XPK Start: Fri Apr 17 15:03:19 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-17 15:03:42.736145: 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)
I0417 15:03:42.914993 134405731637056 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 15:03:51.954269 134405731637056 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 15:03:51.956498 134405731637056 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-rs-8hxvl-slice-job-0-0.mt-16-zarr3-false-rs-8hxvl:8482
I0417 15:03:53.169590 134405731637056 max_utils.py:284] Jax distributed system initialized!
I0417 15:03:59.210675 134405731637056 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 15:03:59.210775 134405731637056 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 15:03:59.210815 134405731637056 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
I0417 15:03:59.210868 134405731637056 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 15:04:00.236668 134405731637056 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 15:04:00.236965 134405731637056 checkpointing.py:677] Setting up checkpoint logger...
I0417 15:04:00.237020 134405731637056 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0417 15:04:00.237065 134405731637056 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 15:04:00.237403 134405731637056 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 0x7a3d0e540680>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 15:04:01.485699 134405731637056 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 15:04:01.485951 134405731637056 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 0x7a27fc649850>}, handler_registry=None
I0417 15:04:01.486194 134405731637056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a27fc649850>` 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`.
I0417 15:04:01.486243 134405731637056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3c01d64ad0>` 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`.
I0417 15:04:01.486279 134405731637056 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 0x7a27fc649850>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a27fc649850>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3c01d64ad0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3c01d64ad0>}).
I0417 15:04:01.486603 134405731637056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 15:04:01.486672 134405731637056 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a3c01f94d60> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 15:04:03.883190 134284987832064 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437956468219271, 'commit_timestamp_nsecs': 1776438003551890020, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0417 15:04:04.136170 134284551640832 checkpoint_manager.py:1871] Missing metrics for step 5
I0417 15:04:04.387790 134405731637056 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints
I0417 15:04:05.441807 134405731637056 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_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a3c01d3c500>
I0417 15:04:05.441994 134405731637056 checkpointing.py:301] Checkpoint manager created!
I0417 15:04:06.353926 134405731637056 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 15:04:06.354035 134405731637056 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 15:04:06.730283 134405731637056 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 15:04:06.730373 134405731637056 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 15:04:06.746317 134405731637056 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 15:04:06.746370 134405731637056 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 15:04:06.769204 134405731637056 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 15:04:06.769271 134405731637056 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 15:04:06.785393 134405731637056 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 15:04:06.785454 134405731637056 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 15:04:06.801501 134405731637056 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 15:04:06.801558 134405731637056 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 15:04:06.825731 134405731637056 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 15:04:06.825796 134405731637056 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 15:04:06.846266 134405731637056 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 15:04:06.846330 134405731637056 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 15:04:07.062230 134405731637056 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 15:04:07.062333 134405731637056 checkpointing.py:581] restoring from this run's directory step 9
I0417 15:04:07.308629 134405731637056 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9.
I0417 15:04:07.550600 134405731637056 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9.
I0417 15:04:07.551036 134405731637056 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.24 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9.
I0417 15:04:09.368383 1448 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 15:04:16.433755 134405731637056 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 223.473 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.0714335441589355 s) (per-host)
I0417 15:04:16.440002 134405731637056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 211.869 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.469680309295654 s) (per-host)
I0417 15:04:16.759704 134405731637056 event_tracking.py:138] [process=0] [sync] Finished load in 9.45 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9
I0417 15:04:16.759881 134405731637056 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776438247.3085315, 'checkpointer_duration_secs': 9.451297998428345, 'checkpoint_manager_start_time': 1776438247.062574, 'checkpoint_manager_duration_secs': 9.697257995605469}
fsdp: 32
I0417 15:04:17.170533 134405731637056 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0417 15:04:17.170652 134405731637056 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 15:04:17.170705 134405731637056 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 15:04:17.170761 134405731637056 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0417 15:04:17.170811 134405731637056 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 15:04:17.170864 134405731637056 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 15:04:17.170925 134405731637056 maxtext_utils.py:1651] 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)
I0417 15:04:17.170978 134405731637056 maxtext_utils.py:1651] 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')
I0417 15:04:17.171016 134405731637056 maxtext_utils.py:1651] 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)
I0417 15:04:17.171051 134405731637056 maxtext_utils.py:1651] 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)
I0417 15:04:17.171097 134405731637056 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0417 15:04:17.171142 134405731637056 maxtext_utils.py:1651] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0417 15:04:17.656490 134405731637056 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 15:04:17.656584 134405731637056 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 15:04:17.671655 134405731637056 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 15:04:17.671717 134405731637056 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 15:04:28.527288 134405731637056 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.
I0417 15:04:28.534577 134405731637056 metric_logger.py:301] number parameters: 1.104 billion
I0417 15:04:39.951343 134405731637056 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0417 15:04:40.106908 134405731637056 checkpointing.py:776] Waited 0.15554523468017578 seconds for step 10 to finish before starting checkpointing.
I0417 15:04:40.109263 134405731637056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 15:04:40.111328 134405731637056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10
I0417 15:04:40.114130 134405731637056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10.
I0417 15:04:40.841073 134405731637056 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 15:04:40.853664 134405731637056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 15:04:40.853729 134405731637056 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
I0417 15:04:41.113887 134283888944896 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10
I0417 15:04:41.130007 134405731637056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277164s
I0417 15:04:41.130881 134405731637056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.468 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28209614753723145 s) (per-host)
I0417 15:04:41.130969 134405731637056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282219s (batch_requests_ready=0.002290s, total_serialization_initiated=0.279123s, others=0.000806s)
I0417 15:04:41.131057 134405731637056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286116s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.286099)
I0417 15:04:41.131810 134405731637056 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10.
I0417 15:04:41.132134 134283897337600 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:24:41.132101
I0417 15:04:41.138089 134405731637056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 15:04:41.138327 134280101484288 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 15:04:41.138477 134405731637056 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776438280.1092446, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776438280.1113682, 'checkpointer_blocking_duration_secs': 1.0209109783172607, 'get_old_steps_start_time': 1776438281.132302, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776438280.107417, 'checkpoint_manager_blocking_duration_secs': 1.0310249328613281}
I0417 15:04:41.138581 134405731637056 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0417 15:04:41.138632 134405731637056 max_utils.py:750]
Memstats: After params initialized:
I0417 15:04:41.138677 134405731637056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 15:04:41.138709 134405731637056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 15:04:41.138735 134405731637056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 15:04:41.138760 134405731637056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 15:04:41.479360 134405731637056 metric_logger.py:196] completed step: 10, seconds: 11.124, TFLOP/s/device: 1.221, Tokens/s/device: 184.100, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0417 15:04:41.624534 134405731637056 metric_logger.py:196] completed step: 11, seconds: 1.526, TFLOP/s/device: 8.903, Tokens/s/device: 1341.993, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0417 15:04:41.772137 134405731637056 metric_logger.py:196] completed step: 12, seconds: 0.011, TFLOP/s/device: 1267.693, Tokens/s/device: 191080.425, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0417 15:04:41.919358 134405731637056 metric_logger.py:196] completed step: 13, seconds: 0.146, TFLOP/s/device: 93.246, Tokens/s/device: 14055.122, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0417 15:04:41.928446 134405731637056 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0417 15:04:42.213330 134405731637056 checkpointing.py:776] Waited 0.28471875190734863 seconds for step 15 to finish before starting checkpointing.
I0417 15:04:42.216477 134405731637056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 15:04:42.873776 134280621885184 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10/items
I0417 15:04:43.058860 134279561303808 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776438282630195176, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776438282630195176, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA
I0417 15:04:45.781351 134280130848512 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10/items/array_metadatas/process_0
I0417 15:05:25.497752 134280122455808 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.481931s (commit=40.271526s, array_metadata_write=1.210405s)
I0417 15:05:25.499118 134283897337600 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.377 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.65030360221863 s) (per-host)
I0417 15:05:25.499288 134283897337600 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.367042s.
I0417 15:05:27.053382 134279561303808 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776438282630195176, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA
I0417 15:05:27.392403 134283897337600 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000376s.
I0417 15:05:28.446673 134283897337600 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.057860s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10/items
I0417 15:05:28.448779 134283897337600 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10/items
I0417 15:05:29.001537 134283897337600 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10
I0417 15:05:30.432140 134283897337600 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10`.
I0417 15:05:30.433283 134283897337600 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.32 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/10
I0417 15:05:30.434830 134283897337600 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.302581s.
I0417 15:05:30.435075 134280101484288 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 15:05:30.435197 134280101484288 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 15:05:30.435299 134280101484288 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0417 15:05:30.437123 134280101484288 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0417 15:05:30.437465 134405731637056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0417 15:05:30.437646 134405731637056 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.221170 seconds. If this number is high, consider checkpointing less frequently.
I0417 15:05:30.439726 134405731637056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15
I0417 15:05:30.442205 134405731637056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15.
I0417 15:05:30.728910 134405731637056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 15:05:30.729042 134405731637056 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
I0417 15:05:30.756608 134405731637056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028694s
I0417 15:05:30.757547 134405731637056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.946 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0376734733581543 s) (per-host)
I0417 15:05:30.757655 134405731637056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037805s (batch_requests_ready=0.004481s, total_serialization_initiated=0.032455s, others=0.000869s)
I0417 15:05:30.757777 134405731637056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042823s (all_items=0.000030s, per_item={'items': '0.00003028'}, temp_paths=0.042793)
I0417 15:05:30.758734 134405731637056 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.32 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15.
I0417 15:05:30.759139 134280101484288 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:25:30.759102
I0417 15:05:30.761546 134405731637056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 15:05:30.761924 134280130848512 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 15:05:30.762157 134405731637056 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776438282.2164319, 'wait_for_prev_duration_secs': 48.22117018699646, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776438330.439798, 'checkpointer_blocking_duration_secs': 0.3195207118988037, 'get_old_steps_start_time': 1776438330.7593453, 'get_old_steps_duration_secs': 4.38690185546875e-05, 'checkpoint_manager_blocking_start_time': 1776438282.2140152, 'checkpoint_manager_blocking_duration_secs': 48.54808521270752}
I0417 15:05:30.762302 134405731637056 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0417 15:05:30.966634 134283897337600 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15
I0417 15:05:32.759592 134283888944896 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15/items
I0417 15:05:36.158747 134280621885184 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15/items/array_metadatas/process_0
I0417 15:05:52.179559 134405731637056 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.040, Tokens/s/device: 13873.271, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0417 15:05:52.188503 134405731637056 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.529, Tokens/s/device: 13946.963, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0417 15:05:52.334011 134405731637056 metric_logger.py:196] completed step: 16, seconds: 70.259, TFLOP/s/device: 0.193, Tokens/s/device: 29.149, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0417 15:05:52.481291 134405731637056 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1922.344, Tokens/s/device: 289756.650, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0417 15:05:52.628597 134405731637056 metric_logger.py:196] completed step: 18, seconds: 0.147, TFLOP/s/device: 92.734, Tokens/s/device: 13977.805, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0417 15:05:52.775374 134405731637056 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0417 15:05:52.776066 134405731637056 checkpointing.py:776] Waited 0.0007116794586181641 seconds for step 19 to finish before starting checkpointing.
I0417 15:05:52.778043 134405731637056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 15:06:15.348759 134280139241216 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.380361s (commit=39.528327s, array_metadata_write=1.852033s)
I0417 15:06:15.349819 134280101484288 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.393 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.62995219230652 s) (per-host)
I0417 15:06:15.349968 134280101484288 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.590684s.
I0417 15:06:16.817906 134280101484288 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000270s.
I0417 15:06:17.674025 134280101484288 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.475372s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15/items
I0417 15:06:17.675726 134280101484288 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15/items
I0417 15:06:18.229603 134280101484288 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15
I0417 15:06:19.638350 134280101484288 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15`.
I0417 15:06:19.639069 134280101484288 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.20 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/15
I0417 15:06:19.640304 134280101484288 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.881017s.
I0417 15:06:19.640524 134280130848512 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 15:06:19.640639 134280130848512 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 15:06:19.640704 134280130848512 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0417 15:06:19.642768 134280130848512 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0417 15:06:19.642947 134405731637056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0417 15:06:19.643084 134405731637056 checkpoint_manager.py:1452] Waiting for previous save to complete took 26.865046 seconds. If this number is high, consider checkpointing less frequently.
I0417 15:06:19.644777 134405731637056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19
I0417 15:06:19.646620 134405731637056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19.
I0417 15:06:19.932187 134405731637056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 15:06:19.932278 134405731637056 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
I0417 15:06:19.957424 134405731637056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026169s
I0417 15:06:19.958143 134405731637056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 50.521 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030533313751220703 s) (per-host)
I0417 15:06:19.958218 134405731637056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030626s (batch_requests_ready=0.001873s, total_serialization_initiated=0.028073s, others=0.000680s)
I0417 15:06:19.958289 134405731637056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034567s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.034549)
I0417 15:06:19.958971 134405731637056 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.31 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19.
I0417 15:06:19.959301 134280130848512 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:26:19.959264
I0417 15:06:19.963294 134405731637056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 15:06:19.963528 134280630277888 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 15:06:19.963655 134405731637056 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776438352.7780094, 'wait_for_prev_duration_secs': 26.86504578590393, 'time_between_consecutive_saves_sec': 22.340813875198364, 'checkpointer_blocking_start_time': 1776438379.644816, 'checkpointer_blocking_duration_secs': 0.314617395401001, 'get_old_steps_start_time': 1776438379.9594524, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776438352.7763023, 'checkpoint_manager_blocking_duration_secs': 27.187318325042725}
I0417 15:06:19.963760 134405731637056 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0417 15:06:19.963802 134405731637056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 15:06:20.169625 134280101484288 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19
I0417 15:06:22.332063 134281212983040 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19/items
I0417 15:06:25.266339 134283888944896 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19/items/array_metadatas/process_0
I0417 15:07:06.452304 134283326895872 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.934892s (commit=40.470116s, array_metadata_write=2.464776s)
I0417 15:07:06.453901 134280130848512 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.951 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.5262930393219 s) (per-host)
I0417 15:07:06.453964 134280130848512 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.494556s.
I0417 15:07:09.979614 134280130848512 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000230s.
I0417 15:07:11.148375 134280130848512 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.206953s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19/items
I0417 15:07:11.151593 134280130848512 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19/items
I0417 15:07:12.089702 134280130848512 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19
I0417 15:07:13.482114 134280130848512 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19`.
I0417 15:07:13.483011 134280130848512 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.84 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/19
I0417 15:07:13.484452 134280130848512 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 53.525040s.
I0417 15:07:13.484655 134280630277888 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 15:07:13.484774 134280630277888 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 15:07:13.484825 134280630277888 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0417 15:07:13.486419 134280630277888 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0417 15:07:13.486555 134405731637056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0417 15:07:13.486652 134405731637056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 15:07:13.487616 134405731637056 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.312, Tokens/s/device: 13914.272, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 17 15:07:24 UTC 2026
EXIT_CODE=0