XPK Start: Sun Apr 19 18:38:45 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 18:39:09.918817: 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) I0419 18:39:10.464323 135763238426432 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 18:39:19.505375 135763238426432 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 18:39:19.507745 135763238426432 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--qnigi-slice-job-0-0.mt-08-checkpoint-async--qnigi:8482 I0419 18:39:20.587832 135763238426432 max_utils.py:284] Jax distributed system initialized! I0419 18:39:25.669207 135763238426432 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 18:39:25.669309 135763238426432 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 18:39:25.669349 135763238426432 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 I0419 18:39:25.669382 135763238426432 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 18:39:26.360594 135763238426432 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 18:39:26.360867 135763238426432 checkpointing.py:677] Setting up checkpoint logger... I0419 18:39:26.360922 135763238426432 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 18:39:26.360968 135763238426432 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 18:39:26.361313 135763238426432 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 0x7b791fd502f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 18:39:28.037431 135763238426432 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0419 18:39:28.037680 135763238426432 checkpoint_manager.py:708] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b659068c9b0>}, handler_registry=None I0419 18:39:28.037917 135763238426432 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b659068c9b0>` 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`. I0419 18:39:28.037963 135763238426432 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b65905ed8e0>` 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`. I0419 18:39:28.037996 135763238426432 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 0x7b659068c9b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b659068c9b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b65905ed8e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b65905ed8e0>}). I0419 18:39:28.038321 135763238426432 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 18:39:28.038391 135763238426432 async_checkpointer.py:192] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b64c01d3600> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 18:39:29.523175 135644948961024 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776623745116142709, 'commit_timestamp_nsecs': 1776623791947603426, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/5/_CHECKPOINT_METADATA I0419 18:39:29.768720 135644948961024 checkpoint_manager.py:1871] Missing metrics for step 5 I0419 18:39:31.009472 135763238426432 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints I0419 18:39:31.454236 135763238426432 checkpoint_manager.py:929] [process=4][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_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b659054e810> I0419 18:39:31.454404 135763238426432 checkpointing.py:301] Checkpoint manager created! I0419 18:39:32.398699 135763238426432 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 18:39:32.398809 135763238426432 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 18:39:32.778748 135763238426432 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 18:39:32.778834 135763238426432 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 18:39:32.795232 135763238426432 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 18:39:32.795290 135763238426432 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 18:39:32.819895 135763238426432 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 18:39:32.819963 135763238426432 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:39:32.836375 135763238426432 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 18:39:32.836436 135763238426432 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:39:32.852777 135763238426432 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 18:39:32.852837 135763238426432 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:39:32.877629 135763238426432 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 18:39:32.877699 135763238426432 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:39:32.899044 135763238426432 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 18:39:32.899107 135763238426432 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 18:39:33.118589 135763238426432 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0419 18:39:33.118694 135763238426432 checkpointing.py:581] restoring from this run's directory step 9 I0419 18:39:33.372455 135763238426432 event_tracking.py:70] [process=4] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/9. I0419 18:39:34.191915 135763238426432 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/9. I0419 18:39:34.192354 135763238426432 event_tracking.py:125] [process=4] [sync] Finished blocking load in 0.82 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/9. I0419 18:39:36.143955 1409 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 18:39:42.089832 135763238426432 jax_array_handlers.py:843] [process=4] /jax/orbax/read/worker/io/requested throughput: 265.512 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.951806545257568 s) (per-host) I0419 18:39:42.094657 135763238426432 base_pytree_checkpoint_handler.py:130] [process=4] /jax/checkpoint/read/gbytes_per_sec: 248.729 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.362717866897583 s) (per-host) I0419 18:39:42.490299 135763238426432 event_tracking.py:138] [process=4] [sync] Finished load in 9.12 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/9 I0419 18:39:42.490453 135763238426432 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776623973.3723578, 'checkpointer_duration_secs': 9.118063688278198, 'checkpoint_manager_start_time': 1776623973.1189368, 'checkpoint_manager_duration_secs': 9.371487379074097} fsdp: 32 I0419 18:39:42.906832 135763238426432 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 18:39:42.906958 135763238426432 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) I0419 18:39:42.907016 135763238426432 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) I0419 18:39:42.907080 135763238426432 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') I0419 18:39:42.907133 135763238426432 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) I0419 18:39:42.907172 135763238426432 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) I0419 18:39:42.907224 135763238426432 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) I0419 18:39:42.907278 135763238426432 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') I0419 18:39:42.907318 135763238426432 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) I0419 18:39:42.907354 135763238426432 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) I0419 18:39:42.907401 135763238426432 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 18:39:42.907448 135763238426432 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 18:39:43.395977 135763238426432 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 18:39:43.396071 135763238426432 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 18:39:43.411602 135763238426432 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 18:39:43.411663 135763238426432 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 18:39:54.406671 135763238426432 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. I0419 18:39:54.407509 135763238426432 metric_logger.py:301] number parameters: 1.104 billion I0419 18:40:05.541618 135763238426432 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0419 18:40:05.698209 135763238426432 checkpointing.py:776] Waited 0.1565713882446289 seconds for step 10 to finish before starting checkpointing. I0419 18:40:05.700677 135763238426432 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 18:40:05.702723 135763238426432 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 10 I0419 18:40:05.704243 135763238426432 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/10. I0419 18:40:06.447688 135763238426432 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 18:40:06.450198 135763238426432 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 18:40:06.450253 135763238426432 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 I0419 18:40:06.734256 135763238426432 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.286646s I0419 18:40:06.734424 135763238426432 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 5.304 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29083847999572754 s) (per-host) I0419 18:40:06.734476 135763238426432 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.290902s (batch_requests_ready=0.002299s, total_serialization_initiated=0.288531s, others=0.000072s) I0419 18:40:06.734580 135763238426432 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.295203s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.295187) I0419 18:40:06.735295 135763238426432 event_tracking.py:125] [process=4] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/10. I0419 18:40:06.735655 135644361766656 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:00:06.735618 I0419 18:40:06.741932 135763238426432 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 18:40:06.742187 135643267053312 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0419 18:40:06.742311 135763238426432 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776624005.7006593, 'wait_for_prev_duration_secs': 5.888938903808594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776624005.702762, 'checkpointer_blocking_duration_secs': 1.0330486297607422, 'get_old_steps_start_time': 1776624006.735837, 'get_old_steps_duration_secs': 3.4332275390625e-05, 'checkpoint_manager_blocking_start_time': 1776624005.6987548, 'checkpoint_manager_blocking_duration_secs': 1.0435171127319336} I0419 18:40:06.742418 135763238426432 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0419 18:40:06.742468 135763238426432 max_utils.py:750] Memstats: After params initialized: I0419 18:40:06.742520 135763238426432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0419 18:40:06.742568 135763238426432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0419 18:40:06.742595 135763238426432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0419 18:40:06.742619 135763238426432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0419 18:40:07.049933 135763238426432 metric_logger.py:196] completed step: 10, seconds: 11.134, TFLOP/s/device: 1.220, Tokens/s/device: 183.941, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0419 18:40:07.311055 135763238426432 metric_logger.py:196] completed step: 11, seconds: 1.507, TFLOP/s/device: 9.017, Tokens/s/device: 1359.183, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0419 18:40:07.458902 135763238426432 metric_logger.py:196] completed step: 12, seconds: 0.114, TFLOP/s/device: 119.396, Tokens/s/device: 17996.643, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0419 18:40:07.606163 135763238426432 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.566, Tokens/s/device: 13349.673, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0419 18:40:07.610634 135763238426432 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0419 18:40:07.900290 135763238426432 checkpointing.py:776] Waited 0.28961849212646484 seconds for step 15 to finish before starting checkpointing. I0419 18:40:07.903334 135763238426432 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 18:40:13.204726 135643288024832 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/10/items/array_metadatas/process_4 I0419 18:40:49.672247 135644361766656 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 36.541 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.22862982749939 s) (per-host) I0419 18:40:49.672359 135644361766656 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.936589s. I0419 18:40:59.292078 135644361766656 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 52.556293s. I0419 18:40:59.292396 135643267053312 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 18:40:59.292513 135643267053312 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0419 18:40:59.292582 135643267053312 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 18:40:59.293997 135643267053312 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 18:40:59.294160 135763238426432 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0419 18:40:59.294296 135763238426432 checkpoint_manager.py:1452] Waiting for previous save to complete took 51.390959 seconds. If this number is high, consider checkpointing less frequently. I0419 18:40:59.296437 135763238426432 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 15 I0419 18:40:59.299151 135763238426432 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/15. I0419 18:41:00.024875 135763238426432 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 18:41:00.024968 135763238426432 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 I0419 18:41:00.063112 135763238426432 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039429s I0419 18:41:00.063275 135763238426432 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 35.912 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04295468330383301 s) (per-host) I0419 18:41:00.063325 135763238426432 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.043013s (batch_requests_ready=0.001889s, total_serialization_initiated=0.041057s, others=0.000067s) I0419 18:41:00.063413 135763238426432 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048039s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.048026) I0419 18:41:00.064069 135763238426432 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/15. I0419 18:41:00.064409 135643267053312 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:01:00.064371 I0419 18:41:00.066335 135763238426432 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 18:41:00.066617 135637374052096 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0419 18:41:00.066775 135763238426432 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776624007.9033043, 'wait_for_prev_duration_secs': 51.3909592628479, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776624059.2964756, 'checkpointer_blocking_duration_secs': 0.7680892944335938, 'get_old_steps_start_time': 1776624060.0645876, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776624007.9009554, 'checkpoint_manager_blocking_duration_secs': 52.16578793525696} I0419 18:41:00.066880 135763238426432 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0419 18:41:00.067643 135763238426432 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.930, Tokens/s/device: 13856.657, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0419 18:41:00.073571 135763238426432 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.370, Tokens/s/device: 13922.975, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0419 18:41:05.148369 135643288024832 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/15/items/array_metadatas/process_4 I0419 18:41:21.300782 135763238426432 metric_logger.py:196] completed step: 16, seconds: 52.461, TFLOP/s/device: 0.259, Tokens/s/device: 39.038, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0419 18:41:21.448269 135763238426432 metric_logger.py:196] completed step: 17, seconds: 21.081, TFLOP/s/device: 0.645, Tokens/s/device: 97.148, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0419 18:41:21.595683 135763238426432 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.368, Tokens/s/device: 13470.494, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0419 18:41:21.742336 135763238426432 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0419 18:41:21.743042 135763238426432 checkpointing.py:776] Waited 0.0007269382476806641 seconds for step 19 to finish before starting checkpointing. I0419 18:41:21.745140 135763238426432 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 18:41:41.476048 135643267053312 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 38.103 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.455676317214966 s) (per-host) I0419 18:41:41.476177 135643267053312 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.411654s. I0419 18:41:52.776838 135643267053312 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 52.712300s. I0419 18:41:52.777124 135637374052096 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 18:41:52.777244 135637374052096 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0419 18:41:52.777293 135637374052096 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 18:41:52.778705 135637374052096 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 18:41:52.778890 135763238426432 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0419 18:41:52.779016 135763238426432 checkpoint_manager.py:1452] Waiting for previous save to complete took 31.033882 seconds. If this number is high, consider checkpointing less frequently. I0419 18:41:52.780759 135763238426432 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 19 I0419 18:41:52.782942 135763238426432 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/19. I0419 18:41:53.497416 135763238426432 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 18:41:53.497508 135763238426432 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 I0419 18:41:53.525145 135763238426432 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028553s I0419 18:41:53.525313 135763238426432 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 48.007 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03213214874267578 s) (per-host) I0419 18:41:53.525363 135763238426432 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032193s (batch_requests_ready=0.001762s, total_serialization_initiated=0.030366s, others=0.000066s) I0419 18:41:53.525448 135763238426432 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036294s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.036283) I0419 18:41:53.526162 135763238426432 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/19. I0419 18:41:53.526496 135643288024832 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:01:53.526458 I0419 18:41:53.530547 135763238426432 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 18:41:53.530798 135637374052096 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0419 18:41:53.530927 135763238426432 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776624081.74511, 'wait_for_prev_duration_secs': 31.033881902694702, 'time_between_consecutive_saves_sec': 22.451072931289673, 'checkpointer_blocking_start_time': 1776624112.7807965, 'checkpointer_blocking_duration_secs': 0.7458615303039551, 'get_old_steps_start_time': 1776624113.5266767, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776624081.7432902, 'checkpoint_manager_blocking_duration_secs': 31.787604570388794} I0419 18:41:53.531038 135763238426432 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0419 18:41:53.531080 135763238426432 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 18:41:59.934389 135636293424896 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/19/items/array_metadatas/process_4 I0419 18:42:36.308921 135643288024832 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 36.893 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.815706729888916 s) (per-host) I0419 18:42:36.309030 135643288024832 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.782404s. I0419 18:42:43.897643 135643288024832 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 50.371001s. I0419 18:42:43.897957 135637374052096 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 18:42:43.898089 135637374052096 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0419 18:42:43.898135 135637374052096 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 18:42:43.899497 135637374052096 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 18:42:43.899678 135763238426432 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0419 18:42:43.899827 135763238426432 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 18:42:43.900754 135763238426432 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.124, Tokens/s/device: 13885.970, 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: Sun Apr 19 18:42:56 UTC 2026 EXIT_CODE=0