XPK Start: Fri Apr 17 14:30:32 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 14:30:56.077493: 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 14:30:56.255972 138022316984128 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 14:31:05.298637 138022316984128 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 14:31:05.301078 138022316984128 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-7kwxg-slice-job-0-0.mt-15-ocdbt-true-sv-7kwxg:8482
I0417 14:31:06.432798 138022316984128 max_utils.py:284] Jax distributed system initialized!
I0417 14:31:12.489245 138022316984128 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 14:31:12.489348 138022316984128 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 14:31:12.489387 138022316984128 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 14:31:12.489421 138022316984128 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 14:31:13.507595 138022316984128 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 14:31:13.507889 138022316984128 checkpointing.py:677] Setting up checkpoint logger...
I0417 14:31:13.507947 138022316984128 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0417 14:31:13.507992 138022316984128 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 14:31:13.508315 138022316984128 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 0x7d87353145f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 14:31:16.120334 138022316984128 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 14:31:16.120665 138022316984128 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 0x7d860fac5a30>}, handler_registry=None
I0417 14:31:16.120928 138022316984128 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d860fac5a30>` 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 14:31:16.120978 138022316984128 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d860fb2f980>` 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 14:31:16.121015 138022316984128 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 0x7d860fac5a30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d860fac5a30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d860fb2f980>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d860fb2f980>}).
I0417 14:31:16.121332 138022316984128 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 14:31:16.121401 138022316984128 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d72d472c180> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 14:31:17.666244 138022316984128 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints
I0417 14:31:18.118571 138022316984128 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_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d860fb1f320>
I0417 14:31:18.118742 138022316984128 checkpointing.py:301] Checkpoint manager created!
I0417 14:31:19.041386 138022316984128 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 14:31:19.041498 138022316984128 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:31:19.417408 138022316984128 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:31:19.417495 138022316984128 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:31:19.433665 138022316984128 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:31:19.433723 138022316984128 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:31:19.457275 138022316984128 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 14:31:19.457339 138022316984128 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:31:19.473438 138022316984128 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 14:31:19.473494 138022316984128 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:31:19.489473 138022316984128 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 14:31:19.489534 138022316984128 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:31:19.513753 138022316984128 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 14:31:19.513820 138022316984128 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:31:19.535380 138022316984128 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 14:31:19.535441 138022316984128 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 14:31:19.747558 138022316984128 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 14:31:19.747663 138022316984128 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0417 14:31:21.156547 138022316984128 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0417 14:31:21.156668 138022316984128 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 14:31:21.156718 138022316984128 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 14:31:21.156775 138022316984128 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 14:31:21.156825 138022316984128 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 14:31:21.156881 138022316984128 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 14:31:21.156940 138022316984128 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 14:31:21.156992 138022316984128 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 14:31:21.157033 138022316984128 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 14:31:21.157070 138022316984128 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 14:31:21.157122 138022316984128 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0417 14:31:21.157173 138022316984128 maxtext_utils.py:1651] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0417 14:31:21.642314 138022316984128 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:31:21.642406 138022316984128 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:31:21.657599 138022316984128 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:31:21.657664 138022316984128 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:31:32.508034 138022316984128 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 14:31:32.515949 138022316984128 metric_logger.py:301] number parameters: 1.104 billion
I0417 14:31:43.695360 138022316984128 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0417 14:31:43.848996 138022316984128 checkpointing.py:776] Waited 0.15361595153808594 seconds for step 0 to finish before starting checkpointing.
I0417 14:31:43.851371 138022316984128 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:31:43.853057 138022316984128 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0417 14:31:43.854536 138022316984128 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0.
I0417 14:31:44.580607 138022316984128 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 14:31:44.595345 138022316984128 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:31:44.595410 138022316984128 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 14:31:44.845532 137893915674368 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0
I0417 14:31:44.875198 138022316984128 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.280574s
I0417 14:31:44.876065 138022316984128 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.368 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2873818874359131 s) (per-host)
I0417 14:31:44.876138 138022316984128 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.287472s (batch_requests_ready=0.002592s, total_serialization_initiated=0.284091s, others=0.000789s)
I0417 14:31:44.876223 138022316984128 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.291543s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.291525)
I0417 14:31:44.876979 138022316984128 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_15_ocdbt_true/checkpoints/0.
I0417 14:31:44.877331 137893924067072 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:51:44.877284
I0417 14:31:44.881487 138022316984128 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:31:44.881801 137893389518592 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:31:44.881980 138022316984128 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776436303.851353, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776436303.8530996, 'checkpointer_blocking_duration_secs': 1.0243988037109375, 'get_old_steps_start_time': 1776436304.8775253, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776436303.8495061, 'checkpoint_manager_blocking_duration_secs': 1.0324366092681885}
I0417 14:31:44.882086 138022316984128 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0417 14:31:44.882140 138022316984128 max_utils.py:750]
Memstats: After params initialized:
I0417 14:31:44.882191 138022316984128 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 14:31:44.882224 138022316984128 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 14:31:44.882251 138022316984128 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 14:31:44.882276 138022316984128 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 14:31:45.233886 138022316984128 metric_logger.py:196] completed step: 0, seconds: 10.885, TFLOP/s/device: 1.248, Tokens/s/device: 188.142, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0417 14:31:45.237303 138022316984128 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/tensorboard/'
I0417 14:31:45.638913 138022316984128 metric_logger.py:196] completed step: 1, seconds: 1.537, TFLOP/s/device: 8.842, Tokens/s/device: 1332.709, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0417 14:31:45.784825 138022316984128 metric_logger.py:196] completed step: 2, seconds: 0.406, TFLOP/s/device: 33.490, Tokens/s/device: 5048.003, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0417 14:31:45.932167 138022316984128 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1492.927, Tokens/s/device: 225030.216, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0417 14:31:45.941443 138022316984128 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0417 14:31:46.226406 138022316984128 checkpointing.py:776] Waited 0.2848188877105713 seconds for step 5 to finish before starting checkpointing.
I0417 14:31:46.229364 138022316984128 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:31:47.026544 137893898888960 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/items
I0417 14:31:47.181521 137888016623360 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776436306772144251, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776436306772144251, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA
I0417 14:31:48.143307 2761 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 14:31:49.938064 137893406304000 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_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_0
I0417 14:32:23.195654 137893397911296 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.055339s (commit=33.814744s, array_metadata_write=1.240595s)
I0417 14:32:23.197052 137893924067072 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 40.913 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.608344316482544 s) (per-host)
I0417 14:32:23.197163 137893924067072 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.319701s.
I0417 14:32:23.892889 137893924067072 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776436306772144251, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA
I0417 14:32:24.661473 137888016623360 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776436306772144251, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA
I0417 14:32:24.906157 137893924067072 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:32:24.940822 137893924067072 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000383s.
I0417 14:32:25.536477 137893924067072 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.524448s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/items
I0417 14:32:25.538638 137893924067072 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/items
I0417 14:32:26.052996 137893924067072 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0
I0417 14:32:27.879556 137893924067072 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_15_ocdbt_true/checkpoints/0`.
I0417 14:32:27.880532 137893924067072 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 44.03 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0
I0417 14:32:27.882093 137893924067072 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 43.004627s.
I0417 14:32:27.882424 137893389518592 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:32:27.882597 137893389518592 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:32:27.882703 137893389518592 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:32:27.884325 137893389518592 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0417 14:32:27.884602 138022316984128 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.
W0417 14:32:27.884720 138022316984128 checkpoint_manager.py:1452] Waiting for previous save to complete took 41.655367 seconds. If this number is high, consider checkpointing less frequently.
I0417 14:32:27.886597 138022316984128 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5
I0417 14:32:27.888979 138022316984128 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5.
I0417 14:32:29.057075 138022316984128 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:32:29.057213 138022316984128 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 14:32:29.095483 138022316984128 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039476s
I0417 14:32:29.096537 138022316984128 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 32.168 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04795384407043457 s) (per-host)
I0417 14:32:29.096672 138022316984128 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.048119s (batch_requests_ready=0.003942s, total_serialization_initiated=0.043216s, others=0.000960s)
I0417 14:32:29.096818 138022316984128 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.053332s (all_items=0.000027s, per_item={'items': '0.00002718'}, temp_paths=0.053305)
I0417 14:32:29.097867 138022316984128 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.21 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5.
I0417 14:32:29.098242 137893389518592 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:52:29.098206
I0417 14:32:29.100716 138022316984128 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:32:29.101102 137892293613312 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:32:29.101335 138022316984128 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776436306.2293186, 'wait_for_prev_duration_secs': 41.655367374420166, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776436347.8866603, 'checkpointer_blocking_duration_secs': 1.2117626667022705, 'get_old_steps_start_time': 1776436349.0984516, 'get_old_steps_duration_secs': 4.1484832763671875e-05, 'checkpoint_manager_blocking_start_time': 1776436306.2270653, 'checkpoint_manager_blocking_duration_secs': 42.8742151260376}
I0417 14:32:29.101479 138022316984128 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0417 14:32:29.295068 137893924067072 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5
I0417 14:32:31.025829 137892869433088 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5/items
I0417 14:32:35.265784 137892310398720 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_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_0
I0417 14:32:51.725194 138022316984128 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.124, Tokens/s/device: 13885.875, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0417 14:32:51.734249 138022316984128 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.463, Tokens/s/device: 13936.998, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0417 14:32:51.879749 138022316984128 metric_logger.py:196] completed step: 6, seconds: 65.791, TFLOP/s/device: 0.207, Tokens/s/device: 31.129, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0417 14:32:52.027217 138022316984128 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1913.950, Tokens/s/device: 288491.337, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0417 14:32:52.174488 138022316984128 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.730, Tokens/s/device: 13977.328, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0417 14:32:52.320935 138022316984128 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0417 14:32:52.321620 138022316984128 checkpointing.py:776] Waited 0.000705718994140625 seconds for step 9 to finish before starting checkpointing.
I0417 14:32:52.323672 138022316984128 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:33:13.365041 137892302006016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.352113s (commit=39.030505s, array_metadata_write=1.321608s)
I0417 14:33:13.367342 137893389518592 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.642 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.318737506866455 s) (per-host)
I0417 14:33:13.367472 137893389518592 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.269082s.
I0417 14:33:14.934422 137893389518592 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000256s.
I0417 14:33:15.184998 137893389518592 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:33:15.812314 137893389518592 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.521834s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5/items
I0417 14:33:15.814127 137893389518592 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5/items
I0417 14:33:16.380372 137893389518592 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5
I0417 14:33:17.795916 137893389518592 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_15_ocdbt_true/checkpoints/5`.
I0417 14:33:17.796722 137893389518592 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.91 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/5
I0417 14:33:17.798083 137893389518592 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.699692s.
I0417 14:33:17.798218 137892293613312 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:33:17.798273 137892293613312 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:33:17.798354 137892293613312 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:33:17.800249 137892293613312 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0417 14:33:17.800419 138022316984128 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.
W0417 14:33:17.800544 138022316984128 checkpoint_manager.py:1452] Waiting for previous save to complete took 25.476879 seconds. If this number is high, consider checkpointing less frequently.
I0417 14:33:17.802010 138022316984128 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0417 14:33:17.803981 138022316984128 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9.
I0417 14:33:18.540040 138022316984128 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:33:18.540138 138022316984128 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 14:33:18.565780 138022316984128 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026570s
I0417 14:33:18.566572 138022316984128 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.932 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030893564224243164 s) (per-host)
I0417 14:33:18.566646 138022316984128 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030988s (batch_requests_ready=0.001782s, total_serialization_initiated=0.028504s, others=0.000702s)
I0417 14:33:18.566728 138022316984128 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034939s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.034924)
I0417 14:33:18.567478 138022316984128 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9.
I0417 14:33:18.567781 137892293613312 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:53:18.567746
I0417 14:33:18.575601 138022316984128 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:33:18.575943 137893368579840 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:33:18.576115 138022316984128 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776436372.3236394, 'wait_for_prev_duration_secs': 25.47687864303589, 'time_between_consecutive_saves_sec': 24.439245223999023, 'checkpointer_blocking_start_time': 1776436397.8020504, 'checkpointer_blocking_duration_secs': 0.7659096717834473, 'get_old_steps_start_time': 1776436398.56798, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776436372.3218637, 'checkpoint_manager_blocking_duration_secs': 26.254212141036987}
I0417 14:33:18.576239 138022316984128 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0417 14:33:18.576283 138022316984128 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:33:18.787573 137893389518592 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9
I0417 14:33:20.561612 137893414696704 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9/items
I0417 14:33:23.604606 137895000131328 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_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_0
I0417 14:34:02.097286 137893924067072 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.324131s (commit=39.029543s, array_metadata_write=1.294588s)
I0417 14:34:02.098428 137892293613312 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 36.260 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.56271243095398 s) (per-host)
I0417 14:34:02.098563 137892293613312 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.530633s.
I0417 14:34:04.318355 137892293613312 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:34:04.496495 137892293613312 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000242s.
I0417 14:34:04.944985 137892293613312 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.539534s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9/items
I0417 14:34:04.946687 137892293613312 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9/items
I0417 14:34:05.519407 137892293613312 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9
I0417 14:34:07.846699 137892293613312 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_15_ocdbt_true/checkpoints/9`.
I0417 14:34:07.847517 137892293613312 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.05 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9
I0417 14:34:07.849044 137892293613312 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.281111s.
I0417 14:34:07.849191 137893368579840 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:34:07.849269 137893368579840 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:34:07.849322 137893368579840 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:34:07.851322 137893368579840 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0417 14:34:07.851468 138022316984128 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.
I0417 14:34:07.851610 138022316984128 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:34:07.852576 138022316984128 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.201, Tokens/s/device: 13897.465, 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: Fri Apr 17 14:34:18 UTC 2026
EXIT_CODE=0