MaxView

← Back to run

Log Summary

XPK Start: Wed Apr 22 18:46:29 UTC 2026
2026-04-22 18:46:33.594925: 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 00:00:1776883593.607914      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776883593.611660      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776883593.623231      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776883593.623252      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776883593.623255      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776883593.623257      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-22 18:46:52.838829: 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)
I0422 18:46:53.369888 133996793280320 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-22 18:47:02,411:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0422 18:47:02.411328 133996793280320 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-22 18:47:02,413:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-sv-o4en0-slice-job-0-0.mt-16-zarr3-true-sv-o4en0:8482
I0422 18:47:02.413726 133996793280320 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-sv-o4en0-slice-job-0-0.mt-16-zarr3-true-sv-o4en0:8482
I0422 18:47:03.634669 133996793280320 max_utils.py:284] Jax distributed system initialized!
I0422 18:47:10.521535 133996793280320 max_utils.py:800] System Information: Jax Version: 0.8.1
I0422 18:47:10.521642 133996793280320 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0422 18:47:10.521682 133996793280320 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
I0422 18:47:10.521716 133996793280320 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 18:47:11.144275 133996793280320 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 18:47:11.257449 133996793280320 checkpointing.py:688] Setting up checkpoint logger...
I0422 18:47:11.257569 133996793280320 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0422 18:47:11.257616 133996793280320 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0422 18:47:11.257824 133996793280320 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 0x79de1a7790d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 18:47:14.307684 133996793280320 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0422 18:47:14.307927 133996793280320 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 0x79c924698530>}, handler_registry=None
I0422 18:47:14.308180 133996793280320 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79c924698530>` 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`.
I0422 18:47:14.308229 133996793280320 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79c93c534170>` 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`.
I0422 18:47:14.308266 133996793280320 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 0x79c924698530>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79c924698530>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79c93c534170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79c93c534170>}).
I0422 18:47:14.308608 133996793280320 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0422 18:47:14.308686 133996793280320 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79c93c3b4f40> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0422 18:47:16.254371 133996793280320 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints
I0422 18:47:16.256515 133996793280320 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=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79c93c5320f0>
I0422 18:47:16.256632 133996793280320 checkpointing.py:302] Checkpoint manager created!
I0422 18:47:17.194854 133996793280320 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 18:47:17.194972 133996793280320 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0422 18:47:18.635979 133996793280320 maxtext_utils.py:1749]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0422 18:47:18.636115 133996793280320 maxtext_utils.py:1749]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 18:47:18.636172 133996793280320 maxtext_utils.py:1749]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 18:47:18.636218 133996793280320 maxtext_utils.py:1749]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0422 18:47:18.636256 133996793280320 maxtext_utils.py:1749]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 18:47:18.636291 133996793280320 maxtext_utils.py:1749]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 18:47:18.636332 133996793280320 maxtext_utils.py:1749]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 18:47:18.636370 133996793280320 maxtext_utils.py:1749]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0422 18:47:18.636406 133996793280320 maxtext_utils.py:1749]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 18:47:18.636440 133996793280320 maxtext_utils.py:1749]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 18:47:18.636472 133996793280320 maxtext_utils.py:1749]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)

I0422 18:47:18.636505 133996793280320 maxtext_utils.py:1749]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0422 18:47:19.030393 133996793280320 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 18:47:19.030542 133996793280320 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 18:47:19.035954 133996793280320 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 18:47:19.036050 133996793280320 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 18:47:19.195391 133996793280320 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 18:47:19.195526 133996793280320 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 18:47:19.349867 133996793280320 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 18:47:19.349997 133996793280320 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 18:47:19.515677 133996793280320 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 18:47:19.515815 133996793280320 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 18:47:19.669263 133996793280320 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 18:47:19.669404 133996793280320 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 18:47:19.821249 133996793280320 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 18:47:19.821390 133996793280320 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 18:47:19.981483 133996793280320 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 18:47:19.981621 133996793280320 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 18:47:20.006047 133996793280320 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 18:47:20.006177 133996793280320 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 18:47:31.219858 133996793280320 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.
I0422 18:47:31.222479 133996793280320 metric_logger.py:289] number parameters: 1.104 billion
I0422 18:47:42.088198 133996793280320 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0422 18:47:42.240247 133996793280320 checkpointing.py:798] Waited 0.15203499794006348 seconds for step 0 to finish before starting checkpointing.
I0422 18:47:42.242624 133996793280320 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 18:47:42.244593 133996793280320 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0
I0422 18:47:42.247368 133996793280320 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/0.
I0422 18:47:42.892706 133996793280320 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 18:47:42.893748 133996793280320 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0422 18:47:42.893807 133996793280320 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
I0422 18:47:43.148142 133996793280320 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256580s
I0422 18:47:43.148317 133996793280320 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.949 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5231194496154785 s) (per-host)
I0422 18:47:43.148370 133996793280320 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.523183s (batch_requests_ready=0.259970s, total_serialization_initiated=0.263142s, others=0.000071s)
I0422 18:47:43.148465 133996793280320 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.527179s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.527159)
I0422 18:47:43.149457 133869953079040 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0422 18:47:43.149588 133996793280320 async_checkpointer.py:561] Finished blocking save. Time taken: 0.904923s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/0.
I0422 18:47:43.472391 133996793280320 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 18:47:43.472769 133868845766400 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 18:47:43.472938 133996793280320 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883662.2426057, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883662.2446318, 'checkpointer_blocking_duration_secs': 0.9051356315612793, 'get_old_steps_start_time': 1776883663.149786, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776883662.240654, 'checkpoint_manager_blocking_duration_secs': 1.2322440147399902}
I0422 18:47:43.473047 133996793280320 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0422 18:47:43.473133 133996793280320 max_utils.py:750] 
Memstats: After params initialized:
I0422 18:47:43.473196 133996793280320 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0422 18:47:43.473230 133996793280320 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0422 18:47:43.473257 133996793280320 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0422 18:47:43.473283 133996793280320 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0422 18:47:43.741055 133996793280320 metric_logger.py:185] completed step: 0, seconds: 10.865, TFLOP/s/device: 1.251, Tokens/s/device: 188.501, total_weights: 65536, loss: 10.874
I0422 18:47:43.906013 133996793280320 metric_logger.py:185] completed step: 1, seconds: 1.651, TFLOP/s/device: 8.229, Tokens/s/device: 1240.393, total_weights: 65536, loss: 10.874
I0422 18:47:44.364608 133996793280320 metric_logger.py:185] completed step: 2, seconds: 0.019, TFLOP/s/device: 705.825, Tokens/s/device: 106389.610, total_weights: 65536, loss: 10.560
I0422 18:47:44.511980 133996793280320 metric_logger.py:185] completed step: 3, seconds: 0.450, TFLOP/s/device: 30.207, Tokens/s/device: 4553.064, total_weights: 65536, loss: 9.979
I0422 18:47:44.517802 133996793280320 checkpointing.py:794] Waiting for step 5 to finish before checkpoint...
I0422 18:47:44.806945 133996793280320 checkpointing.py:798] Waited 0.28911447525024414 seconds for step 5 to finish before starting checkpointing.
I0422 18:47:44.810154 133996793280320 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 18:47:45.870714    2610 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 18:47:47.725112 133869961471744 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/0/items/array_metadatas/process_6
I0422 18:48:18.849815 133869953079040 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 43.606 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.224555015563965 s) (per-host)
I0422 18:48:18.849939 133869953079040 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.700273s.
I0422 18:48:27.608994 133869953079040 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.459311s.
I0422 18:48:27.609332 133868845766400 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 18:48:27.609459 133868845766400 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 18:48:27.609509 133868845766400 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0422 18:48:27.610949 133868845766400 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0422 18:48:27.611148 133996793280320 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.
W0422 18:48:27.611284 133996793280320 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.801126 seconds. If this number is high, consider checkpointing less frequently.
I0422 18:48:27.613380 133996793280320 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5
I0422 18:48:27.617661 133996793280320 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/5.
I0422 18:48:28.483219 133996793280320 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0422 18:48:28.483365 133996793280320 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
I0422 18:48:28.532361 133996793280320 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051364s
I0422 18:48:28.532528 133996793280320 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.995 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3088245391845703 s) (per-host)
I0422 18:48:28.532581 133996793280320 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.308886s (batch_requests_ready=0.252772s, total_serialization_initiated=0.056046s, others=0.000068s)
I0422 18:48:28.532678 133996793280320 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.313879s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.313862)
I0422 18:48:28.533597 133869961471744 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0422 18:48:28.533757 133996793280320 async_checkpointer.py:561] Finished blocking save. Time taken: 0.920304s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/5.
I0422 18:48:28.954775 133996793280320 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 18:48:28.955169 133868845766400 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 18:48:28.955321 133996793280320 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883664.8101258, 'wait_for_prev_duration_secs': 42.80112648010254, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883707.6134193, 'checkpointer_blocking_duration_secs': 0.9204452037811279, 'get_old_steps_start_time': 1776883708.5338838, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776883664.8073058, 'checkpoint_manager_blocking_duration_secs': 44.14798021316528}
I0422 18:48:28.955492 133996793280320 checkpointing.py:409] Started an asynchronous checkpoint save for step 5
I0422 18:48:28.956326 133996793280320 metric_logger.py:185] completed step: 4, seconds: 0.162, TFLOP/s/device: 83.652, Tokens/s/device: 12608.974, total_weights: 65536, loss: 9.457
I0422 18:48:28.963759 133996793280320 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.444, Tokens/s/device: 13934.153, total_weights: 65536, loss: 8.955
I0422 18:48:49.019926 133996793280320 metric_logger.py:185] completed step: 6, seconds: 44.444, TFLOP/s/device: 0.306, Tokens/s/device: 46.080, total_weights: 65536, loss: 8.465
I0422 18:48:49.167361 133996793280320 metric_logger.py:185] completed step: 7, seconds: 19.911, TFLOP/s/device: 0.682, Tokens/s/device: 102.859, total_weights: 65536, loss: 7.998
I0422 18:48:49.314702 133996793280320 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.493, Tokens/s/device: 13338.631, total_weights: 65536, loss: 7.567
I0422 18:48:49.463155 133996793280320 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0422 18:48:49.463791 133996793280320 checkpointing.py:798] Waited 0.00064849853515625 seconds for step 9 to finish before starting checkpointing.
I0422 18:48:49.466189 133996793280320 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 18:48:53.381560 133865674897152 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/5/items/array_metadatas/process_6
I0422 18:49:28.578637 133869961471744 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.172 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.354888916015625 s) (per-host)
I0422 18:49:28.578765 133869961471744 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.045056s.
I0422 18:49:37.966652 133869961471744 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 69.432928s.
I0422 18:49:37.966893 133868845766400 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 18:49:37.966961 133868845766400 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 18:49:37.967007 133868845766400 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0422 18:49:37.968585 133868845766400 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0422 18:49:37.968750 133996793280320 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5.
W0422 18:49:37.968891 133996793280320 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.502721 seconds. If this number is high, consider checkpointing less frequently.
I0422 18:49:37.970401 133996793280320 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9
I0422 18:49:37.973976 133996793280320 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/9.
I0422 18:49:38.518313 133996793280320 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0422 18:49:38.518417 133996793280320 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
I0422 18:49:38.572836 133996793280320 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.056465s
I0422 18:49:38.572996 133996793280320 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.862 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31724095344543457 s) (per-host)
I0422 18:49:38.573044 133996793280320 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.317299s (batch_requests_ready=0.256343s, total_serialization_initiated=0.060894s, others=0.000063s)
I0422 18:49:38.573157 133996793280320 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.321228s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.321216)
I0422 18:49:38.573874 133869919508224 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0422 18:49:38.574028 133996793280320 async_checkpointer.py:561] Finished blocking save. Time taken: 0.603564s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/9.
I0422 18:49:38.586184 133996793280320 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 18:49:38.586414 133868845766400 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 18:49:38.586541 133996793280320 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883729.4661531, 'wait_for_prev_duration_secs': 48.50272059440613, 'time_between_consecutive_saves_sec': 21.855164766311646, 'checkpointer_blocking_start_time': 1776883777.9704397, 'checkpointer_blocking_duration_secs': 0.6036779880523682, 'get_old_steps_start_time': 1776883778.5741315, 'get_old_steps_duration_secs': 1.71661376953125e-05, 'checkpoint_manager_blocking_start_time': 1776883729.46401, 'checkpoint_manager_blocking_duration_secs': 49.122498512268066}
I0422 18:49:38.586660 133996793280320 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0422 18:49:38.586711 133996793280320 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 18:49:43.767752 133869953079040 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/9/items/array_metadatas/process_6
I0422 18:50:20.732581 133869919508224 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.187 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.476784467697144 s) (per-host)
I0422 18:50:20.732719 133869919508224 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.158731s.
I0422 18:50:27.918088 133869919508224 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.344085s.
I0422 18:50:27.918347 133868845766400 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 18:50:27.918416 133868845766400 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 18:50:27.918454 133868845766400 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0422 18:50:27.919865 133868845766400 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0422 18:50:27.920023 133996793280320 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.
I0422 18:50:27.920202 133996793280320 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 18:50:27.920971 133996793280320 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.498, Tokens/s/device: 13942.311, total_weights: 65536, loss: 7.176
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Wed Apr 22 18:50:40 UTC 2026
EXIT_CODE=0