MaxView

XPK Start: Mon Apr 20 22:15:26 UTC 2026
2026-04-20 22:15:30.893822: 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 00:00:1776723330.906749      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776723330.910415      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776723330.921780      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776723330.921800      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776723330.921802      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776723330.921804      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-20 22:15:50.067692: 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 22:15:50.591137 134264641779520 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-20 22:15:59,632:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0420 22:15:59.632708 134264641779520 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-20 22:15:59,634:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-koeqk-slice-job-0-0.mt-16-zarr3-true-rs-koeqk:8482
I0420 22:15:59.634987 134264641779520 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-koeqk-slice-job-0-0.mt-16-zarr3-true-rs-koeqk:8482
I0420 22:16:01.121056 134264641779520 max_utils.py:284] Jax distributed system initialized!
I0420 22:16:07.993635 134264641779520 max_utils.py:800] System Information: Jax Version: 0.8.1
I0420 22:16:07.993737 134264641779520 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0420 22:16:07.993776 134264641779520 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 22:16:07.993810 134264641779520 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 22:16:08.606919 134264641779520 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 22:16:08.720477 134264641779520 checkpointing.py:688] Setting up checkpoint logger...
I0420 22:16:08.720594 134264641779520 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 22:16:08.720640 134264641779520 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0420 22:16:08.720848 134264641779520 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 0x7a1c797791f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 22:16:10.464483 134264641779520 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 22:16:10.464804 134264641779520 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 0x7a0a109a42c0>}, handler_registry=None
I0420 22:16:10.465047 134264641779520 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a0a109a42c0>` 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 22:16:10.465104 134264641779520 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0a109ab830>` 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 22:16:10.465141 134264641779520 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 0x7a0a109a42c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a0a109a42c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0a109ab830>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0a109ab830>}).
I0420 22:16:10.465492 134264641779520 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0420 22:16:10.465569 134264641779520 async_checkpointer.py:177] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a0a11d8cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0420 22:16:14.052926 134147263690496 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776723200544764582, 'commit_timestamp_nsecs': 1776723249450322536, '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_true/checkpoints/9/_CHECKPOINT_METADATA
I0420 22:16:14.333699 134147683096320 checkpoint_manager.py:1877] Missing metrics for step 5
I0420 22:16:14.742629 134264641779520 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_true/checkpoints
I0420 22:16:16.449004 134264641779520 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_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a0a11dcbbf0>
I0420 22:16:16.449181 134264641779520 checkpointing.py:302] Checkpoint manager created!
I0420 22:16:17.377915 134264641779520 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 22:16:17.378033 134264641779520 checkpointing.py:582] restoring from this run's directory step 9
I0420 22:16:17.879610 134264641779520 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_true/checkpoints/9.
I0420 22:16:19.306217    1378 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 22:16:27.218930 134264641779520 jax_array_handlers.py:837] [process=1] /jax/orbax/read/worker/io/requested throughput: 199.562 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.918715953826904 s) (per-host)
I0420 22:16:27.227642 134264641779520 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/read/gbytes_per_sec: 189.781 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 8.339121341705322 s) (per-host)
I0420 22:16:27.966455 134264641779520 checkpointer.py:318] Finished restoring checkpoint in 10.34 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_true/checkpoints/9.
I0420 22:16:27.966607 134264641779520 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_true/checkpoints', 'checkpointer_start_time': 1776723377.6269715, 'checkpointer_duration_secs': 10.33960485458374, 'checkpoint_manager_start_time': 1776723377.37864, 'checkpoint_manager_duration_secs': 10.5879385471344}
fsdp: 32
I0420 22:16:28.047651 134264641779520 maxtext_utils.py:1749]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0420 22:16:28.047768 134264641779520 maxtext_utils.py:1749]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 22:16:28.047822 134264641779520 maxtext_utils.py:1749]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 22:16:28.047873 134264641779520 maxtext_utils.py:1749]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0420 22:16:28.047913 134264641779520 maxtext_utils.py:1749]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 22:16:28.047949 134264641779520 maxtext_utils.py:1749]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 22:16:28.047988 134264641779520 maxtext_utils.py:1749]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 22:16:28.048028 134264641779520 maxtext_utils.py:1749]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0420 22:16:28.048066 134264641779520 maxtext_utils.py:1749]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 22:16:28.048102 134264641779520 maxtext_utils.py:1749]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 22:16:28.048137 134264641779520 maxtext_utils.py:1749]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]

    Physical:  ('fsdp', None)
I0420 22:16:28.048171 134264641779520 maxtext_utils.py:1749]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0420 22:16:28.448393 134264641779520 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 22:16:28.448557 134264641779520 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:16:28.453939 134264641779520 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 22:16:28.454030 134264641779520 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:16:28.598686 134264641779520 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 22:16:28.598814 134264641779520 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:16:28.742582 134264641779520 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 22:16:28.742709 134264641779520 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 22:16:28.896217 134264641779520 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 22:16:28.896350 134264641779520 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:16:29.040520 134264641779520 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 22:16:29.040647 134264641779520 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:16:29.183767 134264641779520 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 22:16:29.183893 134264641779520 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:16:29.336995 134264641779520 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 22:16:29.337129 134264641779520 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 22:16:29.360222 134264641779520 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 22:16:29.360313 134264641779520 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 22:16:40.421198 134264641779520 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 22:16:40.423670 134264641779520 metric_logger.py:289] number parameters: 1.104 billion
I0420 22:16:51.129632 134264641779520 checkpointing.py:794] Waiting for step 10 to finish before checkpoint...
I0420 22:16:51.446277 134264641779520 checkpointing.py:798] Waited 0.31663036346435547 seconds for step 10 to finish before starting checkpointing.
I0420 22:16:51.449433 134264641779520 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 22:16:51.451216 134264641779520 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 10
I0420 22:16:51.453922 134264641779520 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_true/checkpoints/10.
I0420 22:16:52.475589 134264641779520 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 22:16:52.476789 134264641779520 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 22:16:52.476848 134264641779520 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 22:16:52.720748 134264641779520 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.246301s
I0420 22:16:52.720926 134264641779520 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.047 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5063130855560303 s) (per-host)
I0420 22:16:52.720977 134264641779520 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.506374s (batch_requests_ready=0.255211s, total_serialization_initiated=0.251095s, others=0.000069s)
I0420 22:16:52.721069 134264641779520 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.510400s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.510382)
I0420 22:16:52.721965 134146038945536 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0420 22:16:52.722134 134264641779520 async_checkpointer.py:561] Finished blocking save. Time taken: 1.270848s. 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_true/checkpoints/10.
I0420 22:16:53.576220 134264641779520 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 22:16:53.576618 134143321020160 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 22:16:53.576783 134264641779520 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_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776723411.4494157, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776723411.4512541, 'checkpointer_blocking_duration_secs': 1.270986557006836, 'get_old_steps_start_time': 1776723412.7222593, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776723411.4466467, 'checkpoint_manager_blocking_duration_secs': 2.130098342895508}
I0420 22:16:53.576902 134264641779520 checkpointing.py:409] Started an asynchronous checkpoint save for step 10
I0420 22:16:53.576976 134264641779520 max_utils.py:750] 
Memstats: After params initialized:
I0420 22:16:53.577033 134264641779520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_2(process=1,(2,0,0,0))
I0420 22:16:53.577065 134264641779520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_3(process=1,(3,0,0,0))
I0420 22:16:53.577091 134264641779520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_6(process=1,(2,1,0,0))
I0420 22:16:53.577113 134264641779520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_7(process=1,(3,1,0,0))
I0420 22:16:53.848085 134264641779520 metric_logger.py:185] completed step: 10, seconds: 10.705, TFLOP/s/device: 1.269, Tokens/s/device: 191.314, total_weights: 65536, loss: 6.827
I0420 22:16:54.016776 134264641779520 metric_logger.py:185] completed step: 11, seconds: 2.717, TFLOP/s/device: 5.001, Tokens/s/device: 753.874, total_weights: 65536, loss: 6.522
I0420 22:16:54.164325 134264641779520 metric_logger.py:185] completed step: 12, seconds: 0.013, TFLOP/s/device: 1008.321, Tokens/s/device: 151985.158, total_weights: 65536, loss: 6.260
I0420 22:16:54.311986 134264641779520 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.267, Tokens/s/device: 12550.865, total_weights: 65536, loss: 6.039
I0420 22:16:54.319497 134264641779520 checkpointing.py:794] Waiting for step 15 to finish before checkpoint...
I0420 22:16:54.607132 134264641779520 checkpointing.py:798] Waited 0.28760218620300293 seconds for step 15 to finish before starting checkpointing.
I0420 22:16:54.609698 134264641779520 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 22:17:01.625611 134143337805568 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_true/checkpoints/10/items/array_metadatas/process_1
I0420 22:17:34.485611 134146038945536 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 37.368 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.270960330963135 s) (per-host)
I0420 22:17:34.485726 134146038945536 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.763654s.
I0420 22:17:47.849545 134146038945536 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 55.127456s.
I0420 22:17:47.849808 134143321020160 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 22:17:47.849921 134143321020160 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 22:17:47.849968 134143321020160 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0420 22:17:47.851939 134143321020160 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0420 22:17:47.852118 134264641779520 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 22:17:47.852245 134264641779520 checkpoint_manager.py:1458] Waiting for previous save to complete took 53.242549 seconds. If this number is high, consider checkpointing less frequently.
I0420 22:17:47.854153 134264641779520 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 15
I0420 22:17:47.857912 134264641779520 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_true/checkpoints/15.
I0420 22:17:49.704633 134264641779520 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 22:17:49.704731 134264641779520 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 22:17:49.745613 134264641779520 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043067s
I0420 22:17:49.745780 134264641779520 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.051 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7520272731781006 s) (per-host)
I0420 22:17:49.745839 134264641779520 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.752084s (batch_requests_ready=0.704307s, total_serialization_initiated=0.047712s, others=0.000066s)
I0420 22:17:49.745940 134264641779520 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.756782s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.756768)
I0420 22:17:49.746733 134141198714624 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0420 22:17:49.746889 134264641779520 async_checkpointer.py:561] Finished blocking save. Time taken: 1.892666s. 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_true/checkpoints/15.
I0420 22:17:49.748809 134264641779520 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 22:17:49.749050 134143321020160 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 22:17:49.749173 134264641779520 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_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776723414.6096716, 'wait_for_prev_duration_secs': 53.2425491809845, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776723467.8541925, 'checkpointer_blocking_duration_secs': 1.8928093910217285, 'get_old_steps_start_time': 1776723469.7470212, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776723414.6074998, 'checkpoint_manager_blocking_duration_secs': 55.141639709472656}
I0420 22:17:49.749283 134264641779520 checkpointing.py:409] Started an asynchronous checkpoint save for step 15
I0420 22:17:49.750071 134264641779520 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.286, Tokens/s/device: 13910.397, total_weights: 65536, loss: 5.858
I0420 22:17:49.757603 134264641779520 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.202, Tokens/s/device: 13746.996, total_weights: 65536, loss: 5.713
I0420 22:18:09.531937 134264641779520 metric_logger.py:185] completed step: 16, seconds: 55.437, TFLOP/s/device: 0.245, Tokens/s/device: 36.943, total_weights: 65536, loss: 5.599
I0420 22:18:09.679284 134264641779520 metric_logger.py:185] completed step: 17, seconds: 19.629, TFLOP/s/device: 0.692, Tokens/s/device: 104.334, total_weights: 65536, loss: 5.511
I0420 22:18:09.826869 134264641779520 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.616, Tokens/s/device: 13357.161, total_weights: 65536, loss: 5.442
I0420 22:18:09.975459 134264641779520 checkpointing.py:794] Waiting for step 19 to finish before checkpoint...
I0420 22:18:09.976084 134264641779520 checkpointing.py:798] Waited 0.0006353855133056641 seconds for step 19 to finish before starting checkpointing.
I0420 22:18:09.978076 134264641779520 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 22:18:18.847879 134143337805568 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_true/checkpoints/15/items/array_metadatas/process_1
I0420 22:18:54.187791 134141198714624 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 24.229 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 65.19399070739746 s) (per-host)
I0420 22:18:54.187906 134141198714624 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 64.441062s.
I0420 22:19:02.857450 134141198714624 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 73.110592s.
I0420 22:19:02.857743 134143321020160 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 22:19:02.857876 134143321020160 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 22:19:02.857924 134143321020160 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0420 22:19:02.860132 134143321020160 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0420 22:19:02.860289 134264641779520 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 22:19:02.860444 134264641779520 checkpoint_manager.py:1458] Waiting for previous save to complete took 52.882378 seconds. If this number is high, consider checkpointing less frequently.
I0420 22:19:02.862008 134264641779520 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 19
I0420 22:19:02.865387 134264641779520 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_true/checkpoints/19.
I0420 22:19:04.268907 134264641779520 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 22:19:04.269016 134264641779520 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 22:19:04.314051 134264641779520 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047197s
I0420 22:19:04.314215 134264641779520 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.038 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30621337890625 s) (per-host)
I0420 22:19:04.314260 134264641779520 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306268s (batch_requests_ready=0.254455s, total_serialization_initiated=0.051754s, others=0.000060s)
I0420 22:19:04.314350 134264641779520 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.310385s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.310374)
I0420 22:19:04.315240 134144436725504 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0420 22:19:04.315413 134264641779520 async_checkpointer.py:561] Finished blocking save. Time taken: 1.453321s. 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_true/checkpoints/19.
I0420 22:19:04.343247 134264641779520 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 22:19:04.343572 134143321020160 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 22:19:04.343728 134264641779520 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_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776723489.9780507, 'wait_for_prev_duration_secs': 52.88237762451172, 'time_between_consecutive_saves_sec': 22.126072883605957, 'checkpointer_blocking_start_time': 1776723542.8620467, 'checkpointer_blocking_duration_secs': 1.4534764289855957, 'get_old_steps_start_time': 1776723544.315536, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1776723489.9763002, 'checkpoint_manager_blocking_duration_secs': 54.36739659309387}
I0420 22:19:04.343834 134264641779520 checkpointing.py:409] Started an asynchronous checkpoint save for step 19
I0420 22:19:04.343875 134264641779520 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 22:19:12.470168 134146038945536 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_true/checkpoints/19/items/array_metadatas/process_1
I0420 22:19:48.608094 134144436725504 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 35.417 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.6000554561615 s) (per-host)
I0420 22:19:48.608216 134144436725504 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.292863s.
I0420 22:19:57.442078 134144436725504 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 53.126710s.
I0420 22:19:57.442335 134143321020160 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 22:19:57.442468 134143321020160 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 22:19:57.442518 134143321020160 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0420 22:19:57.444460 134143321020160 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0420 22:19:57.444634 134264641779520 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 22:19:57.444802 134264641779520 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 22:19:57.445434 134264641779520 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.416, Tokens/s/device: 13929.888, 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 22:20:07 UTC 2026
EXIT_CODE=0