XPK Start: Thu Apr 23 18:58:18 UTC 2026 2026-04-23 18:58:22.118772: 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 00:00:1776970702.132017 9 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776970702.135783 9 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776970702.147454 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970702.147471 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970702.147473 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970702.147475 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 18:58:41.442817: 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) I0423 18:58:41.980259 135793675622208 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 18:58:51,022:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 18:58:51.022489 135793675622208 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 18:58:51,024:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-con9q-slice-job-0-0.mt-16-zarr3-false-sv-con9q:8482 I0423 18:58:51.024799 135793675622208 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-con9q-slice-job-0-0.mt-16-zarr3-false-sv-con9q:8482 I0423 18:58:51.949445 135793675622208 max_utils.py:284] Jax distributed system initialized! I0423 18:58:58.896691 135793675622208 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 18:58:58.896801 135793675622208 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 18:58:58.896842 135793675622208 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 I0423 18:58:58.896879 135793675622208 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 18:58:59.527162 135793675622208 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 18:58:59.643485 135793675622208 checkpointing.py:688] Setting up checkpoint logger... I0423 18:58:59.643605 135793675622208 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0423 18:58:59.643651 135793675622208 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 18:58:59.643852 135793675622208 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 0x7b8067905dc0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 18:59:02.623558 135793675622208 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 18:59:02.623794 135793675622208 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 0x7b6c082d0ef0>}, handler_registry=None I0423 18:59:02.624028 135793675622208 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b6c082d0ef0>` 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`. I0423 18:59:02.624076 135793675622208 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b6c080eb860>` 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`. I0423 18:59:02.624127 135793675622208 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 0x7b6c082d0ef0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b6c082d0ef0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b6c080eb860>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b6c080eb860>}). I0423 18:59:02.624478 135793675622208 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 18:59:02.624557 135793675622208 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b6bc825cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 18:59:03.688595 135793675622208 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints I0423 18:59:04.128038 135793675622208 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b6c080ff080> I0423 18:59:04.128260 135793675622208 checkpointing.py:302] Checkpoint manager created! I0423 18:59:05.060694 135793675622208 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 18:59:05.060814 135793675622208 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 18:59:06.531208 135793675622208 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0423 18:59:06.531329 135793675622208 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:59:06.531383 135793675622208 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:59:06.531427 135793675622208 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0423 18:59:06.531466 135793675622208 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:59:06.531501 135793675622208 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:59:06.531543 135793675622208 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:59:06.531582 135793675622208 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0423 18:59:06.531617 135793675622208 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:59:06.531652 135793675622208 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:59:06.531687 135793675622208 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0423 18:59:06.531725 135793675622208 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0423 18:59:06.938083 135793675622208 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:59:06.938243 135793675622208 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:59:06.943693 135793675622208 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:59:06.943790 135793675622208 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:59:07.091831 135793675622208 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:59:07.091964 135793675622208 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:59:07.243137 135793675622208 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:59:07.243276 135793675622208 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:59:07.400296 135793675622208 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:59:07.400437 135793675622208 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:59:07.546443 135793675622208 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:59:07.546579 135793675622208 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:59:07.692197 135793675622208 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:59:07.692335 135793675622208 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:59:07.847080 135793675622208 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 18:59:07.847240 135793675622208 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:59:07.871184 135793675622208 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 18:59:07.871284 135793675622208 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 18:59:19.212740 135793675622208 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. I0423 18:59:19.215438 135793675622208 metric_logger.py:289] number parameters: 1.104 billion I0423 18:59:30.295334 135793675622208 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 18:59:30.447242 135793675622208 checkpointing.py:798] Waited 0.1519007682800293 seconds for step 0 to finish before starting checkpointing. I0423 18:59:30.449652 135793675622208 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:59:30.451293 135793675622208 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0423 18:59:30.454261 135793675622208 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/0. I0423 18:59:31.058439 135793675622208 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 18:59:31.059513 135793675622208 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:59:31.059574 135793675622208 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 I0423 18:59:31.312332 135793675622208 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.255094s I0423 18:59:31.312510 135793675622208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.876 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5363996028900146 s) (per-host) I0423 18:59:31.312567 135793675622208 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.536468s (batch_requests_ready=0.274516s, total_serialization_initiated=0.261877s, others=0.000074s) I0423 18:59:31.312660 135793675622208 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.540434s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.540415) I0423 18:59:31.313669 135667816331008 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:59:31.313838 135793675622208 async_checkpointer.py:561] Finished blocking save. Time taken: 0.862464s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/0. I0423 18:59:31.331376 135793675622208 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:59:31.331692 135666717411072 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:59:31.331852 135793675622208 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970770.4496324, 'wait_for_prev_duration_secs': 6.866455078125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776970770.4513395, 'checkpointer_blocking_duration_secs': 0.8626065254211426, 'get_old_steps_start_time': 1776970771.3139656, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776970770.447652, 'checkpoint_manager_blocking_duration_secs': 0.884162187576294} I0423 18:59:31.331964 135793675622208 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 18:59:31.332040 135793675622208 max_utils.py:750] Memstats: After params initialized: I0423 18:59:31.332091 135793675622208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0423 18:59:31.332141 135793675622208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0423 18:59:31.332169 135793675622208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0423 18:59:31.332195 135793675622208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0423 18:59:31.607525 135793675622208 metric_logger.py:185] completed step: 0, seconds: 11.079, TFLOP/s/device: 1.226, Tokens/s/device: 184.857, total_weights: 65536, loss: 10.874 I0423 18:59:31.783695 135793675622208 metric_logger.py:185] completed step: 1, seconds: 1.310, TFLOP/s/device: 10.369, Tokens/s/device: 1562.983, total_weights: 65536, loss: 10.874 I0423 18:59:32.202070 135793675622208 metric_logger.py:185] completed step: 2, seconds: 0.030, TFLOP/s/device: 446.460, Tokens/s/device: 67295.370, total_weights: 65536, loss: 10.560 I0423 18:59:32.349834 135793675622208 metric_logger.py:185] completed step: 3, seconds: 0.409, TFLOP/s/device: 33.196, Tokens/s/device: 5003.653, total_weights: 65536, loss: 9.979 I0423 18:59:32.355853 135793675622208 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0423 18:59:32.644742 135793675622208 checkpointing.py:798] Waited 0.2888600826263428 seconds for step 5 to finish before starting checkpointing. I0423 18:59:32.647912 135793675622208 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:59:34.416462 2594 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 18:59:36.278150 135668302845696 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0423 19:00:13.638021 135667816331008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.853 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.86187148094177 s) (per-host) I0423 19:00:13.638155 135667816331008 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.324372s. I0423 19:00:20.539195 135667816331008 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.225396s. I0423 19:00:20.539454 135666717411072 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 19:00:20.539576 135666717411072 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 19:00:20.539627 135666717411072 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 19:00:20.541248 135666717411072 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 19:00:20.541429 135793675622208 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0423 19:00:20.541558 135793675622208 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.893651 seconds. If this number is high, consider checkpointing less frequently. I0423 19:00:20.543577 135793675622208 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0423 19:00:20.547550 135793675622208 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/5. I0423 19:00:21.121558 135793675622208 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 19:00:21.121710 135793675622208 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 I0423 19:00:21.172377 135793675622208 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.052952s I0423 19:00:21.172547 135793675622208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.842 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31856775283813477 s) (per-host) I0423 19:00:21.172599 135793675622208 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.318633s (batch_requests_ready=0.260175s, total_serialization_initiated=0.058385s, others=0.000073s) I0423 19:00:21.172692 135793675622208 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.323559s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.323543) I0423 19:00:21.173605 135667781572352 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 19:00:21.173759 135793675622208 async_checkpointer.py:561] Finished blocking save. Time taken: 0.630115s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/5. I0423 19:00:21.184733 135793675622208 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0423 19:00:21.185021 135666717411072 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 19:00:21.185168 135793675622208 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970772.6478827, 'wait_for_prev_duration_secs': 47.89365100860596, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776970820.5436149, 'checkpointer_blocking_duration_secs': 0.6302480697631836, 'get_old_steps_start_time': 1776970821.1738813, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776970772.6450896, 'checkpoint_manager_blocking_duration_secs': 48.54004669189453} I0423 19:00:21.185333 135793675622208 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0423 19:00:21.186118 135793675622208 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.334, Tokens/s/device: 12560.949, total_weights: 65536, loss: 9.457 I0423 19:00:21.193464 135793675622208 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.207, Tokens/s/device: 13898.409, total_weights: 65536, loss: 8.955 I0423 19:00:40.981300 135793675622208 metric_logger.py:185] completed step: 6, seconds: 48.836, TFLOP/s/device: 0.278, Tokens/s/device: 41.936, total_weights: 65536, loss: 8.465 I0423 19:00:41.128687 135793675622208 metric_logger.py:185] completed step: 7, seconds: 19.643, TFLOP/s/device: 0.692, Tokens/s/device: 104.263, total_weights: 65536, loss: 7.998 I0423 19:00:41.276155 135793675622208 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.703, Tokens/s/device: 13370.328, total_weights: 65536, loss: 7.567 I0423 19:00:41.424594 135793675622208 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 19:00:41.425220 135793675622208 checkpointing.py:798] Waited 0.0006375312805175781 seconds for step 9 to finish before starting checkpointing. I0423 19:00:41.427042 135793675622208 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 19:00:45.982038 135666147776256 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0423 19:01:27.021795 135667781572352 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 23.873 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 66.16776609420776 s) (per-host) I0423 19:01:27.021919 135667781572352 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 65.848205s. I0423 19:01:31.676670 135667781572352 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 70.502940s. I0423 19:01:31.676946 135666717411072 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 19:01:31.677070 135666717411072 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 19:01:31.677136 135666717411072 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0423 19:01:31.679456 135666717411072 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0423 19:01:31.679628 135793675622208 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0423 19:01:31.679765 135793675622208 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.252737 seconds. If this number is high, consider checkpointing less frequently. I0423 19:01:31.681330 135793675622208 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0423 19:01:31.684751 135793675622208 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/9. I0423 19:01:32.654788 135793675622208 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 19:01:32.654885 135793675622208 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 I0423 19:01:32.700312 135793675622208 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047574s I0423 19:01:32.700477 135793675622208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.044 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3058137893676758 s) (per-host) I0423 19:01:32.700528 135793675622208 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.305875s (batch_requests_ready=0.253608s, total_serialization_initiated=0.052201s, others=0.000066s) I0423 19:01:32.700618 135793675622208 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.309796s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.309785) I0423 19:01:32.701451 135667816331008 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 19:01:32.701610 135793675622208 async_checkpointer.py:561] Finished blocking save. Time taken: 1.020209s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/9. I0423 19:01:32.727735 135793675622208 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 19:01:32.728038 135666717411072 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 19:01:32.728215 135793675622208 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970841.4270134, 'wait_for_prev_duration_secs': 50.25273656845093, 'time_between_consecutive_saves_sec': 20.88572669029236, 'checkpointer_blocking_start_time': 1776970891.6813703, 'checkpointer_blocking_duration_secs': 1.0203461647033691, 'get_old_steps_start_time': 1776970892.7017295, 'get_old_steps_duration_secs': 1.9550323486328125e-05, 'checkpoint_manager_blocking_start_time': 1776970841.4254484, 'checkpoint_manager_blocking_duration_secs': 51.30273365974426} I0423 19:01:32.728321 135793675622208 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 19:01:32.728366 135793675622208 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 19:01:37.511850 135668302845696 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0423 19:02:15.831069 135667816331008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.366 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.4363739490509 s) (per-host) I0423 19:02:15.831188 135667816331008 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.129624s. I0423 19:02:21.916775 135667816331008 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.215194s. I0423 19:02:21.917009 135666717411072 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 19:02:21.917070 135666717411072 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 19:02:21.917143 135666717411072 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 19:02:21.918678 135666717411072 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 19:02:21.918874 135793675622208 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0423 19:02:21.919046 135793675622208 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 19:02:21.919717 135793675622208 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.356, Tokens/s/device: 13920.893, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 19:02:33 UTC 2026 EXIT_CODE=0