MaxView

XPK Start: Mon Apr 20 09:54:36 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 09:55:00.746708: 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 09:55:00.926326 132876582528832 max_utils.py:273] Attempting to initialize the jax distributed system...
I0420 09:55:09.967081 132876582528832 distributed.py:149] Starting JAX distributed service on [::]:8482
I0420 09:55:09.969547 132876582528832 distributed.py:172] Connecting to JAX distributed service on mt-03-dropout-mehea-slice-job-0-0.mt-03-dropout-mehea:8482
I0420 09:55:11.497861 132876582528832 max_utils.py:284] Jax distributed system initialized!
I0420 09:55:17.940822 132876582528832 max_utils.py:800] System Information: Jax Version: 0.9.2
I0420 09:55:17.940921 132876582528832 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0420 09:55:17.940962 132876582528832 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 09:55:17.940998 132876582528832 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 09:55:19.012386 132876582528832 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 09:55:19.127835 132876582528832 checkpointing.py:688] Setting up checkpoint logger...
I0420 09:55:19.127953 132876582528832 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 09:55:19.127998 132876582528832 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0420 09:55:19.128214 132876582528832 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 0x78d91c914470>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 09:55:22.148800 132876582528832 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 09:55:22.149047 132876582528832 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 0x78d7f9fe69c0>}, handler_registry=None
I0420 09:55:22.149296 132876582528832 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78d7f9fe69c0>` 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 09:55:22.149346 132876582528832 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78c51c12a7e0>` 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 09:55:22.149382 132876582528832 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 0x78d7f9fe69c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78d7f9fe69c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78c51c12a7e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78c51c12a7e0>}).
I0420 09:55:22.149693 132876582528832 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0420 09:55:22.149762 132876582528832 async_checkpointer.py:192] [process=2][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78c4146d0fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0420 09:55:23.580363 132876582528832 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_03_dropout/checkpoints
I0420 09:55:24.008584 132876582528832 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_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78c51c10eea0>
I0420 09:55:24.008757 132876582528832 checkpointing.py:302] Checkpoint manager created!
I0420 09:55:24.388612 132876582528832 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 09:55:24.388728 132876582528832 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32

I0420 09:55:26.200598 132876582528832 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 09:55:26.200692 132876582528832 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 09:55:26.205926 132876582528832 nnx_decoders.py:465] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 09:55:26.205983 132876582528832 nnx_decoders.py:465] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 09:55:26.222509 132876582528832 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 09:55:26.222566 132876582528832 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 09:55:26.238478 132876582528832 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 09:55:26.238534 132876582528832 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 09:55:26.267643 132876582528832 attentions.py:1154] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 09:55:26.267735 132876582528832 attentions.py:1154] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 09:55:26.283756 132876582528832 attentions.py:1155] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 09:55:26.283815 132876582528832 attentions.py:1155] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 09:55:26.299810 132876582528832 attentions.py:1156] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 09:55:26.299868 132876582528832 attentions.py:1156] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 09:55:26.330172 132876582528832 attentions.py:1197] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 09:55:26.330248 132876582528832 attentions.py:1197] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 09:55:26.357752 132876582528832 linears.py:525] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 09:55:26.357821 132876582528832 linears.py:525] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None).
I0420 09:55:30.081679 132876582528832 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.
I0420 09:55:30.084082 132876582528832 metric_logger.py:301] number parameters: 1.104 billion
I0420 09:55:34.148683 132876582528832 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 09:55:34.221530 132876582528832 checkpointing.py:798] Waited 0.07283639907836914 seconds for step 0 to finish before starting checkpointing.
I0420 09:55:34.223906 132876582528832 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 09:55:34.225947 132876582528832 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 0
I0420 09:55:34.227915 132876582528832 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_03_dropout/checkpoints/0.
I0420 09:55:34.945411 132876582528832 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 09:55:34.946440 132876582528832 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 09:55:34.946500 132876582528832 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 09:55:35.226735 132876582528832 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.282087s
I0420 09:55:35.226908 132876582528832 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 5.299 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2911038398742676 s) (per-host)
I0420 09:55:35.226960 132876582528832 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.291165s (batch_requests_ready=0.003021s, total_serialization_initiated=0.288075s, others=0.000069s)
I0420 09:55:35.227084 132876582528832 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.295444s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.295427)
I0420 09:55:35.227891 132876582528832 event_tracking.py:125] [process=2] [async] Finished blocking save in 1.00 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_03_dropout/checkpoints/0.
I0420 09:55:35.228204 132748832397056 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 10:15:35.228170
I0420 09:55:35.239461 132876582528832 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 09:55:35.239695 132747760314112 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save.
I0420 09:55:35.239815 132876582528832 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_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776678934.2238884, 'wait_for_prev_duration_secs': 5.91278076171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776678934.2259855, 'checkpointer_blocking_duration_secs': 1.0023713111877441, 'get_old_steps_start_time': 1776678935.228385, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776678934.2219057, 'checkpoint_manager_blocking_duration_secs': 1.017871618270874}
I0420 09:55:35.239917 132876582528832 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 09:55:35.239984 132876582528832 max_utils.py:750] 
Memstats: After params initialized:
I0420 09:55:35.240046 132876582528832 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_8(process=2,(0,2,0,0))
I0420 09:55:35.240094 132876582528832 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_9(process=2,(1,2,0,0))
I0420 09:55:35.240124 132876582528832 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_12(process=2,(0,3,0,0))
I0420 09:55:35.240150 132876582528832 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_13(process=2,(1,3,0,0))
I0420 09:55:35.553933 132876582528832 metric_logger.py:196] completed step: 0, seconds: 4.064, TFLOP/s/device: 0.197, Tokens/s/device: 31.500, total_weights: 4096, loss: 10.874, lm_loss: 10.874, perplexity: 52805.141
I0420 09:55:35.650070 132876582528832 metric_logger.py:196] completed step: 1, seconds: 1.403, TFLOP/s/device: 0.571, Tokens/s/device: 91.215, total_weights: 4096, loss: 10.892, lm_loss: 10.892, perplexity: 53764.488
I0420 09:55:36.053890 132876582528832 metric_logger.py:196] completed step: 2, seconds: 0.027, TFLOP/s/device: 29.656, Tokens/s/device: 4739.687, total_weights: 4096, loss: 9.969, lm_loss: 9.969, perplexity: 21350.100
I0420 09:55:36.124402 132876582528832 metric_logger.py:196] completed step: 3, seconds: 0.394, TFLOP/s/device: 2.034, Tokens/s/device: 325.158, total_weights: 4096, loss: 9.125, lm_loss: 9.125, perplexity: 9186.262
I0420 09:55:36.266247 132876582528832 metric_logger.py:196] completed step: 4, seconds: 0.087, TFLOP/s/device: 9.183, Tokens/s/device: 1467.671, total_weights: 4096, loss: 8.469, lm_loss: 8.469, perplexity: 4765.735
I0420 09:55:36.273695 132876582528832 metric_logger.py:196] completed step: 5, seconds: 0.070, TFLOP/s/device: 11.403, Tokens/s/device: 1822.427, total_weights: 4096, loss: 7.908, lm_loss: 7.908, perplexity: 2719.938
I0420 09:55:38.150903    2803 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 09:55:40.611384 132747768706816 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_03_dropout/checkpoints/0/items/array_metadatas/process_2
I0420 09:55:46.448227 132876582528832 metric_logger.py:196] completed step: 6, seconds: 0.142, TFLOP/s/device: 5.622, Tokens/s/device: 898.574, total_weights: 4096, loss: 7.554, lm_loss: 7.554, perplexity: 1909.063
I0420 09:55:46.518915 132876582528832 metric_logger.py:196] completed step: 7, seconds: 10.105, TFLOP/s/device: 0.079, Tokens/s/device: 12.667, total_weights: 4096, loss: 7.313, lm_loss: 7.313, perplexity: 1500.206
I0420 09:55:46.589523 132876582528832 metric_logger.py:196] completed step: 8, seconds: 0.076, TFLOP/s/device: 10.521, Tokens/s/device: 1681.445, total_weights: 4096, loss: 7.091, lm_loss: 7.091, perplexity: 1200.912
I0420 09:55:46.660857 132876582528832 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 09:55:46.661514 132876582528832 checkpointing.py:798] Waited 0.0006682872772216797 seconds for step 9 to finish before starting checkpointing.
I0420 09:55:46.663600 132876582528832 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 09:56:07.478830 132748832397056 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 48.539 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 32.54299187660217 s) (per-host)
I0420 09:56:07.478937 132748832397056 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.250617s.
I0420 09:56:17.066179 132748832397056 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 41.837841s.
I0420 09:56:17.066432 132747760314112 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 09:56:17.066497 132747760314112 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save.
I0420 09:56:17.066565 132747760314112 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 09:56:17.068511 132747760314112 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 09:56:17.068677 132876582528832 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 09:56:17.068808 132876582528832 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.405211 seconds. If this number is high, consider checkpointing less frequently.
I0420 09:56:17.070912 132876582528832 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 9
I0420 09:56:17.073078 132876582528832 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_03_dropout/checkpoints/9.
I0420 09:56:17.789929 132876582528832 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 09:56:17.790032 132876582528832 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 09:56:17.824619 132876582528832 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036434s
I0420 09:56:17.824765 132876582528832 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 35.475 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04348301887512207 s) (per-host)
I0420 09:56:17.824817 132876582528832 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.043543s (batch_requests_ready=0.002688s, total_serialization_initiated=0.040788s, others=0.000067s)
I0420 09:56:17.824909 132876582528832 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.047586s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.047570)
I0420 09:56:17.825617 132876582528832 event_tracking.py:125] [process=2] [async] Finished blocking save in 0.75 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_03_dropout/checkpoints/9.
I0420 09:56:17.825897 132747760314112 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 10:16:17.825864
I0420 09:56:17.828958 132876582528832 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 09:56:17.829255 132747768706816 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save.
I0420 09:56:17.829416 132876582528832 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_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776678946.6635687, 'wait_for_prev_duration_secs': 30.405210733413696, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776678977.0709622, 'checkpointer_blocking_duration_secs': 0.7551009654998779, 'get_old_steps_start_time': 1776678977.8260858, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776678946.6617439, 'checkpoint_manager_blocking_duration_secs': 31.16763687133789}
I0420 09:56:17.829541 132876582528832 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 09:56:17.829584 132876582528832 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 09:56:23.396596 132741814339328 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_03_dropout/checkpoints/9/items/array_metadatas/process_2
I0420 09:56:58.369312 132747760314112 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 38.918 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.58798861503601 s) (per-host)
I0420 09:56:58.369440 132747760314112 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.543431s.
I0420 09:57:08.016954 132747760314112 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 50.190932s.
I0420 09:57:08.017261 132747768706816 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 09:57:08.017379 132747768706816 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save.
I0420 09:57:08.017427 132747768706816 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 09:57:08.019129 132747768706816 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 09:57:08.019340 132876582528832 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 09:57:08.019527 132876582528832 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 09:57:08.020606 132876582528832 metric_logger.py:196] completed step: 9, seconds: 0.071, TFLOP/s/device: 11.306, Tokens/s/device: 1807.016, total_weights: 4096, loss: 7.016, lm_loss: 7.016, perplexity: 1113.843
Per train step:
 Total TFLOPs: 0.80 
 split as 99.60% learnable weight flops and 0.40% attention flops
XPK End: Mon Apr 20 09:57:16 UTC 2026
EXIT_CODE=0