XPK Start: Sat Apr 18 01:24:19 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 01:24:43.528766: 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 01:24:44.063724 137264542619456 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 01:24:53.104918 137264542619456 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 01:24:53.107288 137264542619456 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-jjeq5-slice-job-0-0.mt-15-ocdbt-true-sv-jjeq5:8482 I0418 01:24:54.059667 137264542619456 max_utils.py:284] Jax distributed system initialized! I0418 01:25:00.487869 137264542619456 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 01:25:00.487971 137264542619456 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 01:25:00.488011 137264542619456 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 01:25:00.488046 137264542619456 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 01:25:01.174411 137264542619456 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 01:25:01.174682 137264542619456 checkpointing.py:677] Setting up checkpoint logger... I0418 01:25:01.174736 137264542619456 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 01:25:01.174780 137264542619456 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 01:25:01.175156 137264542619456 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 0x7cd6c37f2600>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 01:25:04.566096 137264542619456 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 01:25:04.566331 137264542619456 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 0x7cc2f43886e0>}, handler_registry=None I0418 01:25:04.566573 137264542619456 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cc2f43886e0>` 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 01:25:04.566622 137264542619456 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cc2bc740fe0>` 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 01:25:04.566658 137264542619456 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 0x7cc2f43886e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cc2f43886e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cc2bc740fe0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cc2bc740fe0>}). I0418 01:25:04.566984 137264542619456 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 01:25:04.567053 137264542619456 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7cc2bc4afc40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 01:25:06.795861 137264542619456 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints I0418 01:25:07.281636 137264542619456 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_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cc2bc740350> I0418 01:25:07.281805 137264542619456 checkpointing.py:301] Checkpoint manager created! I0418 01:25:08.190171 137264542619456 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 01:25:08.190284 137264542619456 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 01:25:08.570067 137264542619456 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 01:25:08.570162 137264542619456 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 01:25:08.586411 137264542619456 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 01:25:08.586474 137264542619456 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 01:25:08.610159 137264542619456 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 01:25:08.610226 137264542619456 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:25:08.626549 137264542619456 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 01:25:08.626607 137264542619456 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:25:08.642924 137264542619456 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 01:25:08.642984 137264542619456 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:25:08.668280 137264542619456 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 01:25:08.668354 137264542619456 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:25:08.689346 137264542619456 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 01:25:08.689411 137264542619456 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 01:25:08.908134 137264542619456 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 01:25:08.908243 137264542619456 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 01:25:10.334763 137264542619456 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 01:25:10.334912 137264542619456 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 01:25:10.334973 137264542619456 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 01:25:10.335061 137264542619456 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 01:25:10.335154 137264542619456 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 01:25:10.335215 137264542619456 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 01:25:10.335291 137264542619456 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 01:25:10.335367 137264542619456 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 01:25:10.335427 137264542619456 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 01:25:10.335481 137264542619456 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 01:25:10.335545 137264542619456 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 01:25:10.335607 137264542619456 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 01:25:10.821423 137264542619456 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 01:25:10.821515 137264542619456 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 01:25:10.836767 137264542619456 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 01:25:10.836828 137264542619456 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 01:25:21.834348 137264542619456 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 01:25:21.842042 137264542619456 metric_logger.py:301] number parameters: 1.104 billion I0418 01:25:33.413208 137264542619456 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0418 01:25:33.567524 137264542619456 checkpointing.py:776] Waited 0.1542978286743164 seconds for step 0 to finish before starting checkpointing. I0418 01:25:33.569926 137264542619456 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 01:25:33.571476 137264542619456 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0418 01:25:33.572869 137264542619456 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0. I0418 01:25:34.315217 137264542619456 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 01:25:34.329539 137264542619456 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:25:34.329606 137264542619456 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 01:25:34.575811 137137056532224 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0 I0418 01:25:34.597409 137264542619456 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.268617s I0418 01:25:34.598250 137264542619456 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.608 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2750678062438965 s) (per-host) I0418 01:25:34.598322 137264542619456 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.275154s (batch_requests_ready=0.002241s, total_serialization_initiated=0.272140s, others=0.000773s) I0418 01:25:34.598402 137264542619456 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.279138s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.279120) I0418 01:25:34.599192 137264542619456 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_15_ocdbt_true/checkpoints/0. I0418 01:25:34.599525 137137580484352 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:45:34.599487 I0418 01:25:34.611671 137264542619456 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:25:34.611931 137135432644352 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:25:34.612048 137264542619456 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776475533.569903, 'wait_for_prev_duration_secs': 6.961822509765625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776475533.5715146, 'checkpointer_blocking_duration_secs': 1.0281572341918945, 'get_old_steps_start_time': 1776475534.5996969, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776475533.568047, 'checkpoint_manager_blocking_duration_secs': 1.0439631938934326} I0418 01:25:34.612152 137264542619456 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0418 01:25:34.612205 137264542619456 max_utils.py:750] Memstats: After params initialized: I0418 01:25:34.612252 137264542619456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0418 01:25:34.612285 137264542619456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0418 01:25:34.612311 137264542619456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0418 01:25:34.612336 137264542619456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0418 01:25:34.942265 137264542619456 metric_logger.py:196] completed step: 0, seconds: 11.282, TFLOP/s/device: 1.204, Tokens/s/device: 181.524, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 01:25:34.945382 137264542619456 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/tensorboard/' I0418 01:25:35.347887 137264542619456 metric_logger.py:196] completed step: 1, seconds: 1.527, TFLOP/s/device: 8.896, Tokens/s/device: 1340.940, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 01:25:35.493970 137264542619456 metric_logger.py:196] completed step: 2, seconds: 0.406, TFLOP/s/device: 33.443, Tokens/s/device: 5040.821, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0418 01:25:35.641396 137264542619456 metric_logger.py:196] completed step: 3, seconds: 0.008, TFLOP/s/device: 1613.099, Tokens/s/device: 243143.773, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0418 01:25:35.650722 137264542619456 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0418 01:25:35.935530 137264542619456 checkpointing.py:776] Waited 0.2846798896789551 seconds for step 5 to finish before starting checkpointing. I0418 01:25:35.938481 137264542619456 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:25:37.196945 137137039746816 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0/items I0418 01:25:37.388060 137133307283200 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776475536940403603, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776475536940403603, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0418 01:25:38.304722 2798 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 01:25:40.865721 137135449429760 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_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_0 I0418 01:26:16.398192 137135441037056 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.089270s (commit=36.027743s, array_metadata_write=2.061527s) I0418 01:26:16.399681 137137580484352 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.541 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.07645654678345 s) (per-host) I0418 01:26:16.399861 137137580484352 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.800221s. I0418 01:26:17.581292 137137580484352 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776475536940403603, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0418 01:26:18.312111 137133307283200 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776475536940403603, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0418 01:26:18.730645 137137580484352 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000395s. I0418 01:26:19.133319 137137580484352 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 01:26:20.081209 137137580484352 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.375355s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0/items I0418 01:26:20.083300 137137580484352 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0/items I0418 01:26:21.046703 137137580484352 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0 I0418 01:26:22.421551 137137580484352 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_15_ocdbt_true/checkpoints/0`. I0418 01:26:22.422671 137137580484352 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.85 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/0 I0418 01:26:22.424517 137137580484352 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.824873s. I0418 01:26:22.424802 137135432644352 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:26:22.424981 137135432644352 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:26:22.425092 137135432644352 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:26:22.427191 137135432644352 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 01:26:22.427418 137264542619456 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 01:26:22.427549 137264542619456 checkpoint_manager.py:1452] Waiting for previous save to complete took 46.489077 seconds. If this number is high, consider checkpointing less frequently. I0418 01:26:22.429404 137264542619456 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5 I0418 01:26:22.431711 137264542619456 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5. I0418 01:26:23.146435 137264542619456 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:26:23.146568 137264542619456 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 01:26:23.186281 137264542619456 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040896s I0418 01:26:23.187187 137264542619456 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 31.242 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.049375057220458984 s) (per-host) I0418 01:26:23.187296 137264542619456 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.049506s (batch_requests_ready=0.004040s, total_serialization_initiated=0.044659s, others=0.000807s) I0418 01:26:23.187425 137264542619456 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.054403s (all_items=0.000026s, per_item={'items': '0.00002623'}, temp_paths=0.054377) I0418 01:26:23.188380 137264542619456 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5. I0418 01:26:23.188942 137135432644352 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:46:23.188901 I0418 01:26:23.191365 137264542619456 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:26:23.191738 137135462008576 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:26:23.191946 137264542619456 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776475535.9384332, 'wait_for_prev_duration_secs': 46.48907661437988, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776475582.4294722, 'checkpointer_blocking_duration_secs': 0.7596678733825684, 'get_old_steps_start_time': 1776475583.1891665, 'get_old_steps_duration_secs': 4.172325134277344e-05, 'checkpoint_manager_blocking_start_time': 1776475535.9361918, 'checkpoint_manager_blocking_duration_secs': 47.255698919296265} I0418 01:26:23.192166 137264542619456 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0418 01:26:23.386717 137137580484352 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5 I0418 01:26:27.288247 137132744460032 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5/items I0418 01:26:32.281413 137135478793984 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_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_0 I0418 01:26:48.336799 137264542619456 metric_logger.py:196] completed step: 4, seconds: 0.148, TFLOP/s/device: 91.941, Tokens/s/device: 13858.344, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0418 01:26:48.346191 137264542619456 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.038, Tokens/s/device: 13872.895, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0418 01:26:48.491607 137264542619456 metric_logger.py:196] completed step: 6, seconds: 72.694, TFLOP/s/device: 0.187, Tokens/s/device: 28.173, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0418 01:26:48.639085 137264542619456 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1921.800, Tokens/s/device: 289674.682, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0418 01:26:48.786227 137264542619456 metric_logger.py:196] completed step: 8, seconds: 0.146, TFLOP/s/device: 92.794, Tokens/s/device: 13986.874, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0418 01:26:48.932940 137264542619456 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0418 01:26:48.933651 137264542619456 checkpointing.py:776] Waited 0.0007307529449462891 seconds for step 9 to finish before starting checkpointing. I0418 01:26:48.936094 137264542619456 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:27:12.940157 137135470401280 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 44.120059s (commit=41.239269s, array_metadata_write=2.880790s) I0418 01:27:12.941295 137135432644352 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 31.717 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 49.803462743759155 s) (per-host) I0418 01:27:12.941419 137135432644352 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 49.752321s. I0418 01:27:17.226109 137135432644352 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 01:27:17.805190 137135432644352 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000243s. I0418 01:27:18.176434 137135432644352 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.633531s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5/items I0418 01:27:18.178012 137135432644352 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5/items I0418 01:27:19.990321 137135432644352 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5 I0418 01:27:23.254601 137135432644352 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_15_ocdbt_true/checkpoints/5`. I0418 01:27:23.255454 137135432644352 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 60.83 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/5 I0418 01:27:23.256714 137135432644352 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 60.067614s. I0418 01:27:23.256940 137135462008576 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:27:23.257062 137135462008576 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:27:23.257132 137135462008576 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:27:23.259576 137135462008576 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0418 01:27:23.259784 137264542619456 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 01:27:23.259927 137264542619456 checkpoint_manager.py:1452] Waiting for previous save to complete took 34.323841 seconds. If this number is high, consider checkpointing less frequently. I0418 01:27:23.261280 137264542619456 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0418 01:27:23.263360 137264542619456 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9. I0418 01:27:24.407047 137264542619456 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:27:24.407143 137264542619456 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 01:27:24.433672 137264542619456 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027560s I0418 01:27:24.434480 137264542619456 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.200 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03200340270996094 s) (per-host) I0418 01:27:24.434550 137264542619456 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032090s (batch_requests_ready=0.001861s, total_serialization_initiated=0.029475s, others=0.000755s) I0418 01:27:24.434621 137264542619456 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036195s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.036176) I0418 01:27:24.435346 137264542619456 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9. I0418 01:27:24.435637 137135985252096 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:47:24.435593 I0418 01:27:24.443319 137264542619456 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:27:24.443584 137135462008576 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:27:24.443735 137264542619456 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776475608.9360595, 'wait_for_prev_duration_secs': 34.323840618133545, 'time_between_consecutive_saves_sec': 26.508800268173218, 'checkpointer_blocking_start_time': 1776475643.261319, 'checkpointer_blocking_duration_secs': 1.174457311630249, 'get_old_steps_start_time': 1776475644.4357953, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776475608.9339106, 'checkpoint_manager_blocking_duration_secs': 35.50978946685791} I0418 01:27:24.443861 137264542619456 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0418 01:27:24.443906 137264542619456 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:27:24.657172 137135432644352 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9 I0418 01:27:30.678287 137137580484352 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9/items I0418 01:27:36.966365 137137601877760 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_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_0 I0418 01:28:13.645989 137136552535808 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.575765s (commit=39.403185s, array_metadata_write=1.172579s) I0418 01:28:13.647108 137135985252096 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 32.077 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 49.24459362030029 s) (per-host) I0418 01:28:13.647242 137135985252096 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 49.211489s. I0418 01:28:16.172722 137135985252096 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 01:28:16.363513 137135985252096 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000221s. I0418 01:28:17.116914 137135985252096 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.191444s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9/items I0418 01:28:17.118619 137135985252096 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9/items I0418 01:28:17.627686 137135985252096 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9 I0418 01:28:19.004020 137135985252096 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_15_ocdbt_true/checkpoints/9`. I0418 01:28:19.004880 137135985252096 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 55.74 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_true/checkpoints/9 I0418 01:28:19.006490 137135985252096 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 54.570735s. I0418 01:28:19.006701 137135462008576 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:28:19.006815 137135462008576 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:28:19.006884 137135462008576 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:28:19.009093 137135462008576 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 01:28:19.009268 137264542619456 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 01:28:19.009417 137264542619456 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 01:28:19.010302 137264542619456 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.196, Tokens/s/device: 13896.805, 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 01:28:30 UTC 2026 EXIT_CODE=0