XPK Start: Sat Apr 18 00:08:53 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 00:09:18.327773: 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 00:09:18.864995 134515517069120 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 00:09:27.904074 134515517069120 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 00:09:27.906661 134515517069120 distributed.py:172] Connecting to JAX distributed service on mt-03-dropout-28tdt-slice-job-0-0.mt-03-dropout-28tdt:8482
I0418 00:09:28.905533 134515517069120 max_utils.py:284] Jax distributed system initialized!
I0418 00:09:34.954591 134515517069120 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 00:09:34.954694 134515517069120 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 00:09:34.954735 134515517069120 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 00:09:34.954771 134515517069120 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 00:09:35.632255 134515517069120 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 00:09:35.632524 134515517069120 checkpointing.py:677] Setting up checkpoint logger...
I0418 00:09:35.632575 134515517069120 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 00:09:35.632619 134515517069120 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 00:09:35.632960 134515517069120 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 0x7a56cb31de50>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 00:09:39.904072 134515517069120 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 00:09:39.904314 134515517069120 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a416438ab70>}, handler_registry=None
I0418 00:09:39.904554 134515517069120 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a416438ab70>` 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 00:09:39.904602 134515517069120 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a416438e780>` 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 00:09:39.904636 134515517069120 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 0x7a416438ab70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a416438ab70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a416438e780>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a416438e780>}).
I0418 00:09:39.904969 134515517069120 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 00:09:39.905040 134515517069120 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a41406f7ce0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 00:09:43.594601 134515517069120 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints
I0418 00:09:44.020823 134515517069120 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a55926b6e40>
I0418 00:09:44.021006 134515517069120 checkpointing.py:301] Checkpoint manager created!
I0418 00:09:44.347697 134515517069120 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 00:09:44.347813 134515517069120 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0418 00:09:44.725434 134515517069120 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 00:09:44.725523 134515517069120 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0418 00:09:44.741645 134515517069120 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 00:09:44.741706 134515517069120 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0418 00:09:44.765213 134515517069120 attentions.py:1154] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 00:09:44.765281 134515517069120 attentions.py:1154] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0418 00:09:44.781493 134515517069120 attentions.py:1155] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 00:09:44.781552 134515517069120 attentions.py:1155] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0418 00:09:44.797577 134515517069120 attentions.py:1156] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 00:09:44.797635 134515517069120 attentions.py:1156] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0418 00:09:44.822786 134515517069120 attentions.py:1197] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 00:09:44.822866 134515517069120 attentions.py:1197] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0418 00:09:44.849408 134515517069120 linears.py:525] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 00:09:44.849475 134515517069120 linears.py:525] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None).
I0418 00:09:45.073479 134515517069120 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 00:09:45.073590 134515517069120 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 00:09:46.517200 134515517069120 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 00:09:46.517328 134515517069120 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 00:09:46.517380 134515517069120 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 00:09:46.517436 134515517069120 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 00:09:46.517491 134515517069120 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 00:09:46.517529 134515517069120 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 00:09:46.517580 134515517069120 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 00:09:46.517631 134515517069120 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 00:09:46.517671 134515517069120 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 00:09:46.517708 134515517069120 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 00:09:46.517754 134515517069120 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 00:09:46.517801 134515517069120 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 00:09:47.035615 134515517069120 train.py:155] train/xent Logical: float32[32,128]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0418 00:09:47.035712 134515517069120 train.py:155] train/xent Physical: float32[32,128]............................................. ('fsdp', None).
I0418 00:09:47.051007 134515517069120 train.py:162] train/z_loss Logical: float32[32,128]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0418 00:09:47.051069 134515517069120 train.py:162] train/z_loss Physical: float32[32,128]............................................. ('fsdp', None).
I0418 00:09:50.633170 134515517069120 max_utils.py:791] Total memory size: 0.8 GB, Output size: 0.4 GB, Temp size: 0.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0418 00:09:50.640948 134515517069120 metric_logger.py:301] number parameters: 1.104 billion
I0418 00:09:55.027007 134515517069120 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 00:09:55.212514 134515517069120 checkpointing.py:776] Waited 0.1854860782623291 seconds for step 0 to finish before starting checkpointing.
I0418 00:09:55.215270 134515517069120 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:09:55.217461 134515517069120 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 00:09:55.219030 134515517069120 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0.
I0418 00:09:56.339414 134515517069120 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 00:09:56.353276 134515517069120 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 00:09:56.353349 134515517069120 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 00:09:56.593153 134386478556928 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0
I0418 00:09:56.627975 134515517069120 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275382s
I0418 00:09:56.628797 134515517069120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.476 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28169965744018555 s) (per-host)
I0418 00:09:56.628870 134515517069120 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281790s (batch_requests_ready=0.002190s, total_serialization_initiated=0.278858s, others=0.000743s)
I0418 00:09:56.628944 134515517069120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285779s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.285763)
I0418 00:09:56.629752 134515517069120 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.41 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0.
I0418 00:09:56.630115 134385378506496 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 00:29:56.630069
I0418 00:09:56.642407 134515517069120 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 00:09:56.642672 134386486949632 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 00:09:56.642827 134515517069120 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776470995.215251, 'wait_for_prev_duration_secs': 6.365776062011719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776470995.2175007, 'checkpointer_blocking_duration_secs': 1.412766695022583, 'get_old_steps_start_time': 1776470996.6302948, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776470995.213049, 'checkpoint_manager_blocking_duration_secs': 1.4297411441802979}
I0418 00:09:56.642947 134515517069120 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 00:09:56.642999 134515517069120 max_utils.py:750]
Memstats: After params initialized:
I0418 00:09:56.643049 134515517069120 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_0(process=0,(0,0,0,0))
I0418 00:09:56.643083 134515517069120 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_1(process=0,(1,0,0,0))
I0418 00:09:56.643109 134515517069120 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_4(process=0,(0,1,0,0))
I0418 00:09:56.643134 134515517069120 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_5(process=0,(1,1,0,0))
I0418 00:09:56.997865 134515517069120 metric_logger.py:196] completed step: 0, seconds: 4.194, TFLOP/s/device: 0.191, Tokens/s/device: 30.518, total_weights: 4096, loss: 10.889, lm_loss: 10.889, perplexity: 53560.699
I0418 00:09:57.001205 134515517069120 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/tensorboard/'
I0418 00:09:57.806349 134515517069120 metric_logger.py:196] completed step: 1, seconds: 1.969, TFLOP/s/device: 0.407, Tokens/s/device: 65.006, total_weights: 4096, loss: 10.902, lm_loss: 10.902, perplexity: 54298.523
I0418 00:09:57.875196 134515517069120 metric_logger.py:196] completed step: 2, seconds: 0.809, TFLOP/s/device: 0.990, Tokens/s/device: 158.187, total_weights: 4096, loss: 9.956, lm_loss: 9.956, perplexity: 21082.051
I0418 00:09:57.945738 134515517069120 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 84.108, Tokens/s/device: 13442.554, total_weights: 4096, loss: 9.139, lm_loss: 9.139, perplexity: 9314.691
I0418 00:09:59.706390 134386461771520 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/items
I0418 00:09:59.876399 134383822436096 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776470999461610485, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776470999461610485, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:10:01.264874 2820 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 00:10:04.736954 134385433048832 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/items/array_metadatas/process_0
I0418 00:10:10.471171 134515517069120 metric_logger.py:196] completed step: 4, seconds: 0.069, TFLOP/s/device: 11.527, Tokens/s/device: 1842.257, total_weights: 4096, loss: 8.459, lm_loss: 8.459, perplexity: 4715.493
I0418 00:10:10.480270 134515517069120 metric_logger.py:196] completed step: 5, seconds: 0.071, TFLOP/s/device: 11.348, Tokens/s/device: 1813.725, total_weights: 4096, loss: 7.924, lm_loss: 7.924, perplexity: 2762.373
I0418 00:10:10.549869 134515517069120 metric_logger.py:196] completed step: 6, seconds: 12.524, TFLOP/s/device: 0.064, Tokens/s/device: 10.220, total_weights: 4096, loss: 7.553, lm_loss: 7.553, perplexity: 1906.617
I0418 00:10:10.622342 134515517069120 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 122.159, Tokens/s/device: 19524.100, total_weights: 4096, loss: 7.267, lm_loss: 7.267, perplexity: 1431.653
I0418 00:10:10.692856 134515517069120 metric_logger.py:196] completed step: 8, seconds: 0.070, TFLOP/s/device: 11.366, Tokens/s/device: 1816.608, total_weights: 4096, loss: 7.111, lm_loss: 7.111, perplexity: 1225.983
I0418 00:10:10.764097 134515517069120 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 00:10:10.764640 134515517069120 checkpointing.py:776] Waited 0.0005621910095214844 seconds for step 9 to finish before starting checkpointing.
I0418 00:10:10.766786 134515517069120 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 00:10:34.274080 134385386899200 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 33.007566s (commit=30.993762s, array_metadata_write=2.013804s)
I0418 00:10:34.275122 134385378506496 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 41.647 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.928016662597656 s) (per-host)
I0418 00:10:34.275191 134385378506496 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.644963s.
I0418 00:10:35.811470 134385378506496 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776470999461610485, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:10:36.836992 134383822436096 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776470999461610485, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:10:37.443979 134385378506496 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000261s.
I0418 00:10:37.743940 134385378506496 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 00:10:38.687316 134385378506496 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.753287s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/items
I0418 00:10:38.690433 134385378506496 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0/items
I0418 00:10:39.241556 134385378506496 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0
I0418 00:10:41.500242 134385378506496 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0`.
I0418 00:10:41.501093 134385378506496 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 46.28 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/0
I0418 00:10:41.502737 134385378506496 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 44.872507s.
I0418 00:10:41.502942 134386486949632 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 00:10:41.503111 134386486949632 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 00:10:41.503188 134386486949632 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 00:10:41.505024 134386486949632 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 00:10:41.505188 134515517069120 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0418 00:10:41.505320 134515517069120 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.738532 seconds. If this number is high, consider checkpointing less frequently.
I0418 00:10:41.507272 134515517069120 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 00:10:41.509322 134515517069120 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9.
I0418 00:10:42.651767 134515517069120 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 00:10:42.651874 134515517069120 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 00:10:42.689347 134515517069120 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038303s
I0418 00:10:42.690164 134515517069120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 36.050 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.042789459228515625 s) (per-host)
I0418 00:10:42.690235 134515517069120 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042878s (batch_requests_ready=0.001795s, total_serialization_initiated=0.040348s, others=0.000735s)
I0418 00:10:42.690313 134515517069120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.046887s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.046872)
I0418 00:10:42.690983 134515517069120 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.18 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9.
I0418 00:10:42.691309 134385416263424 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 00:30:42.691270
I0418 00:10:42.693251 134515517069120 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 00:10:42.693501 134386486949632 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 00:10:42.693628 134515517069120 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776471010.7667558, 'wait_for_prev_duration_secs': 30.738531827926636, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776471041.5073113, 'checkpointer_blocking_duration_secs': 1.1841351985931396, 'get_old_steps_start_time': 1776471042.6914697, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776471010.7648928, 'checkpoint_manager_blocking_duration_secs': 31.928702354431152}
I0418 00:10:42.693794 134515517069120 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 00:10:42.693853 134515517069120 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 00:10:43.318348 134385378506496 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9
I0418 00:10:46.826145 134385931908864 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9/items
I0418 00:10:51.387240 134385923516160 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9/items/array_metadatas/process_0
I0418 00:11:30.987370 134385433048832 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.203005s (commit=39.730301s, array_metadata_write=2.472704s)
I0418 00:11:30.989621 134385416263424 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 32.675 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 48.34224271774292 s) (per-host)
I0418 00:11:30.989690 134385416263424 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 48.298269s.
I0418 00:11:34.415210 134385416263424 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 00:11:34.624038 134385416263424 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000208s.
I0418 00:11:35.491578 134385416263424 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.346712s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9/items
I0418 00:11:35.494015 134385416263424 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9/items
I0418 00:11:36.028380 134385416263424 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9
I0418 00:11:38.315761 134385416263424 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9`.
I0418 00:11:38.316604 134385416263424 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 56.81 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_03_dropout/checkpoints/9
I0418 00:11:38.318054 134385416263424 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 55.626630s.
I0418 00:11:38.318256 134386486949632 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 00:11:38.318365 134386486949632 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 00:11:38.318429 134386486949632 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 00:11:38.321364 134386486949632 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 00:11:38.321468 134515517069120 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0418 00:11:38.321541 134515517069120 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:11:38.322537 134515517069120 metric_logger.py:196] completed step: 9, seconds: 0.073, TFLOP/s/device: 11.029, Tokens/s/device: 1762.745, total_weights: 4096, loss: 6.981, lm_loss: 6.981, perplexity: 1075.867
Per train step:
Total TFLOPs: 0.80
split as 99.60% learnable weight flops and 0.40% attention flops
XPK End: Sat Apr 18 00:11:49 UTC 2026
EXIT_CODE=0