MaxView

XPK Start: Mon Apr 20 22:26:41 UTC 2026
2026-04-20 22:26:45.530973: 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 00:00:1776724005.543419      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776724005.547061      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776724005.558329      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776724005.558347      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776724005.558349      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776724005.558351      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-20 22:27:04.646835: 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 22:27:05.167347 138383172503360 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-20 22:27:14,209:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0420 22:27:14.209530 138383172503360 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-20 22:27:14,211:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-rs-j3q53-slice-job-0-0.mt-16-zarr3-false-rs-j3q53:8482
I0420 22:27:14.211925 138383172503360 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-rs-j3q53-slice-job-0-0.mt-16-zarr3-false-rs-j3q53:8482
I0420 22:27:16.076707 138383172503360 max_utils.py:284] Jax distributed system initialized!
I0420 22:27:22.919820 138383172503360 max_utils.py:800] System Information: Jax Version: 0.8.1
I0420 22:27:22.919952 138383172503360 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0420 22:27:22.919992 138383172503360 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
I0420 22:27:22.920026 138383172503360 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 22:27:23.544181 138383172503360 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 22:27:23.660619 138383172503360 checkpointing.py:688] Setting up checkpoint logger...
I0420 22:27:23.660734 138383172503360 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False
I0420 22:27:23.660779 138383172503360 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0420 22:27:23.661026 138383172503360 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7ddb58775220>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 22:27:24.916099 138383172503360 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 22:27:24.916439 138383172503360 checkpoint_manager.py:709] [process=1][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dc8540f44a0>}, handler_registry=None
I0420 22:27:24.916701 138383172503360 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dc8540f44a0>` 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 22:27:24.916766 138383172503360 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dc8541177a0>` 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 22:27:24.916819 138383172503360 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 0x7dc8540f44a0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dc8540f44a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dc8541177a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dc8541177a0>}).
I0420 22:27:24.917201 138383172503360 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0420 22:27:24.917282 138383172503360 async_checkpointer.py:177] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dc85465cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0420 22:27:27.660755 138263041926912 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776723830190133517, 'commit_timestamp_nsecs': 1776723877828993600, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA
I0420 22:27:27.904848 138263041926912 checkpoint_manager.py:1877] Missing metrics for step 5
I0420 22:27:28.358947 138383172503360 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints
I0420 22:27:29.205588 138383172503360 checkpoint_manager.py:929] [process=1][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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ddb44335fa0>
I0420 22:27:29.205760 138383172503360 checkpointing.py:302] Checkpoint manager created!
I0420 22:27:30.132277 138383172503360 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 22:27:30.132394 138383172503360 checkpointing.py:582] restoring from this run's directory step 9
I0420 22:27:30.622797 138383172503360 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9.
I0420 22:27:32.059441    1410 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 22:27:39.202618 138383172503360 jax_array_handlers.py:837] [process=1] /jax/orbax/read/worker/io/requested throughput: 221.059 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.148662805557251 s) (per-host)
I0420 22:27:39.210815 138383172503360 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/read/gbytes_per_sec: 208.553 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.588485956192017 s) (per-host)
I0420 22:27:39.822279 138383172503360 checkpointer.py:318] Finished restoring checkpoint in 9.44 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9.
I0420 22:27:39.822474 138383172503360 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776724050.3790917, 'checkpointer_duration_secs': 9.443343162536621, 'checkpoint_manager_start_time': 1776724050.1330338, 'checkpoint_manager_duration_secs': 9.689403772354126}
fsdp: 32
I0420 22:27:39.903416 138383172503360 maxtext_utils.py:1749]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0420 22:27:39.903556 138383172503360 maxtext_utils.py:1749]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 22:27:39.903633 138383172503360 maxtext_utils.py:1749]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 22:27:39.903708 138383172503360 maxtext_utils.py:1749]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0420 22:27:39.903773 138383172503360 maxtext_utils.py:1749]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 22:27:39.903830 138383172503360 maxtext_utils.py:1749]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 22:27:39.903899 138383172503360 maxtext_utils.py:1749]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 22:27:39.903961 138383172503360 maxtext_utils.py:1749]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0420 22:27:39.904019 138383172503360 maxtext_utils.py:1749]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 22:27:39.904076 138383172503360 maxtext_utils.py:1749]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 22:27:39.904132 138383172503360 maxtext_utils.py:1749]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0420 22:27:39.904188 138383172503360 maxtext_utils.py:1749]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0420 22:27:40.299632 138383172503360 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 22:27:40.299793 138383172503360 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:27:40.305248 138383172503360 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 22:27:40.305341 138383172503360 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:27:40.458416 138383172503360 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 22:27:40.458572 138383172503360 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:27:40.612851 138383172503360 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 22:27:40.612983 138383172503360 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:27:40.774621 138383172503360 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 22:27:40.774757 138383172503360 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:27:40.929905 138383172503360 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 22:27:40.930033 138383172503360 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:27:41.085461 138383172503360 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 22:27:41.085587 138383172503360 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:27:41.251058 138383172503360 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 22:27:41.251190 138383172503360 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:27:41.279298 138383172503360 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 22:27:41.279444 138383172503360 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 22:27:52.467095 138383172503360 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 22:27:52.469606 138383172503360 metric_logger.py:289] number parameters: 1.104 billion
I0420 22:28:03.409765 138383172503360 checkpointing.py:794] Waiting for step 10 to finish before checkpoint...
I0420 22:28:03.560215 138383172503360 checkpointing.py:798] Waited 0.15043330192565918 seconds for step 10 to finish before starting checkpointing.
I0420 22:28:03.562679 138383172503360 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 22:28:03.564595 138383172503360 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 10
I0420 22:28:03.567320 138383172503360 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/10.
I0420 22:28:04.573901 138383172503360 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 22:28:04.574889 138383172503360 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 22:28:04.574949 138383172503360 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 22:28:04.828113 138383172503360 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.255332s
I0420 22:28:04.828281 138383172503360 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.043 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.507007360458374 s) (per-host)
I0420 22:28:04.828333 138383172503360 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.507069s (batch_requests_ready=0.247014s, total_serialization_initiated=0.259985s, others=0.000070s)
I0420 22:28:04.828445 138383172503360 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.511113s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.511095)
I0420 22:28:04.829309 138262999963392 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0420 22:28:04.829491 138383172503360 async_checkpointer.py:561] Finished blocking save. Time taken: 1.264828s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/10.
I0420 22:28:05.262125 138383172503360 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 22:28:05.262521 138258210088704 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 22:28:05.262660 138383172503360 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776724083.562661, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776724083.564635, 'checkpointer_blocking_duration_secs': 1.2649633884429932, 'get_old_steps_start_time': 1776724084.8296165, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776724083.5606077, 'checkpoint_manager_blocking_duration_secs': 1.7020156383514404}
I0420 22:28:05.262815 138383172503360 checkpointing.py:409] Started an asynchronous checkpoint save for step 10
I0420 22:28:05.262919 138383172503360 max_utils.py:750] 
Memstats: After params initialized:
I0420 22:28:05.262977 138383172503360 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_2(process=1,(2,0,0,0))
I0420 22:28:05.263009 138383172503360 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_3(process=1,(3,0,0,0))
I0420 22:28:05.263035 138383172503360 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_6(process=1,(2,1,0,0))
I0420 22:28:05.263058 138383172503360 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_7(process=1,(3,1,0,0))
I0420 22:28:05.534885 138383172503360 metric_logger.py:185] completed step: 10, seconds: 10.939, TFLOP/s/device: 1.242, Tokens/s/device: 187.218, total_weights: 65536, loss: 6.827
I0420 22:28:05.703223 138383172503360 metric_logger.py:185] completed step: 11, seconds: 2.123, TFLOP/s/device: 6.399, Tokens/s/device: 964.496, total_weights: 65536, loss: 6.522
I0420 22:28:05.850720 138383172503360 metric_logger.py:185] completed step: 12, seconds: 0.013, TFLOP/s/device: 1070.696, Tokens/s/device: 161386.919, total_weights: 65536, loss: 6.260
I0420 22:28:05.998221 138383172503360 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 82.954, Tokens/s/device: 12503.740, total_weights: 65536, loss: 6.039
I0420 22:28:06.005774 138383172503360 checkpointing.py:794] Waiting for step 15 to finish before checkpoint...
I0420 22:28:06.293287 138383172503360 checkpointing.py:798] Waited 0.28748202323913574 seconds for step 15 to finish before starting checkpointing.
I0420 22:28:06.295868 138383172503360 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 22:28:14.461123 138258738566912 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/10/items/array_metadatas/process_1
I0420 22:28:50.339484 138262999963392 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 34.325 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.018176794052124 s) (per-host)
I0420 22:28:50.339604 138262999963392 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.510170s.
I0420 22:28:59.577516 138262999963392 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 54.748066s.
I0420 22:28:59.577804 138258210088704 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 22:28:59.577936 138258210088704 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 22:28:59.577986 138258210088704 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0420 22:28:59.579536 138258210088704 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0420 22:28:59.579732 138383172503360 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0420 22:28:59.579885 138383172503360 checkpoint_manager.py:1458] Waiting for previous save to complete took 53.284048 seconds. If this number is high, consider checkpointing less frequently.
I0420 22:28:59.582053 138383172503360 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 15
I0420 22:28:59.586180 138383172503360 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/15.
I0420 22:29:00.545206 138383172503360 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 22:29:00.545310 138383172503360 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 22:29:00.591288 138383172503360 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048147s
I0420 22:29:00.591467 138383172503360 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.002 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3083615303039551 s) (per-host)
I0420 22:29:00.591520 138383172503360 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.308423s (batch_requests_ready=0.255648s, total_serialization_initiated=0.052706s, others=0.000070s)
I0420 22:29:00.591612 138383172503360 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.313507s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.313494)
I0420 22:29:00.592485 138258738566912 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0420 22:29:00.592645 138383172503360 async_checkpointer.py:561] Finished blocking save. Time taken: 1.010510s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/15.
I0420 22:29:01.040846 138383172503360 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 22:29:01.041196 138258210088704 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 22:29:01.041369 138383172503360 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776724086.2958145, 'wait_for_prev_duration_secs': 53.284048080444336, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776724139.5821025, 'checkpointer_blocking_duration_secs': 1.0106513500213623, 'get_old_steps_start_time': 1776724140.5927737, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776724086.293674, 'checkpoint_manager_blocking_duration_secs': 54.74765586853027}
I0420 22:29:01.041494 138383172503360 checkpointing.py:409] Started an asynchronous checkpoint save for step 15
I0420 22:29:01.042291 138383172503360 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.437, Tokens/s/device: 13933.110, total_weights: 65536, loss: 5.858
I0420 22:29:01.049848 138383172503360 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.275, Tokens/s/device: 13757.893, total_weights: 65536, loss: 5.713
I0420 22:29:19.413460 138383172503360 metric_logger.py:185] completed step: 16, seconds: 55.043, TFLOP/s/device: 0.247, Tokens/s/device: 37.208, total_weights: 65536, loss: 5.599
I0420 22:29:19.561096 138383172503360 metric_logger.py:185] completed step: 17, seconds: 18.218, TFLOP/s/device: 0.746, Tokens/s/device: 112.414, total_weights: 65536, loss: 5.511
I0420 22:29:19.708606 138383172503360 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.540, Tokens/s/device: 13345.671, total_weights: 65536, loss: 5.442
I0420 22:29:19.857093 138383172503360 checkpointing.py:794] Waiting for step 19 to finish before checkpoint...
I0420 22:29:19.857732 138383172503360 checkpointing.py:798] Waited 0.0006520748138427734 seconds for step 19 to finish before starting checkpointing.
I0420 22:29:19.859995 138383172503360 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 22:29:23.252729 138258746959616 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/15/items/array_metadatas/process_1
I0420 22:30:00.582680 138258738566912 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 26.196 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.29954671859741 s) (per-host)
I0420 22:30:00.582782 138258738566912 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.990187s.
I0420 22:30:07.603062 138258738566912 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 67.010451s.
I0420 22:30:07.603363 138258210088704 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 22:30:07.603443 138258210088704 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 22:30:07.603499 138258210088704 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0420 22:30:07.604888 138258210088704 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0420 22:30:07.605062 138383172503360 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0420 22:30:07.605185 138383172503360 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.745200 seconds. If this number is high, consider checkpointing less frequently.
I0420 22:30:07.606856 138383172503360 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 19
I0420 22:30:07.610270 138383172503360 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/19.
I0420 22:30:08.998944 138383172503360 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 22:30:08.999044 138383172503360 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 22:30:09.051074 138383172503360 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.054371s
I0420 22:30:09.051239 138383172503360 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.081 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.741255521774292 s) (per-host)
I0420 22:30:09.051285 138383172503360 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.741311s (batch_requests_ready=0.682338s, total_serialization_initiated=0.058913s, others=0.000060s)
I0420 22:30:09.051377 138383172503360 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.745370s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.745358)
I0420 22:30:09.053208 138258210088704 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0420 22:30:09.053312 138383172503360 async_checkpointer.py:561] Finished blocking save. Time taken: 1.446337s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/19.
I0420 22:30:09.055188 138383172503360 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 22:30:09.055386 138261959792384 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 22:30:09.055498 138383172503360 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776724159.8599677, 'wait_for_prev_duration_secs': 47.745200395584106, 'time_between_consecutive_saves_sec': 20.2803852558136, 'checkpointer_blocking_start_time': 1776724207.6069438, 'checkpointer_blocking_duration_secs': 1.446488618850708, 'get_old_steps_start_time': 1776724209.053448, 'get_old_steps_duration_secs': 2.0742416381835938e-05, 'checkpoint_manager_blocking_start_time': 1776724159.8579557, 'checkpoint_manager_blocking_duration_secs': 49.19751238822937}
I0420 22:30:09.055657 138383172503360 checkpointing.py:409] Started an asynchronous checkpoint save for step 19
I0420 22:30:09.055698 138383172503360 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 22:30:14.641865 138262999963392 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/19/items/array_metadatas/process_1
I0420 22:30:48.854093 138258210088704 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 38.960 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.5440776348114 s) (per-host)
I0420 22:30:48.854209 138258210088704 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.800941s.
I0420 22:31:01.284707 138258210088704 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 52.231425s.
I0420 22:31:01.284944 138261959792384 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 22:31:01.285005 138261959792384 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 22:31:01.285048 138261959792384 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0420 22:31:01.287842 138261959792384 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0420 22:31:01.288023 138383172503360 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0420 22:31:01.288225 138383172503360 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 22:31:01.288945 138383172503360 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.273, Tokens/s/device: 13908.414, total_weights: 65536, loss: 5.385
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Mon Apr 20 22:31:12 UTC 2026
EXIT_CODE=0