776518674.3665426, 'wait_for_prev_duration_secs': 28.244104862213135, 'time_between_consecutive_saves_sec': 25.29530906677246, 'checkpointer_blocking_start_time': 1776518702.6136026, 'checkpointer_blocking_duration_secs': 0.7445921897888184, 'get_old_steps_start_time': 1776518703.3582141, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776518674.3647125, 'checkpoint_manager_blocking_duration_secs': 29.002243757247925}
I0418 13:25:03.367108 133579444102976 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 13:25:03.367152 133579444102976 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 13:25:03.576714 133448995960576 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9
I0418 13:25:05.771271 133449489856256 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9/items
I0418 13:25:09.491958 133450053990144 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_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9/items/array_metadatas/process_0
I0418 13:25:50.737614 133449481463552 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.435105s (commit=41.330202s, array_metadata_write=2.104903s)
I0418 13:25:50.738735 133448985462528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.315 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.41380214691162 s) (per-host)
I0418 13:25:50.738886 133448985462528 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.380713s.
I0418 13:25:52.355089 133448985462528 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 13:25:52.803328 133448985462528 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000240s.
I0418 13:25:53.445586 133448985462528 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.929780s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9/items
I0418 13:25:53.447418 133448985462528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9/items
I0418 13:25:54.443614 133448985462528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9
I0418 13:25:57.119656 133448985462528 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9`.
I0418 13:25:57.120493 133448985462528 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 54.51 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9
I0418 13:25:57.121886 133448985462528 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 53.763710s.
I0418 13:25:57.122100 133448970782464 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 13:25:57.122214 133448970782464 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 13:25:57.122266 133448970782464 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 13:25:57.124706 133448970782464 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 13:25:57.124895 133579444102976 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 13:25:57.125049 133579444102976 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 13:25:57.125960 133579444102976 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.150, Tokens/s/device: 13889.831, 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 13:26:09 UTC 2026
EXIT_CODE=0