XPK Start: Sat Apr 18 01:42:38 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 01:43:02.663509: 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 01:43:03.198070 136845296777024 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 01:43:12.239162 136845296777024 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 01:43:12.241550 136845296777024 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-sv-dj5ri-slice-job-0-0.mt-16-zarr3-true-sv-dj5ri:8482
I0418 01:43:13.357671 136845296777024 max_utils.py:284] Jax distributed system initialized!
I0418 01:43:19.305716 136845296777024 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 01:43:19.305817 136845296777024 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 01:43:19.305877 136845296777024 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 01:43:19.305916 136845296777024 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 01:43:19.986952 136845296777024 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 01:43:19.987221 136845296777024 checkpointing.py:677] Setting up checkpoint logger...
I0418 01:43:19.987275 136845296777024 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 01:43:19.987319 136845296777024 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 01:43:19.987640 136845296777024 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 0x7c753871d010>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 01:43:23.318122 136845296777024 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 01:43:23.318351 136845296777024 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 0x7c612427bec0>}, handler_registry=None
I0418 01:43:23.318588 136845296777024 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c612427bec0>` 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 01:43:23.318636 136845296777024 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c61241289b0>` 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 01:43:23.318670 136845296777024 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 0x7c612427bec0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c612427bec0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c61241289b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c61241289b0>}).
I0418 01:43:23.318995 136845296777024 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 01:43:23.319065 136845296777024 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c610c7c7c40> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 01:43:24.445107 136845296777024 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints
I0418 01:43:25.727373 136845296777024 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_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c612411f5c0>
I0418 01:43:25.727548 136845296777024 checkpointing.py:301] Checkpoint manager created!
I0418 01:43:26.693775 136845296777024 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 01:43:26.693905 136845296777024 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 01:43:27.070868 136845296777024 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 01:43:27.070956 136845296777024 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 01:43:27.087166 136845296777024 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 01:43:27.087227 136845296777024 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 01:43:27.110860 136845296777024 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 01:43:27.110924 136845296777024 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:43:27.127101 136845296777024 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 01:43:27.127159 136845296777024 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:43:27.143325 136845296777024 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 01:43:27.143386 136845296777024 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:43:27.168710 136845296777024 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 01:43:27.168781 136845296777024 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:43:27.189537 136845296777024 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 01:43:27.189601 136845296777024 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 01:43:27.403407 136845296777024 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 01:43:27.403508 136845296777024 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 01:43:28.834660 136845296777024 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 01:43:28.834787 136845296777024 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 01:43:28.834868 136845296777024 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 01:43:28.834933 136845296777024 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 01:43:28.834987 136845296777024 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 01:43:28.835057 136845296777024 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 01:43:28.835112 136845296777024 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 01:43:28.835174 136845296777024 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 01:43:28.835238 136845296777024 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 01:43:28.835283 136845296777024 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 01:43:28.835334 136845296777024 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 01:43:28.835383 136845296777024 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 01:43:29.325477 136845296777024 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 01:43:29.325573 136845296777024 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 01:43:29.341039 136845296777024 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 01:43:29.341103 136845296777024 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 01:43:39.772479 136845296777024 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 01:43:39.780508 136845296777024 metric_logger.py:301] number parameters: 1.104 billion
I0418 01:43:51.313166 136845296777024 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 01:43:51.466973 136845296777024 checkpointing.py:776] Waited 0.15379095077514648 seconds for step 0 to finish before starting checkpointing.
I0418 01:43:51.469298 136845296777024 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 01:43:51.470992 136845296777024 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 01:43:51.472450 136845296777024 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0.
I0418 01:43:52.211630 136845296777024 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 01:43:52.226067 136845296777024 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 01:43:52.226135 136845296777024 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 01:43:52.495053 136716652373760 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0
I0418 01:43:52.497039 136845296777024 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271713s
I0418 01:43:52.497945 136845296777024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.543 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27829408645629883 s) (per-host)
I0418 01:43:52.498030 136845296777024 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278399s (batch_requests_ready=0.002278s, total_serialization_initiated=0.275297s, others=0.000824s)
I0418 01:43:52.498117 136845296777024 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.282466s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.282448)
I0418 01:43:52.498931 136845296777024 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0.
I0418 01:43:52.499261 136716660766464 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:03:52.499225
I0418 01:43:52.509308 136845296777024 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 01:43:52.509583 136715534763776 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 01:43:52.509742 136845296777024 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776476631.46928, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776476631.4710288, 'checkpointer_blocking_duration_secs': 1.0283820629119873, 'get_old_steps_start_time': 1776476632.4994376, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776476631.4675076, 'checkpoint_manager_blocking_duration_secs': 1.0421957969665527}
I0418 01:43:52.509859 136845296777024 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 01:43:52.509908 136845296777024 max_utils.py:750]
Memstats: After params initialized:
I0418 01:43:52.509958 136845296777024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 01:43:52.509990 136845296777024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 01:43:52.510015 136845296777024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 01:43:52.510040 136845296777024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 01:43:52.848253 136845296777024 metric_logger.py:196] completed step: 0, seconds: 11.242, TFLOP/s/device: 1.209, Tokens/s/device: 182.176, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 01:43:52.851376 136845296777024 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/tensorboard/'
I0418 01:43:53.236274 136845296777024 metric_logger.py:196] completed step: 1, seconds: 1.533, TFLOP/s/device: 8.862, Tokens/s/device: 1335.762, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 01:43:53.382296 136845296777024 metric_logger.py:196] completed step: 2, seconds: 0.389, TFLOP/s/device: 34.950, Tokens/s/device: 5268.099, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0418 01:43:53.529620 136845296777024 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1436.879, Tokens/s/device: 216582.064, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0418 01:43:53.538790 136845296777024 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0418 01:43:53.823626 136845296777024 checkpointing.py:776] Waited 0.28469181060791016 seconds for step 5 to finish before starting checkpointing.
I0418 01:43:53.826556 136845296777024 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 01:43:54.964713 136716635588352 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/items
I0418 01:43:55.138808 136714464392960 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776476634711418923, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776476634711418923, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA
I0418 01:43:56.094169 2817 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 01:43:58.703556 136716094531328 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_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/items/array_metadatas/process_0
I0418 01:44:34.298523 136715580913408 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.214216s (commit=36.171484s, array_metadata_write=2.042732s)
I0418 01:44:34.300084 136716660766464 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.538 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.080384492874146 s) (per-host)
I0418 01:44:34.300244 136716660766464 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.800869s.
I0418 01:44:35.379354 136716660766464 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776476634711418923, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA
I0418 01:44:36.114973 136714464392960 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776476634711418923, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA
I0418 01:44:36.526367 136716660766464 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000364s.
I0418 01:44:37.190711 136716660766464 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 01:44:38.109725 136716660766464 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.608683s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/items
I0418 01:44:38.111827 136716660766464 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0/items
I0418 01:44:38.634534 136716660766464 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0
I0418 01:44:40.868680 136716660766464 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0`.
I0418 01:44:40.869887 136716660766464 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.40 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/0
I0418 01:44:40.871407 136716660766464 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.372029s.
I0418 01:44:40.871690 136715534763776 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 01:44:40.871866 136715534763776 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 01:44:40.871989 136715534763776 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 01:44:40.874353 136715534763776 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 01:44:40.874656 136845296777024 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 01:44:40.874869 136845296777024 checkpoint_manager.py:1452] Waiting for previous save to complete took 47.048284 seconds. If this number is high, consider checkpointing less frequently.
I0418 01:44:40.876879 136845296777024 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5
I0418 01:44:40.879313 136845296777024 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5.
I0418 01:44:41.618178 136845296777024 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 01:44:41.618312 136845296777024 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 01:44:41.657008 136845296777024 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039840s
I0418 01:44:41.657964 136845296777024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 31.891 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.048370361328125 s) (per-host)
I0418 01:44:41.658069 136845296777024 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.048499s (batch_requests_ready=0.004010s, total_serialization_initiated=0.043625s, others=0.000865s)
I0418 01:44:41.658198 136845296777024 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.053483s (all_items=0.000025s, per_item={'items': '0.00002456'}, temp_paths=0.053458)
I0418 01:44:41.659222 136845296777024 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.78 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5.
I0418 01:44:41.659586 136715534763776 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:04:41.659550
I0418 01:44:41.661993 136845296777024 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 01:44:41.662379 136716107110144 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 01:44:41.662612 136845296777024 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776476633.8265097, 'wait_for_prev_duration_secs': 47.04828405380249, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776476680.8769453, 'checkpointer_blocking_duration_secs': 0.7828340530395508, 'get_old_steps_start_time': 1776476681.6598072, 'get_old_steps_duration_secs': 6.747245788574219e-05, 'checkpoint_manager_blocking_start_time': 1776476633.824306, 'checkpoint_manager_blocking_duration_secs': 47.8382523059845}
I0418 01:44:41.662757 136845296777024 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0418 01:44:41.852104 136716660766464 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5
I0418 01:44:44.057249 136716094531328 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5/items
I0418 01:44:46.890635 136716123895552 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_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5/items/array_metadatas/process_0
I0418 01:45:06.687560 136845296777024 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.554, Tokens/s/device: 13950.764, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0418 01:45:06.697436 136845296777024 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.254, Tokens/s/device: 13905.581, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0418 01:45:06.842786 136845296777024 metric_logger.py:196] completed step: 6, seconds: 73.157, TFLOP/s/device: 0.186, Tokens/s/device: 27.995, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0418 01:45:06.990066 136845296777024 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1905.628, Tokens/s/device: 287237.027, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0418 01:45:07.137679 136845296777024 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.716, Tokens/s/device: 13975.134, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0418 01:45:07.284599 136845296777024 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 01:45:07.285379 136845296777024 checkpointing.py:776] Waited 0.0007970333099365234 seconds for step 9 to finish before starting checkpointing.
I0418 01:45:07.287636 136845296777024 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 01:45:28.866918 136716115502848 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.702832s (commit=41.248107s, array_metadata_write=2.454726s)
I0418 01:45:28.868051 136715534763776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.425 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.258439779281616 s) (per-host)
I0418 01:45:28.868182 136715534763776 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.208437s.
I0418 01:45:32.031184 136715534763776 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000250s.
I0418 01:45:32.323422 136715534763776 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 01:45:33.080250 136715534763776 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.023683s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5/items
I0418 01:45:33.081984 136715534763776 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5/items
I0418 01:45:33.614391 136715534763776 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5
I0418 01:45:35.488970 136715534763776 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5`.
I0418 01:45:35.489873 136715534763776 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 54.61 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/5
I0418 01:45:35.491566 136715534763776 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 53.831819s.
I0418 01:45:35.491798 136716107110144 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 01:45:35.491966 136716107110144 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 01:45:35.492056 136716107110144 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0418 01:45:35.493907 136716107110144 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0418 01:45:35.494102 136845296777024 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 01:45:35.494251 136845296777024 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.206620 seconds. If this number is high, consider checkpointing less frequently.
I0418 01:45:35.495927 136845296777024 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 01:45:35.497936 136845296777024 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9.
I0418 01:45:36.216005 136845296777024 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 01:45:36.216103 136845296777024 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 01:45:36.241990 136845296777024 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026973s
I0418 01:45:36.242862 136845296777024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.847 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03157973289489746 s) (per-host)
I0418 01:45:36.242945 136845296777024 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031692s (batch_requests_ready=0.001785s, total_serialization_initiated=0.029111s, others=0.000796s)
I0418 01:45:36.243034 136845296777024 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035746s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.035731)
I0418 01:45:36.243749 136845296777024 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9.
I0418 01:45:36.244086 136716107110144 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:05:36.244048
I0418 01:45:36.251495 136845296777024 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 01:45:36.251775 136716660766464 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 01:45:36.251925 136845296777024 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776476707.2876003, 'wait_for_prev_duration_secs': 28.20661997795105, 'time_between_consecutive_saves_sec': 26.41317892074585, 'checkpointer_blocking_start_time': 1776476735.4959722, 'checkpointer_blocking_duration_secs': 0.7482542991638184, 'get_old_steps_start_time': 1776476736.244246, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776476707.2856421, 'checkpoint_manager_blocking_duration_secs': 28.966248035430908}
I0418 01:45:36.252085 136845296777024 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 01:45:36.252131 136845296777024 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 01:45:36.444823 136715534763776 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9
I0418 01:45:39.445194 136716094531328 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9/items
I0418 01:45:42.826780 136717717878528 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_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9/items/array_metadatas/process_0
I0418 01:46:23.400462 136717197637376 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.392990s (commit=40.771542s, array_metadata_write=1.621448s)
I0418 01:46:23.401565 136716107110144 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.473 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.190263748168945 s) (per-host)
I0418 01:46:23.401688 136716107110144 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.157489s.
I0418 01:46:27.837199 136716107110144 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 01:46:28.299412 136716107110144 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000254s.
I0418 01:46:28.878814 136716107110144 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.906887s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9/items
I0418 01:46:28.880382 136716107110144 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9/items
I0418 01:46:30.253790 136716107110144 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9
I0418 01:46:33.079118 136716107110144 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9`.
I0418 01:46:33.079829 136716107110144 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 57.58 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_true/checkpoints/9
I0418 01:46:33.081228 136716107110144 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 56.837026s.
I0418 01:46:33.081446 136716660766464 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 01:46:33.081565 136716660766464 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 01:46:33.081616 136716660766464 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 01:46:33.083609 136716660766464 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 01:46:33.083750 136845296777024 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 01:46:33.083832 136845296777024 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 01:46:33.084800 136845296777024 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.229, Tokens/s/device: 13901.805, 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 01:46:43 UTC 2026
EXIT_CODE=0