XPK Start: Wed Apr 22 18:51:43 UTC 2026 2026-04-22 18:51:47.319057: 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 00:00:1776883907.331892 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776883907.335881 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776883907.347225 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883907.347243 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883907.347246 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883907.347248 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 18:52:06.634328: 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 18:52:07.169570 140033603671872 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 18:52:16,211:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 18:52:16.211902 140033603671872 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 18:52:16,215:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-lhdjf-slice-job-0-0.mt-16-zarr3-true-rs-lhdjf:8482 I0422 18:52:16.215191 140033603671872 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-lhdjf-slice-job-0-0.mt-16-zarr3-true-rs-lhdjf:8482 I0422 18:52:17.472432 140033603671872 max_utils.py:284] Jax distributed system initialized! I0422 18:52:24.649293 140033603671872 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:52:24.649398 140033603671872 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:52:24.649438 140033603671872 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 18:52:24.649471 140033603671872 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:52:25.320905 140033603671872 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:52:25.437304 140033603671872 checkpointing.py:688] Setting up checkpoint logger... I0422 18:52:25.437423 140033603671872 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 18:52:25.437485 140033603671872 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 18:52:25.437738 140033603671872 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 0x7f5b96d78fb0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:52:26.780837 140033603671872 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 18:52:26.781189 140033603671872 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 0x7f476425cc20>}, handler_registry=None I0422 18:52:26.781493 140033603671872 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f476425cc20>` 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 18:52:26.781550 140033603671872 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f4764087530>` 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 18:52:26.781588 140033603671872 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 0x7f476425cc20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f476425cc20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f4764087530>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f4764087530>}). I0422 18:52:26.781922 140033603671872 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 18:52:26.781997 140033603671872 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f471c4e0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:52:28.254827 139913768978176 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776883664400641161, 'commit_timestamp_nsecs': 1776883706378991793, 'custom_metadata': {}} from 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/_CHECKPOINT_METADATA I0422 18:52:28.494749 139913768978176 checkpoint_manager.py:1877] Missing metrics for step 0 I0422 18:52:30.217137 140033603671872 checkpoint_manager.py:1818] Found 3 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 18:52:30.219554 140033603671872 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 0x7f4764087950> I0422 18:52:30.219671 140033603671872 checkpointing.py:302] Checkpoint manager created! I0422 18:52:31.176630 140033603671872 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:52:31.176752 140033603671872 checkpointing.py:582] restoring from this run's directory step 9 I0422 18:52:31.678656 140033603671872 checkpointer.py:304] Restoring checkpoint from 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 18:52:33.102302 1426 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 18:52:40.239614 140033603671872 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 221.233 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.143032073974609 s) (per-host) I0422 18:52:40.248279 140033603671872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 209.481 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.554903030395508 s) (per-host) I0422 18:52:41.484134 140033603671872 checkpointer.py:318] Finished restoring checkpoint in 10.06 seconds from 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 18:52:41.484291 140033603671872 standard_logger.py:34] {'step': 9, 'event_type': 'restore', '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', 'checkpointer_start_time': 1776883951.4254181, 'checkpointer_duration_secs': 10.058839082717896, 'checkpoint_manager_start_time': 1776883951.1773653, 'checkpoint_manager_duration_secs': 10.306893825531006} fsdp: 32 I0422 18:52:41.565188 140033603671872 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 18:52:41.565309 140033603671872 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:52:41.565362 140033603671872 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:52:41.565407 140033603671872 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 18:52:41.565446 140033603671872 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:52:41.565483 140033603671872 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:52:41.565520 140033603671872 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:52:41.565565 140033603671872 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 18:52:41.565601 140033603671872 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:52:41.565636 140033603671872 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:52:41.565670 140033603671872 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 18:52:41.565703 140033603671872 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 18:52:41.957984 140033603671872 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:52:41.958151 140033603671872 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:52:41.963697 140033603671872 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:52:41.963792 140033603671872 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:52:42.109675 140033603671872 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:52:42.109804 140033603671872 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:52:42.248494 140033603671872 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:52:42.248623 140033603671872 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:52:42.396706 140033603671872 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 18:52:42.396841 140033603671872 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:52:42.535883 140033603671872 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 18:52:42.536011 140033603671872 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:52:42.674868 140033603671872 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 18:52:42.674995 140033603671872 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:52:42.826568 140033603671872 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:52:42.826696 140033603671872 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:52:42.849518 140033603671872 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:52:42.849611 140033603671872 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 18:52:53.978967 140033603671872 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 18:52:53.981348 140033603671872 metric_logger.py:289] number parameters: 1.104 billion I0422 18:53:05.143921 140033603671872 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0422 18:53:05.297764 140033603671872 checkpointing.py:798] Waited 0.15382957458496094 seconds for step 10 to finish before starting checkpointing. I0422 18:53:05.300385 140033603671872 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:53:05.302305 140033603671872 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0422 18:53:05.304913 140033603671872 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/10. I0422 18:53:05.880496 140033603671872 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 18:53:05.881633 140033603671872 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:53:05.881692 140033603671872 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 18:53:06.134229 140033603671872 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254880s I0422 18:53:06.134404 140033603671872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.999 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.514415979385376 s) (per-host) I0422 18:53:06.134456 140033603671872 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.514477s (batch_requests_ready=0.254398s, total_serialization_initiated=0.260009s, others=0.000070s) I0422 18:53:06.134551 140033603671872 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.518469s (all_items=0.000021s, per_item={'items': '0.00002050'}, temp_paths=0.518448) I0422 18:53:06.135448 139913164998400 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:53:06.135573 140033603671872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.833202s. 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/10. I0422 18:53:06.585302 140033603671872 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:53:06.585669 139908870031104 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:53:06.585835 140033603671872 standard_logger.py:34] {'step': 10, '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': 1776883985.3003664, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883985.302343, 'checkpointer_blocking_duration_secs': 0.8334105014801025, 'get_old_steps_start_time': 1776883986.1357737, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776883985.2981575, 'checkpoint_manager_blocking_duration_secs': 1.2876405715942383} I0422 18:53:06.585951 140033603671872 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0422 18:53:06.586025 140033603671872 max_utils.py:750] Memstats: After params initialized: I0422 18:53:06.586084 140033603671872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 18:53:06.586135 140033603671872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 18:53:06.586164 140033603671872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 18:53:06.586189 140033603671872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 18:53:06.855158 140033603671872 metric_logger.py:185] completed step: 10, seconds: 11.161, TFLOP/s/device: 1.217, Tokens/s/device: 183.491, total_weights: 65536, loss: 6.827 I0422 18:53:07.033875 140033603671872 metric_logger.py:185] completed step: 11, seconds: 1.710, TFLOP/s/device: 7.947, Tokens/s/device: 1197.827, total_weights: 65536, loss: 6.522 I0422 18:53:07.181330 140033603671872 metric_logger.py:185] completed step: 12, seconds: 0.023, TFLOP/s/device: 589.208, Tokens/s/device: 88811.795, total_weights: 65536, loss: 6.260 I0422 18:53:07.328939 140033603671872 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 82.964, Tokens/s/device: 12505.267, total_weights: 65536, loss: 6.039 I0422 18:53:07.335594 140033603671872 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0422 18:53:07.624011 140033603671872 checkpointing.py:798] Waited 0.28838634490966797 seconds for step 15 to finish before starting checkpointing. I0422 18:53:07.626994 140033603671872 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:53:10.706522 139909381723904 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/10/items/array_metadatas/process_6 I0422 18:53:47.070307 139913164998400 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.108 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.45027995109558 s) (per-host) I0422 18:53:47.070432 139913164998400 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.934777s. I0422 18:53:55.975729 139913164998400 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.840058s. I0422 18:53:55.975942 139908870031104 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:53:55.976001 139908870031104 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:53:55.976047 139908870031104 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:53:55.977724 139908870031104 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0422 18:53:55.977895 140033603671872 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0422 18:53:55.978034 140033603671872 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.351044 seconds. If this number is high, consider checkpointing less frequently. I0422 18:53:55.980064 140033603671872 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0422 18:53:55.984122 140033603671872 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/15. I0422 18:53:56.540961 140033603671872 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:53:56.541060 140033603671872 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 18:53:56.592271 140033603671872 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.053311s I0422 18:53:56.592449 140033603671872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.912 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31406426429748535 s) (per-host) I0422 18:53:56.592501 140033603671872 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314126s (batch_requests_ready=0.256176s, total_serialization_initiated=0.057880s, others=0.000069s) I0422 18:53:56.592594 140033603671872 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.319140s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.319127) I0422 18:53:56.593409 139909423687424 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:53:56.593569 140033603671872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.613417s. 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/15. I0422 18:53:56.604147 140033603671872 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:53:56.604447 139908870031104 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:53:56.604604 140033603671872 standard_logger.py:34] {'step': 15, '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': 1776883987.6269658, 'wait_for_prev_duration_secs': 48.35104417800903, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776884035.980118, 'checkpointer_blocking_duration_secs': 0.6135585308074951, 'get_old_steps_start_time': 1776884036.5936966, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776883987.6244383, 'checkpoint_manager_blocking_duration_secs': 48.98013424873352} I0422 18:53:56.604710 140033603671872 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0422 18:53:56.605519 140033603671872 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.439, Tokens/s/device: 13933.395, total_weights: 65536, loss: 5.858 I0422 18:53:56.612969 140033603671872 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.781, Tokens/s/device: 13834.192, total_weights: 65536, loss: 5.713 I0422 18:54:14.737636 140033603671872 metric_logger.py:185] completed step: 16, seconds: 49.276, TFLOP/s/device: 0.276, Tokens/s/device: 41.562, total_weights: 65536, loss: 5.599 I0422 18:54:14.884851 140033603671872 metric_logger.py:185] completed step: 17, seconds: 17.979, TFLOP/s/device: 0.756, Tokens/s/device: 113.912, total_weights: 65536, loss: 5.511 I0422 18:54:15.032297 140033603671872 metric_logger.py:185] completed step: 18, seconds: 0.154, TFLOP/s/device: 88.313, Tokens/s/device: 13311.494, total_weights: 65536, loss: 5.442 I0422 18:54:15.180791 140033603671872 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0422 18:54:15.181456 140033603671872 checkpointing.py:798] Waited 0.0006763935089111328 seconds for step 19 to finish before starting checkpointing. I0422 18:54:15.183419 140033603671872 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:54:19.099334 139909381723904 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/15/items/array_metadatas/process_6 I0422 18:54:55.781197 139909423687424 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.547 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.50277328491211 s) (per-host) I0422 18:54:55.781300 139909423687424 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.187780s. I0422 18:55:05.611406 139909423687424 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 69.017871s. I0422 18:55:05.611697 139908870031104 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:55:05.611827 139908870031104 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:55:05.611877 139908870031104 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:55:05.613589 139908870031104 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0422 18:55:05.613751 140033603671872 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0422 18:55:05.613889 140033603671872 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.430481 seconds. If this number is high, consider checkpointing less frequently. I0422 18:55:05.615528 140033603671872 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0422 18:55:05.639445 140033603671872 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/19. I0422 18:55:06.178644 140033603671872 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:55:06.178745 140033603671872 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 18:55:06.232116 140033603671872 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.055505s I0422 18:55:06.232282 140033603671872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.972 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3102443218231201 s) (per-host) I0422 18:55:06.232334 140033603671872 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.310307s (batch_requests_ready=0.250189s, total_serialization_initiated=0.060050s, others=0.000068s) I0422 18:55:06.232434 140033603671872 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.314341s (all_items=0.000013s, per_item={'items': '0.00001264'}, temp_paths=0.314328) I0422 18:55:06.233332 139912602949376 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:55:06.233492 140033603671872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.617895s. 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/19. I0422 18:55:06.663270 140033603671872 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:55:06.663622 139908870031104 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:55:06.663791 140033603671872 standard_logger.py:34] {'step': 19, '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': 1776884055.183391, 'wait_for_prev_duration_secs': 50.43048119544983, 'time_between_consecutive_saves_sec': 19.20562720298767, 'checkpointer_blocking_start_time': 1776884105.6155667, 'checkpointer_blocking_duration_secs': 0.6180346012115479, 'get_old_steps_start_time': 1776884106.2336137, 'get_old_steps_duration_secs': 2.0742416381835938e-05, 'checkpoint_manager_blocking_start_time': 1776884055.1816802, 'checkpoint_manager_blocking_duration_secs': 51.48207712173462} I0422 18:55:06.663915 140033603671872 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0422 18:55:06.663957 140033603671872 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:55:12.790731 139913164998400 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/19/items/array_metadatas/process_6 I0422 18:55:48.643550 139912602949376 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.974 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.72146487236023 s) (per-host) I0422 18:55:48.643669 139912602949376 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.410225s. I0422 18:55:57.480631 139912602949376 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.247171s. I0422 18:55:57.480917 139908870031104 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:55:57.481034 139908870031104 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:55:57.481078 139908870031104 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:55:57.482620 139908870031104 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0422 18:55:57.482783 140033603671872 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0422 18:55:57.482951 140033603671872 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:55:57.483693 140033603671872 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.522, Tokens/s/device: 13945.919, 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: Wed Apr 22 18:56:09 UTC 2026 EXIT_CODE=0