MaxView

← Back to run

Log Summary

XPK Start: Sun Apr 19 05:40:01 UTC 2026
2026-04-19 05:40:05.790324: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1776577205.803201      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776577205.807129      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776577205.818451      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776577205.818469      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776577205.818472      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776577205.818474      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-19 05:40:24.909943: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0419 05:40:25.439411 137579062097728 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-19 05:40:34,480:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0419 05:40:34.480662 137579062097728 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-19 05:40:34,482:jax._src.distributed:157: Connecting to JAX distributed service on mt-13-scan-layers-true-eby1t-slice-job-0-0.mt-13-scan-layers-true-eby1t:8482
I0419 05:40:34.482941 137579062097728 distributed.py:157] Connecting to JAX distributed service on mt-13-scan-layers-true-eby1t-slice-job-0-0.mt-13-scan-layers-true-eby1t:8482
I0419 05:40:36.362952 137579062097728 max_utils.py:284] Jax distributed system initialized!
I0419 05:40:43.168860 137579062097728 max_utils.py:800] System Information: Jax Version: 0.8.1
I0419 05:40:43.168967 137579062097728 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0419 05:40:43.169007 137579062097728 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709
I0419 05:40:43.169041 137579062097728 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 05:40:43.825979 137579062097728 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 05:40:43.938530 137579062097728 checkpointing.py:688] Setting up checkpoint logger...
I0419 05:40:43.938642 137579062097728 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 05:40:43.938685 137579062097728 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0419 05:40:43.938879 137579062097728 base_pytree_checkpoint_handler.py:415] 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 0x7d200c354da0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 05:40:47.563124 137579062097728 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0419 05:40:47.563388 137579062097728 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d0c6f959e50>}, handler_registry=None
I0419 05:40:47.563627 137579062097728 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d0c6f959e50>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0419 05:40:47.563676 137579062097728 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d0c6f958170>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0419 05:40:47.563713 137579062097728 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 0x7d0c6f959e50>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d0c6f959e50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d0c6f958170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d0c6f958170>}).
I0419 05:40:47.564047 137579062097728 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0419 05:40:47.564124 137579062097728 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d0c6f9a0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0419 05:40:49.654669 137579062097728 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints
I0419 05:40:49.662381 137579062097728 checkpoint_manager.py:929] [process=6][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_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d0ba8199e80>
I0419 05:40:49.662499 137579062097728 checkpointing.py:302] Checkpoint manager created!
I0419 05:40:50.596129 137579062097728 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 05:40:50.596275 137579062097728 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 05:40:52.004963 137579062097728 maxtext_utils.py:1749]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0419 05:40:52.005086 137579062097728 maxtext_utils.py:1749]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0419 05:40:52.005142 137579062097728 maxtext_utils.py:1749]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0419 05:40:52.005187 137579062097728 maxtext_utils.py:1749]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0419 05:40:52.005224 137579062097728 maxtext_utils.py:1749]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0419 05:40:52.005280 137579062097728 maxtext_utils.py:1749]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0419 05:40:52.005319 137579062097728 maxtext_utils.py:1749]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 05:40:52.005355 137579062097728 maxtext_utils.py:1749]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0419 05:40:52.005390 137579062097728 maxtext_utils.py:1749]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 05:40:52.005423 137579062097728 maxtext_utils.py:1749]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 05:40:52.005459 137579062097728 maxtext_utils.py:1749]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0419 05:40:52.005494 137579062097728 maxtext_utils.py:1749]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0419 05:40:52.402820 137579062097728 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 05:40:52.402970 137579062097728 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 05:40:52.408457 137579062097728 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 05:40:52.408560 137579062097728 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 05:40:52.563460 137579062097728 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 05:40:52.563596 137579062097728 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 05:40:52.716920 137579062097728 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 05:40:52.717056 137579062097728 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 05:40:52.879664 137579062097728 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 05:40:52.879803 137579062097728 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 05:40:53.026416 137579062097728 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 05:40:53.026554 137579062097728 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 05:40:53.170758 137579062097728 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 05:40:53.170895 137579062097728 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 05:40:53.327190 137579062097728 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 05:40:53.327366 137579062097728 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 05:40:53.349660 137579062097728 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 05:40:53.349758 137579062097728 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 05:41:04.111167 137579062097728 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0419 05:41:04.113621 137579062097728 metric_logger.py:289] number parameters: 1.104 billion
I0419 05:41:15.127252 137579062097728 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0419 05:41:15.295954 137579062097728 checkpointing.py:798] Waited 0.1687002182006836 seconds for step 0 to finish before starting checkpointing.
I0419 05:41:15.298331 137579062097728 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 05:41:15.300124 137579062097728 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0
I0419 05:41:15.302931 137579062097728 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints/0.
I0419 05:41:16.254077 137579062097728 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 05:41:16.255325 137579062097728 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0419 05:41:16.255386 137579062097728 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0419 05:41:16.504750 137579062097728 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251795s
I0419 05:41:16.504920 137579062097728 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.055 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5049099922180176 s) (per-host)
I0419 05:41:16.504971 137579062097728 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.504972s (batch_requests_ready=0.246622s, total_serialization_initiated=0.258280s, others=0.000070s)
I0419 05:41:16.505066 137579062097728 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.509199s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.509179)
I0419 05:41:16.506077 137456108160768 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0419 05:41:16.506266 137579062097728 async_checkpointer.py:561] Finished blocking save. Time taken: 1.206056s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints/0.
I0419 05:41:16.956622 137579062097728 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 05:41:16.956996 137451813193472 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 05:41:16.957172 137579062097728 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776577275.298313, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776577275.3001623, 'checkpointer_blocking_duration_secs': 1.2062156200408936, 'get_old_steps_start_time': 1776577276.506399, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1776577275.2963731, 'checkpoint_manager_blocking_duration_secs': 1.6607568264007568}
I0419 05:41:16.957296 137579062097728 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0419 05:41:16.957377 137579062097728 max_utils.py:750] 
Memstats: After params initialized:
I0419 05:41:16.957437 137579062097728 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0419 05:41:16.957476 137579062097728 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0419 05:41:16.957503 137579062097728 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0419 05:41:16.957529 137579062097728 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0419 05:41:17.228837 137579062097728 metric_logger.py:185] completed step: 0, seconds: 11.013, TFLOP/s/device: 1.234, Tokens/s/device: 185.969, total_weights: 65536, loss: 10.874
I0419 05:41:17.403303 137579062097728 metric_logger.py:185] completed step: 1, seconds: 2.100, TFLOP/s/device: 6.471, Tokens/s/device: 975.333, total_weights: 65536, loss: 10.874
I0419 05:41:17.828419 137579062097728 metric_logger.py:185] completed step: 2, seconds: 0.029, TFLOP/s/device: 473.997, Tokens/s/device: 71446.014, total_weights: 65536, loss: 10.262
I0419 05:41:17.975805 137579062097728 metric_logger.py:185] completed step: 3, seconds: 0.416, TFLOP/s/device: 32.698, Tokens/s/device: 4928.586, total_weights: 65536, loss: 9.732
I0419 05:41:18.271598 137579062097728 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.291, Tokens/s/device: 12554.558, total_weights: 65536, loss: 9.274
I0419 05:41:18.278583 137579062097728 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.270, Tokens/s/device: 13907.941, total_weights: 65536, loss: 8.888
I0419 05:41:38.198118 137579062097728 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.897, Tokens/s/device: 6918.101, total_weights: 65536, loss: 8.589
I0419 05:41:38.345376 137579062097728 metric_logger.py:185] completed step: 7, seconds: 19.773, TFLOP/s/device: 0.687, Tokens/s/device: 103.576, total_weights: 65536, loss: 8.380
I0419 05:41:38.493070 137579062097728 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.637, Tokens/s/device: 13360.385, total_weights: 65536, loss: 8.251
I0419 05:41:38.641502 137579062097728 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0419 05:41:38.642063 137579062097728 checkpointing.py:798] Waited 0.0005764961242675781 seconds for step 9 to finish before starting checkpointing.
I0419 05:41:38.644007 137579062097728 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 05:41:40.604926    2666 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 05:41:43.380691 137456116553472 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_6
I0419 05:42:13.777441 137456108160768 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 27.339 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 57.77739357948303 s) (per-host)
I0419 05:42:13.777544 137456108160768 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 57.271355s.
I0419 05:42:22.816299 137456108160768 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 66.310090s.
I0419 05:42:22.816602 137451813193472 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 05:42:22.816725 137451813193472 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 05:42:22.816793 137451813193472 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 05:42:22.818775 137451813193472 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 05:42:22.818961 137579062097728 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0419 05:42:22.819097 137579062097728 checkpoint_manager.py:1458] Waiting for previous save to complete took 44.175093 seconds. If this number is high, consider checkpointing less frequently.
I0419 05:42:22.821060 137579062097728 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9
I0419 05:42:22.824638 137579062097728 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints/9.
I0419 05:42:24.243746 137579062097728 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0419 05:42:24.243901 137579062097728 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0419 05:42:24.287676 137579062097728 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046148s
I0419 05:42:24.287839 137579062097728 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.094 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.736635684967041 s) (per-host)
I0419 05:42:24.287888 137579062097728 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.736693s (batch_requests_ready=0.685836s, total_serialization_initiated=0.050792s, others=0.000065s)
I0419 05:42:24.287981 137579062097728 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.740805s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.740792)
I0419 05:42:24.288866 137453391578880 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0419 05:42:24.288995 137579062097728 async_checkpointer.py:561] Finished blocking save. Time taken: 1.467865s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints/9.
I0419 05:42:24.291934 137579062097728 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 05:42:24.292163 137451813193472 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 05:42:24.292310 137579062097728 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776577298.6439788, 'wait_for_prev_duration_secs': 44.17509317398071, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776577342.8210988, 'checkpointer_blocking_duration_secs': 1.4680805206298828, 'get_old_steps_start_time': 1776577344.2891984, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1776577298.6423042, 'checkpoint_manager_blocking_duration_secs': 45.64996099472046}
I0419 05:42:24.292424 137579062097728 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0419 05:42:24.292477 137579062097728 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 05:42:30.354058 137456116553472 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_6
I0419 05:43:05.517984 137453391578880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.639 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.9667284488678 s) (per-host)
I0419 05:43:05.518090 137453391578880 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.229014s.
I0419 05:43:15.350470 137453391578880 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.061378s.
I0419 05:43:15.350704 137451813193472 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 05:43:15.350762 137451813193472 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 05:43:15.350807 137451813193472 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0419 05:43:15.352466 137451813193472 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0419 05:43:15.352642 137579062097728 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0419 05:43:15.352809 137579062097728 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 05:43:15.353462 137579062097728 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.404, Tokens/s/device: 13928.088, total_weights: 65536, loss: 8.175
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 05:43:27 UTC 2026
EXIT_CODE=0