XPK Start: Fri Apr 17 14:58:01 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 14:58:25.232659: 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 14:58:25.412262 132699222746944 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 14:58:34.454018 132699222746944 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 14:58:34.456340 132699222746944 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-o3qra-slice-job-0-0.mt-16-zarr3-false-sv-o3qra:8482
I0417 14:58:35.491609 132699222746944 max_utils.py:284] Jax distributed system initialized!
I0417 14:58:41.482849 132699222746944 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 14:58:41.482954 132699222746944 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 14:58:41.482993 132699222746944 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 14:58:41.483027 132699222746944 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 14:58:42.162251 132699222746944 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 14:58:42.162514 132699222746944 checkpointing.py:677] Setting up checkpoint logger...
I0417 14:58:42.162568 132699222746944 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0417 14:58:42.162612 132699222746944 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 14:58:42.162945 132699222746944 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 0x78afde1184d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 14:58:45.018424 132699222746944 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 14:58:45.018656 132699222746944 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 0x789b98307290>}, handler_registry=None
I0417 14:58:45.018912 132699222746944 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x789b98307290>` 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 14:58:45.018963 132699222746944 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x789b980eca70>` 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 14:58:45.019003 132699222746944 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 0x789b98307290>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x789b98307290>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x789b980eca70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x789b980eca70>}).
I0417 14:58:45.019328 132699222746944 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 14:58:45.019401 132699222746944 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x789b785b0a40> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 14:58:46.582377 132699222746944 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints
I0417 14:58:47.008466 132699222746944 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 0x789b980eaf00>
I0417 14:58:47.008645 132699222746944 checkpointing.py:301] Checkpoint manager created!
I0417 14:58:47.921893 132699222746944 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 14:58:47.922008 132699222746944 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:58:48.298122 132699222746944 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:58:48.298215 132699222746944 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:58:48.314275 132699222746944 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:58:48.314332 132699222746944 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:58:48.337824 132699222746944 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 14:58:48.337909 132699222746944 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:58:48.354054 132699222746944 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 14:58:48.354116 132699222746944 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:58:48.370131 132699222746944 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 14:58:48.370190 132699222746944 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:58:48.394554 132699222746944 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 14:58:48.394621 132699222746944 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:58:48.416323 132699222746944 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 14:58:48.416385 132699222746944 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 14:58:48.634903 132699222746944 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 14:58:48.635010 132699222746944 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0417 14:58:50.062247 132699222746944 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0417 14:58:50.062369 132699222746944 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 14:58:50.062422 132699222746944 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 14:58:50.062478 132699222746944 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 14:58:50.062531 132699222746944 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 14:58:50.062573 132699222746944 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 14:58:50.062627 132699222746944 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 14:58:50.062678 132699222746944 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 14:58:50.062716 132699222746944 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 14:58:50.062751 132699222746944 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 14:58:50.062798 132699222746944 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0417 14:58:50.062863 132699222746944 maxtext_utils.py:1651] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0417 14:58:50.546272 132699222746944 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:58:50.546367 132699222746944 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:58:50.561471 132699222746944 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:58:50.561530 132699222746944 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:59:01.360567 132699222746944 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 14:59:01.368326 132699222746944 metric_logger.py:301] number parameters: 1.104 billion
I0417 14:59:12.914681 132699222746944 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0417 14:59:13.068197 132699222746944 checkpointing.py:776] Waited 0.15349102020263672 seconds for step 0 to finish before starting checkpointing.
I0417 14:59:13.070858 132699222746944 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:59:13.072578 132699222746944 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0417 14:59:13.073865 132699222746944 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/0.
I0417 14:59:13.816482 132699222746944 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 14:59:13.830916 132699222746944 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:59:13.830981 132699222746944 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 14:59:14.099711 132699222746944 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.269569s
I0417 14:59:14.100551 132699222746944 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.588 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2760436534881592 s) (per-host)
I0417 14:59:14.100615 132699222746944 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.276126s (batch_requests_ready=0.002245s, total_serialization_initiated=0.273145s, others=0.000736s)
I0417 14:59:14.100691 132699222746944 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.280101s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.280083)
I0417 14:59:14.101452 132699222746944 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0.
I0417 14:59:14.101751 132571361441536 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:19:14.101716
I0417 14:59:14.118097 132699222746944 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:59:14.118367 132569214691072 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:59:14.118504 132699222746944 standard_logger.py:34] {'step': 0, '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': 1776437953.0708241, 'wait_for_prev_duration_secs': 7.62939453125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776437953.0726151, 'checkpointer_blocking_duration_secs': 1.029249668121338, 'get_old_steps_start_time': 1776437954.1018953, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776437953.068746, 'checkpoint_manager_blocking_duration_secs': 1.049717903137207}
I0417 14:59:14.118605 132699222746944 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0417 14:59:14.118653 132699222746944 max_utils.py:750]
Memstats: After params initialized:
I0417 14:59:14.118701 132699222746944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 14:59:14.118734 132699222746944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 14:59:14.118759 132699222746944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 14:59:14.118782 132699222746944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 14:59:14.455806 132699222746944 metric_logger.py:196] completed step: 0, seconds: 11.238, TFLOP/s/device: 1.209, Tokens/s/device: 182.233, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0417 14:59:14.458877 132699222746944 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/tensorboard/'
I0417 14:59:14.503480 132570835408640 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0
I0417 14:59:14.837180 132699222746944 metric_logger.py:196] completed step: 1, seconds: 1.539, TFLOP/s/device: 8.827, Tokens/s/device: 1330.452, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0417 14:59:14.983264 132699222746944 metric_logger.py:196] completed step: 2, seconds: 0.382, TFLOP/s/device: 35.558, Tokens/s/device: 5359.713, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0417 14:59:15.130647 132699222746944 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1590.626, Tokens/s/device: 239756.497, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0417 14:59:15.139144 132699222746944 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0417 14:59:15.424636 132699222746944 checkpointing.py:776] Waited 0.28533124923706055 seconds for step 5 to finish before starting checkpointing.
I0417 14:59:15.427587 132699222746944 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:59:16.719297 132570818623232 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0/items
I0417 14:59:17.306368 132564434024192 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437956468219271, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776437956468219271, "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/0/_CHECKPOINT_METADATA
I0417 14:59:17.893606 2832 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 14:59:19.642956 132569265862400 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/0/items/array_metadatas/process_0
I0417 14:59:59.389122 132569223083776 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.499383s (commit=40.266284s, array_metadata_write=1.233099s)
I0417 14:59:59.390696 132571361441536 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.666 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.56614542007446 s) (per-host)
I0417 14:59:59.390878 132571361441536 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.288477s.
I0417 15:00:00.572086 132571361441536 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437956468219271, 'commit_timestamp_nsecs': None, '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 15:00:01.313146 132564434024192 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437956468219271, '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/0/_CHECKPOINT_METADATA
I0417 15:00:01.470177 132571361441536 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000437s.
I0417 15:00:02.894262 132571361441536 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.189942s. 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/0/items
I0417 15:00:02.896576 132571361441536 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0/items
I0417 15:00:03.424153 132571361441536 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0
I0417 15:00:04.863794 132571361441536 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/0`.
I0417 15:00:04.865001 132571361441536 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.79 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/0
I0417 15:00:04.866666 132571361441536 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.764262s.
I0417 15:00:04.866970 132569214691072 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 15:00:04.867139 132569214691072 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 15:00:04.867260 132569214691072 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0417 15:00:04.868983 132569214691072 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0417 15:00:04.869329 132699222746944 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 15:00:04.869506 132699222746944 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.441924 seconds. If this number is high, consider checkpointing less frequently.
I0417 15:00:04.871364 132699222746944 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5
I0417 15:00:04.873666 132699222746944 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/5.
I0417 15:00:05.592717 132699222746944 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 15:00:05.592869 132699222746944 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 15:00:05.630309 132699222746944 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038677s
I0417 15:00:05.631291 132699222746944 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 32.628 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.047277212142944336 s) (per-host)
I0417 15:00:05.631394 132699222746944 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.047402s (batch_requests_ready=0.004057s, total_serialization_initiated=0.042449s, others=0.000896s)
I0417 15:00:05.631517 132699222746944 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.052347s (all_items=0.000026s, per_item={'items': '0.00002599'}, temp_paths=0.052321)
I0417 15:00:05.632576 132699222746944 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/5.
I0417 15:00:05.632977 132569214691072 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:20:05.632941
I0417 15:00:05.635492 132699222746944 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 15:00:05.635883 132569265862400 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 15:00:05.636129 132699222746944 standard_logger.py:34] {'step': 5, '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': 1776437955.4275389, 'wait_for_prev_duration_secs': 49.44192361831665, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776438004.871434, 'checkpointer_blocking_duration_secs': 0.7617306709289551, 'get_old_steps_start_time': 1776438005.6331933, 'get_old_steps_duration_secs': 4.267692565917969e-05, 'checkpoint_manager_blocking_start_time': 1776437955.4253209, 'checkpoint_manager_blocking_duration_secs': 50.21075177192688}
I0417 15:00:05.636284 132699222746944 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0417 15:00:05.825240 132571361441536 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/5
I0417 15:00:08.339534 132570331211520 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/5/items
I0417 15:00:11.668977 132569766741760 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/5/items/array_metadatas/process_0
I0417 15:00:28.403905 132699222746944 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.623, Tokens/s/device: 13961.130, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0417 15:00:28.412807 132699222746944 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.001, Tokens/s/device: 13867.447, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0417 15:00:28.558301 132699222746944 metric_logger.py:196] completed step: 6, seconds: 73.272, TFLOP/s/device: 0.185, Tokens/s/device: 27.951, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0417 15:00:28.705801 132699222746944 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1973.439, Tokens/s/device: 297458.243, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0417 15:00:28.853031 132699222746944 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.691, Tokens/s/device: 13971.416, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0417 15:00:28.999715 132699222746944 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0417 15:00:29.000497 132699222746944 checkpointing.py:776] Waited 0.0008027553558349609 seconds for step 9 to finish before starting checkpointing.
I0417 15:00:29.002465 132699222746944 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 15:00:52.228680 132569758349056 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.329413s (commit=40.686473s, array_metadata_write=1.642940s)
I0417 15:00:52.229937 132569214691072 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.864 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.645904541015625 s) (per-host)
I0417 15:00:52.230073 132569214691072 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.596935s.
I0417 15:00:54.164616 132569214691072 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000227s.
I0417 15:00:55.381287 132569214691072 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.189763s. 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/5/items
I0417 15:00:55.387598 132569214691072 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/5/items
I0417 15:00:55.932403 132569214691072 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/5
I0417 15:00:58.197927 132569214691072 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/5`.
I0417 15:00:58.198777 132569214691072 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.33 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/5
I0417 15:00:58.200267 132569214691072 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.567136s.
I0417 15:00:58.200468 132569265862400 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 15:00:58.200585 132569265862400 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 15:00:58.200650 132569265862400 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0417 15:00:58.203588 132569265862400 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0417 15:00:58.203758 132699222746944 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 15:00:58.203907 132699222746944 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.201446 seconds. If this number is high, consider checkpointing less frequently.
I0417 15:00:58.205505 132699222746944 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0417 15:00:58.207593 132699222746944 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/9.
I0417 15:00:58.932381 132699222746944 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 15:00:58.932477 132699222746944 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 15:00:58.958297 132699222746944 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026856s
I0417 15:00:58.959152 132699222746944 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.445 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031197786331176758 s) (per-host)
I0417 15:00:58.959239 132699222746944 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031306s (batch_requests_ready=0.001753s, total_serialization_initiated=0.028776s, others=0.000777s)
I0417 15:00:58.959326 132699222746944 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035326s (all_items=0.000031s, per_item={'items': '0.00003076'}, temp_paths=0.035295)
I0417 15:00:58.960185 132699222746944 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9.
I0417 15:00:58.960465 132569786046208 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:20:58.960426
I0417 15:00:58.973525 132699222746944 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 15:00:58.973792 132569265862400 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 15:00:58.973946 132699222746944 standard_logger.py:34] {'step': 9, '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': 1776438029.002431, 'wait_for_prev_duration_secs': 29.201446056365967, 'time_between_consecutive_saves_sec': 24.133383989334106, 'checkpointer_blocking_start_time': 1776438058.2055433, 'checkpointer_blocking_duration_secs': 0.7550582885742188, 'get_old_steps_start_time': 1776438058.9606192, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776438029.0007226, 'checkpoint_manager_blocking_duration_secs': 29.973188877105713}
I0417 15:00:58.974056 132699222746944 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0417 15:00:58.974097 132699222746944 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 15:00:59.176570 132569214691072 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9
I0417 15:01:01.346659 132569749956352 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9/items
I0417 15:01:04.613857 132570331211520 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/9/items/array_metadatas/process_0
I0417 15:01:44.579921 132570321630976 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.143483s (commit=40.485430s, array_metadata_write=1.658054s)
I0417 15:01:44.581039 132569786046208 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.600 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.653056621551514 s) (per-host)
I0417 15:01:44.581167 132569786046208 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.620590s.
I0417 15:01:46.109683 132569786046208 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000255s.
I0417 15:01:47.439538 132569786046208 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.951132s. 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/9/items
I0417 15:01:47.441178 132569786046208 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9/items
I0417 15:01:47.972278 132569786046208 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9
I0417 15:01:49.386942 132569786046208 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/9`.
I0417 15:01:49.387769 132569786046208 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.18 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_false/checkpoints/9
I0417 15:01:49.389269 132569786046208 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.428678s.
I0417 15:01:49.389548 132569265862400 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 15:01:49.389670 132569265862400 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 15:01:49.389721 132569265862400 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0417 15:01:49.391415 132569265862400 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0417 15:01:49.391579 132699222746944 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 15:01:49.391726 132699222746944 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 15:01:49.392651 132699222746944 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.717, Tokens/s/device: 13824.574, 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 15:02:06 UTC 2026
EXIT_CODE=0