XPK Start: Sat Apr 18 18:33:03 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 18:33:27.853020: 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 18:33:28.381724 137847920035648 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 18:33:37.424176 137847920035648 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 18:33:37.426787 137847920035648 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--8vjyl-slice-job-0-0.mt-08-checkpoint-async--8vjyl:8482
I0418 18:33:38.438400 137847920035648 max_utils.py:284] Jax distributed system initialized!
I0418 18:33:43.601900 137847920035648 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 18:33:43.601999 137847920035648 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 18:33:43.602039 137847920035648 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 18:33:43.602084 137847920035648 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 18:33:44.297493 137847920035648 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 18:33:44.297768 137847920035648 checkpointing.py:677] Setting up checkpoint logger...
I0418 18:33:44.297824 137847920035648 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 18:33:44.297874 137847920035648 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 18:33:44.298233 137847920035648 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 0x7d5e80d38aa0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 18:33:45.942964 137847920035648 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 18:33:45.943490 137847920035648 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 0x7d5d750f2330>}, handler_registry=None
I0418 18:33:45.943815 137847920035648 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d5d750f2330>` 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 18:33:45.943866 137847920035648 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d4ae82472f0>` 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 18:33:45.943903 137847920035648 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 0x7d5d750f2330>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d5d750f2330>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d4ae82472f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d4ae82472f0>}).
I0418 18:33:45.944257 137847920035648 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 18:33:45.944328 137847920035648 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d4ac86ef600> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 18:33:48.731057 137727177647872 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776537055405831817, 'commit_timestamp_nsecs': 1776537103052477371, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9/_CHECKPOINT_METADATA
I0418 18:33:49.191414 137727177647872 checkpoint_manager.py:1871] Missing metrics for step 9
I0418 18:33:49.817632 137847920035648 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints
I0418 18:33:51.151622 137847920035648 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 0x7d4ae8240ec0>
I0418 18:33:51.151795 137847920035648 checkpointing.py:301] Checkpoint manager created!
I0418 18:33:52.075474 137847920035648 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 18:33:52.075590 137847920035648 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 18:33:52.453515 137847920035648 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 18:33:52.453606 137847920035648 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 18:33:52.469862 137847920035648 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 18:33:52.469922 137847920035648 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 18:33:52.494181 137847920035648 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 18:33:52.494248 137847920035648 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 18:33:52.510545 137847920035648 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 18:33:52.510607 137847920035648 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 18:33:52.526691 137847920035648 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 18:33:52.526750 137847920035648 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 18:33:52.551372 137847920035648 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 18:33:52.551444 137847920035648 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 18:33:52.572393 137847920035648 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 18:33:52.572457 137847920035648 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 18:33:52.796594 137847920035648 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 18:33:52.796702 137847920035648 checkpointing.py:581] restoring from this run's directory step 9
I0418 18:33:53.474409 137847920035648 event_tracking.py:70] [process=7] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9.
I0418 18:33:54.404339 137847920035648 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9.
I0418 18:33:54.404759 137847920035648 event_tracking.py:125] [process=7] [sync] Finished blocking load in 0.93 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9.
I0418 18:33:57.612554 1446 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 18:34:04.482312 137847920035648 jax_array_handlers.py:843] [process=7] /jax/orbax/read/worker/io/requested throughput: 229.829 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.875851631164551 s) (per-host)
I0418 18:34:04.487152 137847920035648 base_pytree_checkpoint_handler.py:130] [process=7] /jax/checkpoint/read/gbytes_per_sec: 205.412 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.70447850227356 s) (per-host)
I0418 18:34:04.599676 137847920035648 event_tracking.py:138] [process=7] [sync] Finished load in 11.13 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/9
I0418 18:34:04.599841 137847920035648 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776537233.4743109, 'checkpointer_duration_secs': 11.12549877166748, 'checkpoint_manager_start_time': 1776537232.7969484, 'checkpoint_manager_duration_secs': 11.802863597869873}
fsdp: 32
I0418 18:34:05.019143 137847920035648 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 18:34:05.019268 137847920035648 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 18:34:05.019325 137847920035648 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 18:34:05.019384 137847920035648 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 18:34:05.019435 137847920035648 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 18:34:05.019472 137847920035648 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 18:34:05.019524 137847920035648 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 18:34:05.019576 137847920035648 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 18:34:05.019616 137847920035648 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 18:34:05.019652 137847920035648 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 18:34:05.019697 137847920035648 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 18:34:05.019743 137847920035648 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 18:34:05.504129 137847920035648 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 18:34:05.504225 137847920035648 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 18:34:05.519479 137847920035648 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 18:34:05.519538 137847920035648 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 18:34:16.413207 137847920035648 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 18:34:16.414027 137847920035648 metric_logger.py:301] number parameters: 1.104 billion
I0418 18:34:27.828023 137847920035648 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0418 18:34:27.981729 137847920035648 checkpointing.py:776] Waited 0.15368342399597168 seconds for step 10 to finish before starting checkpointing.
I0418 18:34:27.984490 137847920035648 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 18:34:27.986517 137847920035648 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 10
I0418 18:34:27.988293 137847920035648 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/10.
I0418 18:34:28.736420 137847920035648 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 18:34:28.737320 137847920035648 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 18:34:28.737375 137847920035648 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 18:34:29.018396 137847920035648 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.282051s
I0418 18:34:29.018575 137847920035648 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.389 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28623104095458984 s) (per-host)
I0418 18:34:29.018630 137847920035648 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286297s (batch_requests_ready=0.002246s, total_serialization_initiated=0.283975s, others=0.000075s)
I0418 18:34:29.018723 137847920035648 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290435s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.290417)
I0418 18:34:29.019478 137847920035648 event_tracking.py:125] [process=7] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/10.
I0418 18:34:29.019787 137727144077056 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 18:54:29.019750
I0418 18:34:29.024242 137847920035648 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 18:34:29.024508 137725206312704 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0418 18:34:29.024653 137847920035648 standard_logger.py:34] {'step': 10, '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': 1776537267.9844723, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776537267.9865556, 'checkpointer_blocking_duration_secs': 1.0333905220031738, 'get_old_steps_start_time': 1776537269.019972, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776537267.9822428, 'checkpoint_manager_blocking_duration_secs': 1.0423762798309326}
I0418 18:34:29.024762 137847920035648 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0418 18:34:29.024810 137847920035648 max_utils.py:750]
Memstats: After params initialized:
I0418 18:34:29.024857 137847920035648 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0))
I0418 18:34:29.024888 137847920035648 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0))
I0418 18:34:29.024916 137847920035648 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0))
I0418 18:34:29.024940 137847920035648 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0))
I0418 18:34:29.336672 137847920035648 metric_logger.py:196] completed step: 10, seconds: 11.414, TFLOP/s/device: 1.190, Tokens/s/device: 179.431, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0418 18:34:29.521330 137847920035648 metric_logger.py:196] completed step: 11, seconds: 1.507, TFLOP/s/device: 9.015, Tokens/s/device: 1358.891, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0418 18:34:29.668805 137847920035648 metric_logger.py:196] completed step: 12, seconds: 0.037, TFLOP/s/device: 363.273, Tokens/s/device: 54756.430, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0418 18:34:29.816143 137847920035648 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.706, Tokens/s/device: 13370.677, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0418 18:34:29.820447 137847920035648 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0418 18:34:30.110246 137847920035648 checkpointing.py:776] Waited 0.2897622585296631 seconds for step 15 to finish before starting checkpointing.
I0418 18:34:30.113471 137847920035648 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 18:34:35.585212 137725214705408 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/10/items/array_metadatas/process_7
I0418 18:35:12.881748 137727144077056 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 35.778 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.14936971664429 s) (per-host)
I0418 18:35:12.881872 137727144077056 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.861977s.
I0418 18:35:22.985413 137727144077056 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 53.965501s.
I0418 18:35:22.985708 137725206312704 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 18:35:22.985826 137725206312704 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0418 18:35:22.985888 137725206312704 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0418 18:35:22.987403 137725206312704 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0418 18:35:22.987581 137847920035648 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0418 18:35:22.987719 137847920035648 checkpoint_manager.py:1452] Waiting for previous save to complete took 52.874246 seconds. If this number is high, consider checkpointing less frequently.
I0418 18:35:22.989871 137847920035648 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 15
I0418 18:35:22.992207 137847920035648 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/15.
I0418 18:35:23.710903 137847920035648 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 18:35:23.710997 137847920035648 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 18:35:23.736687 137847920035648 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026796s
I0418 18:35:23.736873 137847920035648 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 50.778 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03037881851196289 s) (per-host)
I0418 18:35:23.736927 137847920035648 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030451s (batch_requests_ready=0.001894s, total_serialization_initiated=0.028476s, others=0.000081s)
I0418 18:35:23.737023 137847920035648 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035553s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.035540)
I0418 18:35:23.737764 137847920035648 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/15.
I0418 18:35:23.738103 137725206312704 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 18:55:23.738057
I0418 18:35:23.749223 137847920035648 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 18:35:23.749483 137721692940032 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0418 18:35:23.749640 137847920035648 standard_logger.py:34] {'step': 15, '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': 1776537270.1134424, 'wait_for_prev_duration_secs': 52.87424612045288, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776537322.9899154, 'checkpointer_blocking_duration_secs': 0.7483294010162354, 'get_old_steps_start_time': 1776537323.7382689, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776537270.1108596, 'checkpoint_manager_blocking_duration_secs': 53.638747453689575}
I0418 18:35:23.749750 137847920035648 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0418 18:35:23.750664 137847920035648 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.203, Tokens/s/device: 13897.843, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0418 18:35:23.756923 137847920035648 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.252, Tokens/s/device: 13905.297, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0418 18:35:30.478736 137725214705408 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/15/items/array_metadatas/process_7
I0418 18:35:45.509814 137847920035648 metric_logger.py:196] completed step: 16, seconds: 53.935, TFLOP/s/device: 0.252, Tokens/s/device: 37.972, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0418 18:35:45.657498 137847920035648 metric_logger.py:196] completed step: 17, seconds: 21.607, TFLOP/s/device: 0.629, Tokens/s/device: 94.784, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0418 18:35:45.805046 137847920035648 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.318, Tokens/s/device: 13463.055, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0418 18:35:45.951828 137847920035648 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0418 18:35:45.952572 137847920035648 checkpointing.py:776] Waited 0.0007646083831787109 seconds for step 19 to finish before starting checkpointing.
I0418 18:35:45.954615 137847920035648 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 18:36:06.306602 137725206312704 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.080 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.60008144378662 s) (per-host)
I0418 18:36:06.306712 137725206312704 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.568498s.
I0418 18:36:14.506515 137725206312704 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 50.768284s.
I0418 18:36:14.506796 137721692940032 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 18:36:14.506920 137721692940032 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0418 18:36:14.506968 137721692940032 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0418 18:36:14.509443 137721692940032 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0418 18:36:14.509611 137847920035648 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0418 18:36:14.509732 137847920035648 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.555124 seconds. If this number is high, consider checkpointing less frequently.
I0418 18:36:14.511178 137847920035648 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 19
I0418 18:36:14.513114 137847920035648 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/19.
I0418 18:36:15.229169 137847920035648 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 18:36:15.229263 137847920035648 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 18:36:15.255196 137847920035648 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026986s
I0418 18:36:15.255322 137847920035648 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 50.461 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030569791793823242 s) (per-host)
I0418 18:36:15.255380 137847920035648 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030634s (batch_requests_ready=0.001788s, total_serialization_initiated=0.028778s, others=0.000068s)
I0418 18:36:15.255470 137847920035648 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035262s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.035251)
I0418 18:36:15.256220 137847920035648 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_08_checkpoint_async_true/checkpoints/19.
I0418 18:36:15.256544 137721692940032 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 18:56:15.256508
I0418 18:36:15.681108 137847920035648 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 18:36:15.681468 137722541025024 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0418 18:36:15.681638 137847920035648 standard_logger.py:34] {'step': 19, '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': 1776537345.9545863, 'wait_for_prev_duration_secs': 28.555123567581177, 'time_between_consecutive_saves_sec': 22.967143297195435, 'checkpointer_blocking_start_time': 1776537374.5112166, 'checkpointer_blocking_duration_secs': 0.7454667091369629, 'get_old_steps_start_time': 1776537375.2567012, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776537345.9528065, 'checkpoint_manager_blocking_duration_secs': 29.728797912597656}
I0418 18:36:15.681759 137847920035648 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0418 18:36:15.681802 137847920035648 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 18:36:22.296629 137725214705408 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/19/items/array_metadatas/process_7
I0418 18:36:58.074257 137721692940032 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 36.864 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.84945726394653 s) (per-host)
I0418 18:36:58.074365 137721692940032 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.817711s.
I0418 18:37:09.340821 137721692940032 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 54.084151s.
I0418 18:37:09.341133 137722541025024 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 18:37:09.341249 137722541025024 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0418 18:37:09.341289 137722541025024 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0418 18:37:09.343876 137722541025024 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0418 18:37:09.344047 137847920035648 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0418 18:37:09.344214 137847920035648 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 18:37:09.345130 137847920035648 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.026, Tokens/s/device: 13871.110, 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 18:37:18 UTC 2026
EXIT_CODE=0