MaxView

XPK Start: Mon Apr 20 10:06:42 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-20 10:07:06.733774: 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)
I0420 10:07:06.914080 132753142945600 max_utils.py:273] Attempting to initialize the jax distributed system...
I0420 10:07:15.954757 132753142945600 distributed.py:149] Starting JAX distributed service on [::]:8482
I0420 10:07:15.957065 132753142945600 distributed.py:172] Connecting to JAX distributed service on mt-07-eval-ayswe-slice-job-0-0.mt-07-eval-ayswe:8482
I0420 10:07:17.021776 132753142945600 max_utils.py:284] Jax distributed system initialized!
I0420 10:07:23.142708 132753142945600 max_utils.py:800] System Information: Jax Version: 0.9.2
I0420 10:07:23.142806 132753142945600 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0420 10:07:23.142857 132753142945600 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
I0420 10:07:23.142902 132753142945600 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 10:07:23.841476 132753142945600 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 10:07:23.957708 132753142945600 checkpointing.py:688] Setting up checkpoint logger...
I0420 10:07:23.957822 132753142945600 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 10:07:23.957871 132753142945600 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0420 10:07:23.958165 132753142945600 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 0x78bc6bf1c1d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 10:07:27.360535 132753142945600 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 10:07:27.360766 132753142945600 checkpoint_manager.py:708] [process=2][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a71836e540>}, handler_registry=None
I0420 10:07:27.361003 132753142945600 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a71836e540>` 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`.
I0420 10:07:27.361066 132753142945600 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a71836c620>` 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`.
I0420 10:07:27.361102 132753142945600 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 0x78a71836e540>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a71836e540>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a71836c620>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a71836c620>}).
I0420 10:07:27.361421 132753142945600 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0420 10:07:27.361490 132753142945600 async_checkpointer.py:192] [process=2][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78a7182c1b20> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0420 10:07:28.154250 132753142945600 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints
I0420 10:07:28.589335 132753142945600 checkpoint_manager.py:929] [process=2][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/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78a7382b7770>
I0420 10:07:28.589510 132753142945600 checkpointing.py:302] Checkpoint manager created!
I0420 10:07:29.556398 132753142945600 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 10:07:29.556511 132753142945600 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32

I0420 10:07:31.406064 132753142945600 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 10:07:31.406160 132753142945600 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 10:07:31.411943 132753142945600 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 10:07:31.411999 132753142945600 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 10:07:31.428691 132753142945600 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 10:07:31.428751 132753142945600 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 10:07:31.444606 132753142945600 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 10:07:31.444663 132753142945600 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 10:07:31.469949 132753142945600 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 10:07:31.470029 132753142945600 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 10:07:31.485866 132753142945600 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 10:07:31.485929 132753142945600 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 10:07:31.501814 132753142945600 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 10:07:31.501876 132753142945600 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 10:07:31.532741 132753142945600 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 10:07:31.532811 132753142945600 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 10:07:31.553088 132753142945600 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 10:07:31.553155 132753142945600 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 10:07:42.469414 132753142945600 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.
I0420 10:07:42.471825 132753142945600 metric_logger.py:301] number parameters: 1.104 billion
I0420 10:07:53.811386 132753142945600 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 10:07:53.963183 132753142945600 checkpointing.py:798] Waited 0.15178298950195312 seconds for step 0 to finish before starting checkpointing.
I0420 10:07:53.965526 132753142945600 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 10:07:53.967279 132753142945600 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 0
I0420 10:07:53.968710 132753142945600 event_tracking.py:70] [process=2] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints/0.
I0420 10:07:54.624308 132753142945600 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 10:07:54.625297 132753142945600 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 10:07:54.625355 132753142945600 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
I0420 10:07:54.900514 132753142945600 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277023s
I0420 10:07:54.900689 132753142945600 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 5.385 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2864553928375244 s) (per-host)
I0420 10:07:54.900743 132753142945600 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286521s (batch_requests_ready=0.003101s, total_serialization_initiated=0.283347s, others=0.000073s)
I0420 10:07:54.900841 132753142945600 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290637s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.290620)
I0420 10:07:54.901708 132753142945600 event_tracking.py:125] [process=2] [async] Finished blocking save in 0.93 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints/0.
I0420 10:07:54.901998 132624359184128 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 10:27:54.901966
I0420 10:07:54.909843 132753142945600 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 10:07:54.910138 132623771010816 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save.
I0420 10:07:54.910279 132753142945600 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776679673.9655085, 'wait_for_prev_duration_secs': 6.222724914550781e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776679673.9673193, 'checkpointer_blocking_duration_secs': 0.9348447322845459, 'get_old_steps_start_time': 1776679674.90219, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1776679673.9635682, 'checkpoint_manager_blocking_duration_secs': 0.9466674327850342}
I0420 10:07:54.910393 132753142945600 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 10:07:54.910462 132753142945600 max_utils.py:750] 
Memstats: After params initialized:
I0420 10:07:54.910515 132753142945600 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_8(process=2,(0,2,0,0))
I0420 10:07:54.910547 132753142945600 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_9(process=2,(1,2,0,0))
I0420 10:07:54.910579 132753142945600 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_12(process=2,(0,3,0,0))
I0420 10:07:54.910604 132753142945600 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_13(process=2,(1,3,0,0))
I0420 10:07:55.225981 132753142945600 metric_logger.py:196] completed step: 0, seconds: 11.338, TFLOP/s/device: 1.198, Tokens/s/device: 180.625, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0420 10:07:55.400658 132753142945600 metric_logger.py:196] completed step: 1, seconds: 1.413, TFLOP/s/device: 9.618, Tokens/s/device: 1449.758, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0420 10:07:55.852893 132753142945600 metric_logger.py:196] completed step: 2, seconds: 0.029, TFLOP/s/device: 474.543, Tokens/s/device: 71528.360, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133
I0420 10:07:56.000507 132753142945600 metric_logger.py:196] completed step: 3, seconds: 0.442, TFLOP/s/device: 30.748, Tokens/s/device: 4634.701, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789
I0420 10:07:56.296472 132753142945600 metric_logger.py:196] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.649, Tokens/s/device: 12457.724, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594
I0420 10:07:56.304220 132753142945600 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 91.980, Tokens/s/device: 13864.255, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978
I0420 10:07:58.089088    2815 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 10:08:00.109687 132623779403520 array_metadata_store.py:203] [process=2][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints/0/items/array_metadatas/process_2
I0420 10:08:18.550441 132753142945600 metric_logger.py:196] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.843, Tokens/s/device: 6909.885, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156
I0420 10:08:18.698262 132753142945600 metric_logger.py:196] completed step: 7, seconds: 22.100, TFLOP/s/device: 0.615, Tokens/s/device: 92.670, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465
I0420 10:08:18.845872 132753142945600 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.628, Tokens/s/device: 13358.903, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256
I0420 10:08:18.994158 132753142945600 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 10:08:18.994889 132753142945600 checkpointing.py:798] Waited 0.0007412433624267578 seconds for step 9 to finish before starting checkpointing.
I0420 10:08:18.997045 132753142945600 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 10:08:32.372757 132624359184128 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 41.834 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.758503437042236 s) (per-host)
I0420 10:08:32.372864 132624359184128 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.470737s.
I0420 10:08:40.490641 132624359184128 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 45.588494s.
I0420 10:08:40.490928 132623771010816 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 10:08:40.491075 132623771010816 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save.
I0420 10:08:40.491127 132623771010816 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 10:08:40.492908 132623771010816 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 10:08:40.493045 132753142945600 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0420 10:08:40.493127 132753142945600 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.496104 seconds. If this number is high, consider checkpointing less frequently.
I0420 10:08:40.496254 132753142945600 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 9
I0420 10:08:40.498327 132753142945600 event_tracking.py:70] [process=2] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints/9.
I0420 10:08:40.803662 132753142945600 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 10:08:40.803756 132753142945600 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
I0420 10:08:40.836473 132753142945600 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034448s
I0420 10:08:40.836643 132753142945600 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 36.947 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04175114631652832 s) (per-host)
I0420 10:08:40.836695 132753142945600 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.041813s (batch_requests_ready=0.002801s, total_serialization_initiated=0.038944s, others=0.000068s)
I0420 10:08:40.836792 132753142945600 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045694s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.045681)
I0420 10:08:40.837551 132753142945600 event_tracking.py:125] [process=2] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints/9.
I0420 10:08:40.837883 132623771010816 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 10:28:40.837846
I0420 10:08:40.847997 132753142945600 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 10:08:40.848282 132602128201472 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save.
I0420 10:08:40.848441 132753142945600 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776679698.9969966, 'wait_for_prev_duration_secs': 21.49610424041748, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776679720.496295, 'checkpointer_blocking_duration_secs': 0.3417022228240967, 'get_old_steps_start_time': 1776679720.83803, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776679698.9951391, 'checkpoint_manager_blocking_duration_secs': 21.853267431259155}
I0420 10:08:40.848551 132753142945600 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 10:08:40.848595 132753142945600 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 10:08:46.164978 132623779403520 array_metadata_store.py:203] [process=2][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_07_eval/checkpoints/9/items/array_metadatas/process_2
I0420 10:09:22.849786 132623771010816 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 37.560 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.054848194122314 s) (per-host)
I0420 10:09:22.849913 132623771010816 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.011380s.
I0420 10:09:31.283992 132623771010816 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 50.445445s.
I0420 10:09:31.284286 132602128201472 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 10:09:31.284429 132602128201472 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save.
I0420 10:09:31.284478 132602128201472 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 10:09:31.286234 132602128201472 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 10:09:31.286404 132753142945600 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0420 10:09:31.286568 132753142945600 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 10:09:31.287587 132753142945600 metric_logger.py:196] completed step: 9, seconds: 0.149, TFLOP/s/device: 91.365, Tokens/s/device: 13771.493, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3551.438
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Mon Apr 20 10:09:43 UTC 2026
EXIT_CODE=0