XPK Start: Sat Apr 18 13:27:11 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 13:27:35.372452: 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 13:27:35.553712 135031345899328 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 13:27:44.595873 135031345899328 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 13:27:44.598170 135031345899328 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--v6haw-slice-job-0-0.mt-08-checkpoint-async--v6haw:8482 I0418 13:27:46.289463 135031345899328 max_utils.py:284] Jax distributed system initialized! I0418 13:27:52.437690 135031345899328 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 13:27:52.437790 135031345899328 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 13:27:52.437830 135031345899328 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 13:27:52.437865 135031345899328 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 13:27:53.130947 135031345899328 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 13:27:53.131227 135031345899328 checkpointing.py:677] Setting up checkpoint logger... I0418 13:27:53.131281 135031345899328 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 13:27:53.131326 135031345899328 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 13:27:53.131661 135031345899328 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 0x7aced3be75c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 13:27:54.507142 135031345899328 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 13:27:54.507378 135031345899328 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aba8435df70>}, handler_registry=None I0418 13:27:54.507629 135031345899328 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aba8435df70>` 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 13:27:54.507679 135031345899328 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7acdabf94c80>` 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 13:27:54.507715 135031345899328 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 0x7aba8435df70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aba8435df70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7acdabf94c80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7acdabf94c80>}). I0418 13:27:54.508044 135031345899328 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 13:27:54.508113 135031345899328 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7aba746dbec0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 13:27:57.618472 134910274688768 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776518601917348229, 'commit_timestamp_nsecs': 1776518647359629552, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA I0418 13:27:57.861287 134910266296064 checkpoint_manager.py:1871] Missing metrics for step 5 I0418 13:27:59.146941 135031345899328 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints I0418 13:27:59.977359 135031345899328 checkpoint_manager.py:929] [process=3][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_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7aba8435fb60> I0418 13:27:59.977533 135031345899328 checkpointing.py:301] Checkpoint manager created! I0418 13:28:00.902467 135031345899328 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 13:28:00.902579 135031345899328 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 13:28:01.284157 135031345899328 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 13:28:01.284252 135031345899328 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 13:28:01.300533 135031345899328 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 13:28:01.300592 135031345899328 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 13:28:01.324205 135031345899328 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 13:28:01.324277 135031345899328 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:28:01.340576 135031345899328 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 13:28:01.340645 135031345899328 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:28:01.356898 135031345899328 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 13:28:01.356958 135031345899328 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:28:01.382440 135031345899328 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 13:28:01.382511 135031345899328 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:28:01.403563 135031345899328 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 13:28:01.403641 135031345899328 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 13:28:01.621595 135031345899328 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 13:28:01.621710 135031345899328 checkpointing.py:581] restoring from this run's directory step 9 I0418 13:28:01.866072 135031345899328 event_tracking.py:70] [process=3] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9. I0418 13:28:02.106400 135031345899328 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9. I0418 13:28:02.106847 135031345899328 event_tracking.py:125] [process=3] [sync] Finished blocking load in 0.24 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9. I0418 13:28:05.314530 1450 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 13:28:12.523158 135031345899328 jax_array_handlers.py:843] [process=3] /jax/orbax/read/worker/io/requested throughput: 219.041 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.214517831802368 s) (per-host) I0418 13:28:12.528803 135031345899328 base_pytree_checkpoint_handler.py:130] [process=3] /jax/checkpoint/read/gbytes_per_sec: 196.884 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 8.038224458694458 s) (per-host) I0418 13:28:12.829091 135031345899328 event_tracking.py:138] [process=3] [sync] Finished load in 10.96 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/9 I0418 13:28:12.829245 135031345899328 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776518881.865975, 'checkpointer_duration_secs': 10.96323823928833, 'checkpoint_manager_start_time': 1776518881.6219559, 'checkpoint_manager_duration_secs': 11.20725965499878} fsdp: 32 I0418 13:28:13.244451 135031345899328 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 13:28:13.244575 135031345899328 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 13:28:13.244643 135031345899328 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 13:28:13.244703 135031345899328 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 13:28:13.244754 135031345899328 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 13:28:13.244791 135031345899328 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 13:28:13.244842 135031345899328 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 13:28:13.244893 135031345899328 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 13:28:13.244933 135031345899328 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 13:28:13.244969 135031345899328 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 13:28:13.245013 135031345899328 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 13:28:13.245062 135031345899328 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 13:28:13.726984 135031345899328 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 13:28:13.727077 135031345899328 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 13:28:13.742753 135031345899328 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 13:28:13.742811 135031345899328 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 13:28:24.645495 135031345899328 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 13:28:24.646301 135031345899328 metric_logger.py:301] number parameters: 1.104 billion I0418 13:28:35.968492 135031345899328 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0418 13:28:36.159229 135031345899328 checkpointing.py:776] Waited 0.1907215118408203 seconds for step 10 to finish before starting checkpointing. I0418 13:28:36.161635 135031345899328 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 13:28:36.163724 135031345899328 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 10 I0418 13:28:36.165129 135031345899328 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/10. I0418 13:28:37.326140 135031345899328 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 13:28:37.327155 135031345899328 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 13:28:37.327210 135031345899328 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 13:28:37.607939 135031345899328 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281871s I0418 13:28:37.608106 135031345899328 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.399 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2857239246368408 s) (per-host) I0418 13:28:37.608159 135031345899328 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285787s (batch_requests_ready=0.002146s, total_serialization_initiated=0.283570s, others=0.000072s) I0418 13:28:37.608231 135031345899328 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289915s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.289900) I0418 13:28:37.609000 135031345899328 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.45 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/10. I0418 13:28:37.609356 134909746210560 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:48:37.609316 I0418 13:28:37.615899 135031345899328 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0418 13:28:37.616190 134906441090816 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 13:28:37.616343 135031345899328 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776518916.1616042, 'wait_for_prev_duration_secs': 7.510185241699219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776518916.1637623, 'checkpointer_blocking_duration_secs': 1.4457650184631348, 'get_old_steps_start_time': 1776518917.6095514, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776518916.1597521, 'checkpoint_manager_blocking_duration_secs': 1.456554889678955} I0418 13:28:37.616454 135031345899328 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0418 13:28:37.616504 135031345899328 max_utils.py:750] Memstats: After params initialized: I0418 13:28:37.616551 135031345899328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0418 13:28:37.616584 135031345899328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0418 13:28:37.616610 135031345899328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0418 13:28:37.616647 135031345899328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0418 13:28:37.928068 135031345899328 metric_logger.py:196] completed step: 10, seconds: 11.322, TFLOP/s/device: 1.200, Tokens/s/device: 180.886, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0418 13:28:38.113716 135031345899328 metric_logger.py:196] completed step: 11, seconds: 1.958, TFLOP/s/device: 6.939, Tokens/s/device: 1045.977, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0418 13:28:38.261203 135031345899328 metric_logger.py:196] completed step: 12, seconds: 0.039, TFLOP/s/device: 352.391, Tokens/s/device: 53116.166, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0418 13:28:38.408653 135031345899328 metric_logger.py:196] completed step: 13, seconds: 0.154, TFLOP/s/device: 88.504, Tokens/s/device: 13340.281, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0418 13:28:38.413423 135031345899328 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0418 13:28:38.702971 135031345899328 checkpointing.py:776] Waited 0.28951430320739746 seconds for step 15 to finish before starting checkpointing. I0418 13:28:38.705988 135031345899328 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 13:28:44.662558 134906449483520 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/10/items/array_metadatas/process_3 I0418 13:29:20.912366 134909746210560 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.238 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.589948892593384 s) (per-host) I0418 13:29:20.912493 134909746210560 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.303014s. I0418 13:29:29.679650 134909746210560 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 52.070139s. I0418 13:29:29.679924 134906441090816 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 13:29:29.680048 134906441090816 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 13:29:29.680095 134906441090816 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0418 13:29:29.681555 134906441090816 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0418 13:29:29.681745 135031345899328 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0418 13:29:29.681878 135031345899328 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.975889 seconds. If this number is high, consider checkpointing less frequently. I0418 13:29:29.683807 135031345899328 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 15 I0418 13:29:29.686354 135031345899328 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/15. I0418 13:29:30.403565 135031345899328 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 13:29:30.403676 135031345899328 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 13:29:30.428580 135031345899328 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026189s I0418 13:29:30.428724 135031345899328 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 52.123 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.029595136642456055 s) (per-host) I0418 13:29:30.428771 135031345899328 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.029651s (batch_requests_ready=0.001772s, total_serialization_initiated=0.027817s, others=0.000062s) I0418 13:29:30.428863 135031345899328 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034632s (all_items=0.000013s, per_item={'items': '0.00001264'}, temp_paths=0.034619) I0418 13:29:30.429524 135031345899328 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/15. I0418 13:29:30.429819 134906449483520 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:49:30.429797 I0418 13:29:30.441788 135031345899328 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0418 13:29:30.442092 134906441090816 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 13:29:30.442258 135031345899328 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776518918.7059581, 'wait_for_prev_duration_secs': 50.97588872909546, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776518969.683847, 'checkpointer_blocking_duration_secs': 0.7460718154907227, 'get_old_steps_start_time': 1776518970.4299428, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776518918.703584, 'checkpoint_manager_blocking_duration_secs': 51.73863458633423} I0418 13:29:30.442363 135031345899328 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0418 13:29:30.443277 135031345899328 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.237, Tokens/s/device: 13903.032, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0418 13:29:30.450159 135031345899328 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.135, Tokens/s/device: 13887.665, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0418 13:29:37.389668 134904822818560 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/15/items/array_metadatas/process_3 I0418 13:29:52.441539 135031345899328 metric_logger.py:196] completed step: 16, seconds: 52.035, TFLOP/s/device: 0.261, Tokens/s/device: 39.358, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0418 13:29:52.589063 135031345899328 metric_logger.py:196] completed step: 17, seconds: 21.845, TFLOP/s/device: 0.622, Tokens/s/device: 93.751, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0418 13:29:52.736463 135031345899328 metric_logger.py:196] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.915, Tokens/s/device: 13402.177, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0418 13:29:52.883064 135031345899328 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0418 13:29:52.884015 135031345899328 checkpointing.py:776] Waited 0.0009698867797851562 seconds for step 19 to finish before starting checkpointing. I0418 13:29:52.886351 135031345899328 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 13:30:13.378345 134906449483520 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.753 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.979177951812744 s) (per-host) I0418 13:30:13.378465 134906449483520 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.948569s. I0418 13:30:21.743053 134906449483520 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 51.313143s. I0418 13:30:21.743308 134906441090816 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 13:30:21.743437 134906441090816 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 13:30:21.743484 134906441090816 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0418 13:30:21.744961 134906441090816 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0418 13:30:21.745141 135031345899328 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0418 13:30:21.745263 135031345899328 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.858918 seconds. If this number is high, consider checkpointing less frequently. I0418 13:30:21.746588 135031345899328 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 19 I0418 13:30:21.748649 135031345899328 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/19. I0418 13:30:22.504878 135031345899328 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 13:30:22.504969 135031345899328 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 13:30:22.530594 135031345899328 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026634s I0418 13:30:22.530730 135031345899328 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 51.384 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030020713806152344 s) (per-host) I0418 13:30:22.530777 135031345899328 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030076s (batch_requests_ready=0.001621s, total_serialization_initiated=0.028395s, others=0.000060s) I0418 13:30:22.530849 135031345899328 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034379s (all_items=0.000010s, per_item={'items': '0.00000954'}, temp_paths=0.034369) I0418 13:30:22.531534 135031345899328 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.78 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/19. I0418 13:30:22.531899 134906441090816 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:50:22.531854 I0418 13:30:22.535412 135031345899328 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0418 13:30:22.535648 134905351259904 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 13:30:22.535767 135031345899328 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776518992.88632, 'wait_for_prev_duration_secs': 28.85891842842102, 'time_between_consecutive_saves_sec': 23.204723834991455, 'checkpointer_blocking_start_time': 1776519021.7466433, 'checkpointer_blocking_duration_secs': 0.7854135036468506, 'get_old_steps_start_time': 1776519022.5320728, 'get_old_steps_duration_secs': 3.457069396972656e-05, 'checkpoint_manager_blocking_start_time': 1776518992.8842564, 'checkpoint_manager_blocking_duration_secs': 29.65147876739502} I0418 13:30:22.535891 135031345899328 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0418 13:30:22.535934 135031345899328 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 13:30:30.268745 134904822818560 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_08_checkpoint_async_true/checkpoints/19/items/array_metadatas/process_3 I0418 13:31:06.453058 134906441090816 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 35.939 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.95230746269226 s) (per-host) I0418 13:31:06.453185 134906441090816 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.921154s. I0418 13:31:15.948058 134906441090816 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 53.416016s. I0418 13:31:15.948349 134905351259904 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 13:31:15.948477 134905351259904 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 13:31:15.948522 134905351259904 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0418 13:31:15.950194 134905351259904 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0418 13:31:15.950360 135031345899328 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0418 13:31:15.950512 135031345899328 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 13:31:15.951526 135031345899328 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.164, Tokens/s/device: 13891.904, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 13:31:26 UTC 2026 EXIT_CODE=0