XPK Start: Sat Apr 18 18:27:57 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 18:28:21.237767: 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 18:28:21.776747 132574670817088 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 18:28:30.818197 132574670817088 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 18:28:30.820647 132574670817088 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--z5lnl-slice-job-0-0.mt-08-checkpoint-async--z5lnl:8482 I0418 18:28:31.745230 132574670817088 max_utils.py:284] Jax distributed system initialized! I0418 18:28:37.797868 132574670817088 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 18:28:37.797971 132574670817088 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 18:28:37.798011 132574670817088 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 18:28:37.798045 132574670817088 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 18:28:38.495049 132574670817088 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 18:28:38.495357 132574670817088 checkpointing.py:677] Setting up checkpoint logger... I0418 18:28:38.495410 132574670817088 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 18:28:38.495454 132574670817088 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 18:28:38.495784 132574670817088 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 0x7892ba84c170>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 18:28:42.196826 132574670817088 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 18:28:42.197061 132574670817088 checkpoint_manager.py:708] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x787e7848f140>}, handler_registry=None I0418 18:28:42.197313 132574670817088 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x787e7848f140>` 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 18:28:42.197367 132574670817088 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x787e78492a50>` 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 18:28:42.197402 132574670817088 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 0x787e7848f140>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x787e7848f140>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x787e78492a50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x787e78492a50>}). I0418 18:28:42.197716 132574670817088 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 18:28:42.197786 132574670817088 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x787db07c7600> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 18:28:43.455608 132574670817088 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints I0418 18:28:44.771598 132574670817088 checkpoint_manager.py:929] [process=7][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_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7891af0c8740> I0418 18:28:44.771769 132574670817088 checkpointing.py:301] Checkpoint manager created! I0418 18:28:45.688111 132574670817088 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 18:28:45.688223 132574670817088 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 18:28:46.066898 132574670817088 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 18:28:46.066985 132574670817088 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 18:28:46.083353 132574670817088 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 18:28:46.083413 132574670817088 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 18:28:46.107329 132574670817088 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 18:28:46.107397 132574670817088 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 18:28:46.123526 132574670817088 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 18:28:46.123585 132574670817088 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 18:28:46.139661 132574670817088 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 18:28:46.139720 132574670817088 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 18:28:46.164100 132574670817088 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 18:28:46.164166 132574670817088 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 18:28:46.185165 132574670817088 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 18:28:46.185226 132574670817088 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 18:28:46.401657 132574670817088 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 18:28:46.401766 132574670817088 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 18:28:47.818270 132574670817088 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 18:28:47.818393 132574670817088 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 18:28:47.818446 132574670817088 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 18:28:47.818503 132574670817088 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 18:28:47.818555 132574670817088 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 18:28:47.818593 132574670817088 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 18:28:47.818643 132574670817088 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 18:28:47.818692 132574670817088 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 18:28:47.818730 132574670817088 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 18:28:47.818767 132574670817088 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 18:28:47.818812 132574670817088 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 18:28:47.818866 132574670817088 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 18:28:48.304119 132574670817088 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 18:28:48.304211 132574670817088 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 18:28:48.319483 132574670817088 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 18:28:48.319546 132574670817088 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 18:28:59.260764 132574670817088 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 18:28:59.261606 132574670817088 metric_logger.py:301] number parameters: 1.104 billion I0418 18:29:10.649387 132574670817088 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0418 18:29:10.804337 132574670817088 checkpointing.py:776] Waited 0.15493249893188477 seconds for step 0 to finish before starting checkpointing. I0418 18:29:10.806760 132574670817088 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 18:29:10.808538 132574670817088 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 0 I0418 18:29:10.809932 132574670817088 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/0. I0418 18:29:12.007992 132574670817088 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 18:29:12.008984 132574670817088 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 18:29:12.009042 132574670817088 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 18:29:12.292635 132574670817088 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.284722s I0418 18:29:12.292803 132574670817088 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.316 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2901487350463867 s) (per-host) I0418 18:29:12.292853 132574670817088 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.290208s (batch_requests_ready=0.002192s, total_serialization_initiated=0.287949s, others=0.000066s) I0418 18:29:12.292940 132574670817088 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.294337s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.294318) I0418 18:29:12.293761 132574670817088 event_tracking.py:125] [process=7] [async] Finished blocking save in 1.49 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/0. I0418 18:29:12.294131 132446938552064 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 18:49:12.294091 I0418 18:29:12.298395 132574670817088 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 18:29:12.298689 132446439855872 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0418 18:29:12.298842 132574670817088 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776536950.8067417, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776536950.808577, 'checkpointer_blocking_duration_secs': 1.4857075214385986, 'get_old_steps_start_time': 1776536952.2943108, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776536950.8048866, 'checkpoint_manager_blocking_duration_secs': 1.4939169883728027} I0418 18:29:12.298953 132574670817088 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0418 18:29:12.299003 132574670817088 max_utils.py:750] Memstats: After params initialized: I0418 18:29:12.299051 132574670817088 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0418 18:29:12.299095 132574670817088 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0418 18:29:12.299123 132574670817088 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0418 18:29:12.299148 132574670817088 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0418 18:29:12.609538 132574670817088 metric_logger.py:196] completed step: 0, seconds: 11.388, TFLOP/s/device: 1.193, Tokens/s/device: 179.844, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 18:29:12.784892 132574670817088 metric_logger.py:196] completed step: 1, seconds: 1.958, TFLOP/s/device: 6.938, Tokens/s/device: 1045.767, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 18:29:13.189715 132574670817088 metric_logger.py:196] completed step: 2, seconds: 0.028, TFLOP/s/device: 476.992, Tokens/s/device: 71897.490, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0418 18:29:13.337103 132574670817088 metric_logger.py:196] completed step: 3, seconds: 0.405, TFLOP/s/device: 33.569, Tokens/s/device: 5059.951, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0418 18:29:13.341547 132574670817088 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0418 18:29:13.631338 132574670817088 checkpointing.py:776] Waited 0.28975391387939453 seconds for step 5 to finish before starting checkpointing. I0418 18:29:13.634217 132574670817088 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 18:29:15.996151 2757 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 18:29:18.884357 132446448248576 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/0/items/array_metadatas/process_7 I0418 18:29:48.461236 132446938552064 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 43.326 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.45853900909424 s) (per-host) I0418 18:29:48.461359 132446938552064 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.167114s. I0418 18:29:57.706493 132446938552064 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 45.412232s. I0418 18:29:57.706776 132446439855872 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 18:29:57.706893 132446439855872 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0418 18:29:57.706943 132446439855872 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 18:29:57.708966 132446439855872 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 18:29:57.709083 132574670817088 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0418 18:29:57.709154 132574670817088 checkpoint_manager.py:1452] Waiting for previous save to complete took 44.074941 seconds. If this number is high, consider checkpointing less frequently. I0418 18:29:57.711159 132574670817088 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 5 I0418 18:29:57.713717 132574670817088 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/5. I0418 18:29:58.874671 132574670817088 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 18:29:58.874763 132574670817088 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 18:29:58.924510 132574670817088 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.050913s I0418 18:29:58.924671 132574670817088 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 28.420 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05427718162536621 s) (per-host) I0418 18:29:58.924719 132574670817088 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.054335s (batch_requests_ready=0.001739s, total_serialization_initiated=0.052531s, others=0.000065s) I0418 18:29:58.924805 132574670817088 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.059390s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.059375) I0418 18:29:58.925482 132574670817088 event_tracking.py:125] [process=7] [async] Finished blocking save in 1.21 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/5. I0418 18:29:58.925825 132446448248576 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 18:49:58.925788 I0418 18:29:58.927812 132574670817088 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0418 18:29:58.928125 132446439855872 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0418 18:29:58.928285 132574670817088 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776536953.6341896, 'wait_for_prev_duration_secs': 44.07494115829468, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776536997.711196, 'checkpointer_blocking_duration_secs': 1.2147679328918457, 'get_old_steps_start_time': 1776536998.9259868, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776536953.631946, 'checkpoint_manager_blocking_duration_secs': 45.29630422592163} I0418 18:29:58.928399 132574670817088 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0418 18:29:58.929236 132574670817088 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.822, Tokens/s/device: 13388.159, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0418 18:29:58.935419 132574670817088 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.122, Tokens/s/device: 13885.687, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0418 18:30:06.163572 132424084199168 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/5/items/array_metadatas/process_7 I0418 18:30:22.342277 132574670817088 metric_logger.py:196] completed step: 6, seconds: 45.592, TFLOP/s/device: 0.298, Tokens/s/device: 44.920, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0418 18:30:22.489804 132574670817088 metric_logger.py:196] completed step: 7, seconds: 23.261, TFLOP/s/device: 0.584, Tokens/s/device: 88.045, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0418 18:30:22.637305 132574670817088 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.354, Tokens/s/device: 13468.456, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0418 18:30:22.784133 132574670817088 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0418 18:30:22.784820 132574670817088 checkpointing.py:776] Waited 0.0007138252258300781 seconds for step 9 to finish before starting checkpointing. I0418 18:30:22.787127 132574670817088 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 18:30:40.960010 132446448248576 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.529 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.08957839012146 s) (per-host) I0418 18:30:40.960157 132446448248576 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.034222s. I0418 18:30:51.208843 132446448248576 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 52.282891s. I0418 18:30:51.209142 132446439855872 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 18:30:51.209261 132446439855872 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0418 18:30:51.209311 132446439855872 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0418 18:30:51.210919 132446439855872 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0418 18:30:51.211128 132574670817088 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0418 18:30:51.211251 132574670817088 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.424131 seconds. If this number is high, consider checkpointing less frequently. I0418 18:30:51.212728 132574670817088 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 9 I0418 18:30:51.214944 132574670817088 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9. I0418 18:30:52.372566 132574670817088 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 18:30:52.372654 132574670817088 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 18:30:52.421012 132574670817088 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049388s I0418 18:30:52.421192 132574670817088 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 29.235 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0527651309967041 s) (per-host) I0418 18:30:52.421243 132574670817088 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.052827s (batch_requests_ready=0.001602s, total_serialization_initiated=0.051158s, others=0.000067s) I0418 18:30:52.421334 132574670817088 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.056937s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.056926) I0418 18:30:52.422032 132574670817088 event_tracking.py:125] [process=7] [async] Finished blocking save in 1.21 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9. I0418 18:30:52.422379 132446439855872 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 18:50:52.422342 I0418 18:30:52.424278 132574670817088 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 18:30:52.424505 132443755501312 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0418 18:30:52.424628 132574670817088 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776537022.7870958, 'wait_for_prev_duration_secs': 28.424131393432617, 'time_between_consecutive_saves_sec': 25.078092098236084, 'checkpointer_blocking_start_time': 1776537051.212766, 'checkpointer_blocking_duration_secs': 1.2097578048706055, 'get_old_steps_start_time': 1776537052.4225416, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776537022.7850442, 'checkpoint_manager_blocking_duration_secs': 29.639551401138306} I0418 18:30:52.424737 132574670817088 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0418 18:30:52.424778 132574670817088 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 18:31:00.194608 132446406285056 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9/items/array_metadatas/process_7 I0418 18:31:34.873636 132446439855872 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.162 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.50517749786377 s) (per-host) I0418 18:31:34.873746 132446439855872 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.451258s. I0418 18:31:44.769047 132446439855872 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 52.346543s. I0418 18:31:44.769276 132443755501312 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 18:31:44.769338 132443755501312 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0418 18:31:44.769379 132443755501312 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 18:31:44.771227 132443755501312 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 18:31:44.771396 132574670817088 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0418 18:31:44.771546 132574670817088 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 18:31:44.772477 132574670817088 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.119, Tokens/s/device: 13885.122, 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 18:31:58 UTC 2026 EXIT_CODE=0