XPK Start: Fri Apr 17 13:35:25 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 13:35:49.487251: 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 13:35:49.665647 135324625303360 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 13:35:58.705704 135324625303360 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 13:35:58.708020 135324625303360 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--i6fpu-slice-job-0-0.mt-08-checkpoint-async--i6fpu:8482 I0417 13:35:59.741573 135324625303360 max_utils.py:284] Jax distributed system initialized! I0417 13:36:05.340727 135324625303360 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 13:36:05.340829 135324625303360 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 13:36:05.340894 135324625303360 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 13:36:05.340928 135324625303360 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 13:36:06.006956 135324625303360 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 13:36:06.007223 135324625303360 checkpointing.py:677] Setting up checkpoint logger... I0417 13:36:06.007276 135324625303360 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 13:36:06.007318 135324625303360 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 13:36:06.007642 135324625303360 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 0x7b1300750c80>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 13:36:12.214932 135324625303360 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0417 13:36:12.215159 135324625303360 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 0x7aff006b8380>}, handler_registry=None I0417 13:36:12.215392 135324625303360 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aff006b8380>` 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 13:36:12.215443 135324625303360 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aff006ba690>` 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 13:36:12.215478 135324625303360 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 0x7aff006b8380>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aff006b8380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aff006ba690>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aff006ba690>}). I0417 13:36:12.215799 135324625303360 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 13:36:12.215886 135324625303360 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7afe2060c7c0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 13:36:14.435723 135324625303360 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints I0417 13:36:15.516538 135324625303360 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_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7aff006ba480> I0417 13:36:15.516713 135324625303360 checkpointing.py:301] Checkpoint manager created! I0417 13:36:16.427875 135324625303360 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 13:36:16.427987 135324625303360 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 13:36:16.805231 135324625303360 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 13:36:16.805319 135324625303360 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 13:36:16.821445 135324625303360 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 13:36:16.821500 135324625303360 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 13:36:16.845083 135324625303360 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 13:36:16.845150 135324625303360 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:36:16.861273 135324625303360 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 13:36:16.861337 135324625303360 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:36:16.877617 135324625303360 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 13:36:16.877684 135324625303360 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:36:16.902541 135324625303360 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 13:36:16.902607 135324625303360 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:36:16.924555 135324625303360 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 13:36:16.924619 135324625303360 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 13:36:17.140488 135324625303360 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0417 13:36:17.140593 135324625303360 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 13:36:18.563393 135324625303360 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 13:36:18.563515 135324625303360 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 13:36:18.563568 135324625303360 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 13:36:18.563625 135324625303360 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 13:36:18.563678 135324625303360 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 13:36:18.563721 135324625303360 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 13:36:18.563774 135324625303360 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 13:36:18.563827 135324625303360 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 13:36:18.563894 135324625303360 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 13:36:18.563942 135324625303360 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 13:36:18.563994 135324625303360 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 13:36:18.564044 135324625303360 maxtext_utils.py:1651] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 13:36:19.062427 135324625303360 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 13:36:19.062523 135324625303360 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0417 13:36:19.077772 135324625303360 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 13:36:19.077833 135324625303360 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0417 13:36:29.706403 135324625303360 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 13:36:29.714081 135324625303360 metric_logger.py:301] number parameters: 1.104 billion I0417 13:36:40.970558 135324625303360 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0417 13:36:41.358175 135324625303360 checkpointing.py:776] Waited 0.38759279251098633 seconds for step 0 to finish before starting checkpointing. I0417 13:36:41.360593 135324625303360 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 13:36:41.362565 135324625303360 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 13:36:41.363953 135324625303360 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0. I0417 13:36:42.550222 135324625303360 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 13:36:42.564594 135324625303360 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 13:36:42.564659 135324625303360 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 13:36:42.837736 135324625303360 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273848s I0417 13:36:42.839963 135324625303360 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.476 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2816884517669678 s) (per-host) I0417 13:36:42.840049 135324625303360 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281798s (batch_requests_ready=0.002228s, total_serialization_initiated=0.277429s, others=0.002141s) I0417 13:36:42.840136 135324625303360 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285806s (all_items=0.000023s, per_item={'items': '0.00002337'}, temp_paths=0.285783) I0417 13:36:42.840849 135324625303360 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.48 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0. I0417 13:36:42.841072 135198133540608 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:56:42.841040 I0417 13:36:42.854107 135324625303360 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:36:42.854383 135194385762048 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:36:42.854535 135324625303360 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776433001.3605742, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776433001.3626046, 'checkpointer_blocking_duration_secs': 1.4786145687103271, 'get_old_steps_start_time': 1776433002.841243, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776433001.3587036, 'checkpoint_manager_blocking_duration_secs': 1.495793342590332} I0417 13:36:42.854637 135324625303360 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0417 13:36:42.854685 135324625303360 max_utils.py:750] Memstats: After params initialized: I0417 13:36:42.854733 135324625303360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 13:36:42.854765 135324625303360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 13:36:42.854790 135324625303360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 13:36:42.854825 135324625303360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 13:36:42.891383 135197610563328 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0 I0417 13:36:43.190215 135324625303360 metric_logger.py:196] completed step: 0, seconds: 10.957, TFLOP/s/device: 1.240, Tokens/s/device: 186.906, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0417 13:36:43.193469 135324625303360 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/tensorboard/' I0417 13:36:44.026332 135324625303360 metric_logger.py:196] completed step: 1, seconds: 2.218, TFLOP/s/device: 6.127, Tokens/s/device: 923.466, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0417 13:36:44.172310 135324625303360 metric_logger.py:196] completed step: 2, seconds: 0.837, TFLOP/s/device: 16.236, Tokens/s/device: 2447.302, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0417 13:36:44.319597 135324625303360 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 1380.245, Tokens/s/device: 208045.510, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0417 13:36:44.328879 135324625303360 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0417 13:36:44.613778 135324625303360 checkpointing.py:776] Waited 0.2848629951477051 seconds for step 5 to finish before starting checkpointing. I0417 13:36:44.616805 135324625303360 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:36:46.288719 135197593777920 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0/items I0417 13:36:46.896551 135193327298304 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776433005626338797, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776433005626338797, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA I0417 13:36:48.300027 2811 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 13:36:52.287972 135194937911040 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_08_checkpoint_async_true/checkpoints/0/items/array_metadatas/process_0 I0417 13:37:28.056761 135194929518336 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 39.763907s (commit=35.801991s, array_metadata_write=3.961916s) I0417 13:37:28.059133 135198133540608 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.716 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.50082230567932 s) (per-host) I0417 13:37:28.059313 135198133540608 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.218120s. I0417 13:37:30.896749 135198133540608 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776433005626338797, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA I0417 13:37:33.076455 135198133540608 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:37:33.353579 135193327298304 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776433005626338797, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA I0417 13:37:34.167449 135198133540608 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000353s. I0417 13:37:34.167686 135198133540608 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.737084s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0/items I0417 13:37:34.169468 135198133540608 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0/items I0417 13:37:35.938276 135198133540608 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0 I0417 13:37:40.209008 135198133540608 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_08_checkpoint_async_true/checkpoints/0`. I0417 13:37:40.210148 135198133540608 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 58.85 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/0 I0417 13:37:40.211848 135198133540608 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 57.370640s. I0417 13:37:40.212131 135194385762048 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:37:40.212291 135194385762048 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:37:40.212398 135194385762048 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:37:40.213930 135194385762048 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 13:37:40.214203 135324625303360 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 13:37:40.214325 135324625303360 checkpoint_manager.py:1452] Waiting for previous save to complete took 55.597528 seconds. If this number is high, consider checkpointing less frequently. I0417 13:37:40.216232 135324625303360 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5 I0417 13:37:40.218519 135324625303360 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5. I0417 13:37:41.369515 135324625303360 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 13:37:41.369646 135324625303360 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 13:37:41.405959 135324625303360 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037423s I0417 13:37:41.406849 135324625303360 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 33.609 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0458981990814209 s) (per-host) I0417 13:37:41.406943 135324625303360 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046026s (batch_requests_ready=0.004028s, total_serialization_initiated=0.041227s, others=0.000771s) I0417 13:37:41.407065 135324625303360 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050836s (all_items=0.000026s, per_item={'items': '0.00002575'}, temp_paths=0.050810) I0417 13:37:41.407997 135324625303360 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.19 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5. I0417 13:37:41.408445 135194385762048 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:57:41.408403 I0417 13:37:41.410871 135324625303360 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:37:41.411248 135196552709888 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:37:41.411473 135324625303360 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776433004.6167605, 'wait_for_prev_duration_secs': 55.597527742385864, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776433060.2162983, 'checkpointer_blocking_duration_secs': 1.19227933883667, 'get_old_steps_start_time': 1776433061.4086058, 'get_old_steps_duration_secs': 4.2438507080078125e-05, 'checkpoint_manager_blocking_start_time': 1776433004.6144931, 'checkpoint_manager_blocking_duration_secs': 56.79692769050598} I0417 13:37:41.411614 135324625303360 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0417 13:37:42.115075 135197072795392 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5 I0417 13:37:47.740429 135196577888000 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5/items I0417 13:37:57.674745 135196569495296 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_08_checkpoint_async_true/checkpoints/5/items/array_metadatas/process_0 I0417 13:38:11.767880 135324625303360 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.637, Tokens/s/device: 13963.224, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0417 13:38:11.777092 135324625303360 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.370, Tokens/s/device: 13923.070, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0417 13:38:11.922569 135324625303360 metric_logger.py:196] completed step: 6, seconds: 87.447, TFLOP/s/device: 0.155, Tokens/s/device: 23.420, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0417 13:38:12.070333 135324625303360 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1903.226, Tokens/s/device: 286874.912, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0417 13:38:12.217267 135324625303360 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.665, Tokens/s/device: 13967.414, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0417 13:38:12.363938 135324625303360 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0417 13:38:12.364676 135324625303360 checkpointing.py:776] Waited 0.0007605552673339844 seconds for step 9 to finish before starting checkpointing. I0417 13:38:12.366696 135324625303360 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:38:35.523112 135196561102592 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.255291s (commit=39.820024s, array_metadata_write=3.435267s) I0417 13:38:35.524258 135194385762048 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 29.164 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 54.16329884529114 s) (per-host) I0417 13:38:35.524390 135194385762048 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 54.115137s. I0417 13:38:39.187290 135194385762048 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:38:39.534005 135194385762048 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000258s. I0417 13:38:40.109887 135194385762048 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.833297s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5/items I0417 13:38:40.115924 135194385762048 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5/items I0417 13:38:41.578679 135194385762048 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5 I0417 13:38:44.938232 135194385762048 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_08_checkpoint_async_true/checkpoints/5`. I0417 13:38:44.939126 135194385762048 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 64.72 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/5 I0417 13:38:44.940468 135194385762048 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 63.531215s. I0417 13:38:44.940658 135196552709888 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:38:44.940772 135196552709888 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:38:44.940849 135196552709888 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:38:44.946520 135196552709888 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0417 13:38:44.946703 135324625303360 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 13:38:44.946833 135324625303360 checkpoint_manager.py:1452] Waiting for previous save to complete took 32.580143 seconds. If this number is high, consider checkpointing less frequently. I0417 13:38:44.948507 135324625303360 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 13:38:44.950535 135324625303360 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9. I0417 13:38:45.745705 135324625303360 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 13:38:45.745798 135324625303360 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 13:38:45.771406 135324625303360 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026622s I0417 13:38:45.772229 135324625303360 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.834 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03095412254333496 s) (per-host) I0417 13:38:45.772299 135324625303360 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031043s (batch_requests_ready=0.001760s, total_serialization_initiated=0.028503s, others=0.000779s) I0417 13:38:45.772372 135324625303360 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034991s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.034976) I0417 13:38:45.773098 135324625303360 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.82 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9. I0417 13:38:45.773401 135196552709888 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:58:45.773365 I0417 13:38:45.787739 135324625303360 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:38:45.788024 135196569495296 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:38:45.788214 135324625303360 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776433092.3666635, 'wait_for_prev_duration_secs': 32.58014273643494, 'time_between_consecutive_saves_sec': 32.15266418457031, 'checkpointer_blocking_start_time': 1776433124.9485466, 'checkpointer_blocking_duration_secs': 0.8249998092651367, 'get_old_steps_start_time': 1776433125.7735655, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776433092.364913, 'checkpoint_manager_blocking_duration_secs': 33.42325758934021} I0417 13:38:45.788359 135324625303360 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0417 13:38:45.788414 135324625303360 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:38:45.970451 135194385762048 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9 I0417 13:38:50.858821 135197072795392 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9/items I0417 13:38:57.261120 135197114758912 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_08_checkpoint_async_true/checkpoints/9/items/array_metadatas/process_0 I0417 13:39:36.075446 135196577888000 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.000194s (commit=39.844292s, array_metadata_write=3.155901s) I0417 13:39:36.081714 135196552709888 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 31.378 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 50.34041094779968 s) (per-host) I0417 13:39:36.081861 135196552709888 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 50.308335s. I0417 13:39:41.014146 135196552709888 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:39:42.206726 135196552709888 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000236s. I0417 13:39:42.206928 135196552709888 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.911511s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9/items I0417 13:39:42.208309 135196552709888 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9/items I0417 13:39:43.326708 135196552709888 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9 I0417 13:39:47.035980 135196552709888 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_08_checkpoint_async_true/checkpoints/9`. I0417 13:39:47.036699 135196552709888 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 62.09 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_08_checkpoint_async_true/checkpoints/9 I0417 13:39:47.038350 135196552709888 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 61.264826s. I0417 13:39:47.038515 135196569495296 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:39:47.038637 135196569495296 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:39:47.038688 135196569495296 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:39:47.041578 135196569495296 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 13:39:47.041788 135324625303360 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 13:39:47.041981 135324625303360 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 13:39:47.042898 135324625303360 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.074, Tokens/s/device: 13878.348, 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 13:39:56 UTC 2026 EXIT_CODE=0