XPK Start: Mon Apr 20 21:56:24 UTC 2026 2026-04-20 21:56:28.738845: 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 00:00:1776722188.751711 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776722188.755440 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776722188.767009 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776722188.767027 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776722188.767030 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776722188.767032 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 21:56:47.840079: 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 21:56:48.366564 132436679788352 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 21:56:57,406:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 21:56:57.406268 132436679788352 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 21:56:57,408:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-tosaj-slice-job-0-0.mt-15-ocdbt-true-rs-tosaj:8482 I0420 21:56:57.408583 132436679788352 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-tosaj-slice-job-0-0.mt-15-ocdbt-true-rs-tosaj:8482 I0420 21:56:58.641983 132436679788352 max_utils.py:284] Jax distributed system initialized! I0420 21:57:05.606314 132436679788352 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 21:57:05.606421 132436679788352 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 21:57:05.606480 132436679788352 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 21:57:05.606515 132436679788352 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 21:57:06.227766 132436679788352 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 21:57:06.339790 132436679788352 checkpointing.py:688] Setting up checkpoint logger... I0420 21:57:06.339910 132436679788352 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 21:57:06.339957 132436679788352 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 21:57:06.340158 132436679788352 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 0x78732fd4bb90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 21:57:07.679408 132436679788352 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 21:57:07.679621 132436679788352 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 0x785ebb6ddb80>}, handler_registry=None I0420 21:57:07.679855 132436679788352 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x785ebb6ddb80>` 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 21:57:07.679903 132436679788352 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78603a5067b0>` 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 21:57:07.679939 132436679788352 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 0x785ebb6ddb80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x785ebb6ddb80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78603a5067b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78603a5067b0>}). I0420 21:57:07.680262 132436679788352 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 21:57:07.680338 132436679788352 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x786038170f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 21:57:13.092311 132315527247616 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776721933033808300, 'commit_timestamp_nsecs': 1776721980828144461, '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_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0420 21:57:13.754127 132315518854912 checkpoint_manager.py:1877] Missing metrics for step 5 I0420 21:57:14.739063 132436679788352 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_15_ocdbt_true/checkpoints I0420 21:57:14.741894 132436679788352 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78603a510230> I0420 21:57:14.742074 132436679788352 checkpointing.py:302] Checkpoint manager created! I0420 21:57:15.669743 132436679788352 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 21:57:15.669867 132436679788352 checkpointing.py:582] restoring from this run's directory step 9 I0420 21:57:17.026769 132436679788352 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_15_ocdbt_true/checkpoints/9. I0420 21:57:20.013815 1438 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 21:57:27.523840 132436679788352 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 210.260 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.5158164501190186 s) (per-host) I0420 21:57:27.531929 132436679788352 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 199.598 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.928980350494385 s) (per-host) I0420 21:57:27.534163 132436679788352 checkpointer.py:318] Finished restoring checkpoint in 10.75 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9. I0420 21:57:27.534256 132436679788352 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_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776722236.7801058, 'checkpointer_duration_secs': 10.754122257232666, 'checkpoint_manager_start_time': 1776722235.6705394, 'checkpoint_manager_duration_secs': 11.863690614700317} fsdp: 32 I0420 21:57:27.614027 132436679788352 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 21:57:27.614145 132436679788352 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:57:27.614200 132436679788352 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:57:27.614242 132436679788352 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 21:57:27.614280 132436679788352 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:57:27.614314 132436679788352 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:57:27.614351 132436679788352 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:57:27.614386 132436679788352 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 21:57:27.614427 132436679788352 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:57:27.614477 132436679788352 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:57:27.614511 132436679788352 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 21:57:27.614544 132436679788352 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 21:57:28.010554 132436679788352 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:57:28.010700 132436679788352 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:57:28.016083 132436679788352 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:57:28.016179 132436679788352 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:57:28.166749 132436679788352 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:57:28.166886 132436679788352 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:57:28.313465 132436679788352 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:57:28.313605 132436679788352 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:57:28.462610 132436679788352 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 21:57:28.462768 132436679788352 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:57:28.601870 132436679788352 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 21:57:28.602011 132436679788352 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:57:28.742231 132436679788352 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 21:57:28.742369 132436679788352 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:57:28.892223 132436679788352 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 21:57:28.892362 132436679788352 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:57:28.914806 132436679788352 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 21:57:28.914898 132436679788352 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 21:57:40.188010 132436679788352 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 21:57:40.190377 132436679788352 metric_logger.py:289] number parameters: 1.104 billion I0420 21:57:50.885134 132436679788352 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 21:57:51.036593 132436679788352 checkpointing.py:798] Waited 0.15143799781799316 seconds for step 10 to finish before starting checkpointing. I0420 21:57:51.039371 132436679788352 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:57:51.041072 132436679788352 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0420 21:57:51.043649 132436679788352 async_checkpointer.py:452] [process=6] 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_15_ocdbt_true/checkpoints/10. I0420 21:57:52.059677 132436679788352 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 21:57:52.060766 132436679788352 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:57:52.060826 132436679788352 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 21:57:52.315108 132436679788352 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256557s I0420 21:57:52.315278 132436679788352 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: 0.5143380165100098 s) (per-host) I0420 21:57:52.315335 132436679788352 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.514404s (batch_requests_ready=0.253129s, total_serialization_initiated=0.261201s, others=0.000074s) I0420 21:57:52.315442 132436679788352 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.518432s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.518413) I0420 21:57:52.316319 132314017281792 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0420 21:57:52.316512 132436679788352 async_checkpointer.py:561] Finished blocking save. Time taken: 1.275367s. 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_15_ocdbt_true/checkpoints/10. I0420 21:57:53.168710 132436679788352 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:57:53.169107 132310284363520 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:57:53.169266 132436679788352 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722271.0393527, 'wait_for_prev_duration_secs': 6.628036499023438e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776722271.0411115, 'checkpointer_blocking_duration_secs': 1.2755131721496582, 'get_old_steps_start_time': 1776722272.3166444, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776722271.0369658, 'checkpoint_manager_blocking_duration_secs': 2.132263422012329} I0420 21:57:53.169381 132436679788352 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 21:57:53.169473 132436679788352 max_utils.py:750] Memstats: After params initialized: I0420 21:57:53.169533 132436679788352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0420 21:57:53.169566 132436679788352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0420 21:57:53.169593 132436679788352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0420 21:57:53.169626 132436679788352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0420 21:57:53.444763 132436679788352 metric_logger.py:185] completed step: 10, seconds: 10.694, TFLOP/s/device: 1.271, Tokens/s/device: 191.514, total_weights: 65536, loss: 6.827 I0420 21:57:53.610820 132436679788352 metric_logger.py:185] completed step: 11, seconds: 2.558, TFLOP/s/device: 5.312, Tokens/s/device: 800.650, total_weights: 65536, loss: 6.522 I0420 21:57:53.758260 132436679788352 metric_logger.py:185] completed step: 12, seconds: 0.011, TFLOP/s/device: 1291.552, Tokens/s/device: 194676.806, total_weights: 65536, loss: 6.260 I0420 21:57:53.905712 132436679788352 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 83.095, Tokens/s/device: 12524.921, total_weights: 65536, loss: 6.039 I0420 21:57:53.913158 132436679788352 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0420 21:57:54.200821 132436679788352 checkpointing.py:798] Waited 0.28763675689697266 seconds for step 15 to finish before starting checkpointing. I0420 21:57:54.203427 132436679788352 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:58:01.246884 132310787663616 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6 I0420 21:58:35.693951 132314017281792 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 35.987 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.89297533035278 s) (per-host) I0420 21:58:35.694073 132314017281792 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.377621s. I0420 21:58:45.032107 132314017281792 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 52.715640s. I0420 21:58:45.032394 132310284363520 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:58:45.032526 132310284363520 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:58:45.032574 132310284363520 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:58:45.034213 132310284363520 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 21:58:45.034389 132436679788352 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. W0420 21:58:45.034539 132436679788352 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.831118 seconds. If this number is high, consider checkpointing less frequently. I0420 21:58:45.036661 132436679788352 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0420 21:58:45.040719 132436679788352 async_checkpointer.py:452] [process=6] 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_15_ocdbt_true/checkpoints/15. I0420 21:58:46.036418 132436679788352 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:58:46.036533 132436679788352 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 21:58:46.084641 132436679788352 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.050262s I0420 21:58:46.084827 132436679788352 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.929 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31295084953308105 s) (per-host) I0420 21:58:46.084883 132436679788352 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313022s (batch_requests_ready=0.258108s, total_serialization_initiated=0.054832s, others=0.000082s) I0420 21:58:46.084977 132436679788352 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318382s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=0.318368) I0420 21:58:46.085850 132311869798144 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0420 21:58:46.086007 132436679788352 async_checkpointer.py:561] Finished blocking save. Time taken: 1.049273s. 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_15_ocdbt_true/checkpoints/15. I0420 21:58:46.519167 132436679788352 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:58:46.519572 132310284363520 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:58:46.519745 132436679788352 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722274.2033978, 'wait_for_prev_duration_secs': 50.83111810684204, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776722325.036701, 'checkpointer_blocking_duration_secs': 1.04941725730896, 'get_old_steps_start_time': 1776722326.086138, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776722274.201182, 'checkpoint_manager_blocking_duration_secs': 52.31852722167969} I0420 21:58:46.519867 132436679788352 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0420 21:58:46.520673 132436679788352 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.366, Tokens/s/device: 13922.407, total_weights: 65536, loss: 5.858 I0420 21:58:46.528172 132436679788352 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.384, Tokens/s/device: 13774.364, total_weights: 65536, loss: 5.713 I0420 21:59:05.344497 132436679788352 metric_logger.py:185] completed step: 16, seconds: 52.613, TFLOP/s/device: 0.258, Tokens/s/device: 38.925, total_weights: 65536, loss: 5.599 I0420 21:59:05.492109 132436679788352 metric_logger.py:185] completed step: 17, seconds: 18.671, TFLOP/s/device: 0.728, Tokens/s/device: 109.687, total_weights: 65536, loss: 5.511 I0420 21:59:05.639541 132436679788352 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.729, Tokens/s/device: 13374.170, total_weights: 65536, loss: 5.442 I0420 21:59:05.788123 132436679788352 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 21:59:05.788779 132436679788352 checkpointing.py:798] Waited 0.0006666183471679688 seconds for step 19 to finish before starting checkpointing. I0420 21:59:05.790764 132436679788352 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:59:13.011060 132311886583552 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6 I0420 21:59:47.608890 132311869798144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.545 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.83696484565735 s) (per-host) I0420 21:59:47.609013 132311869798144 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.523053s. I0420 21:59:57.923382 132311869798144 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 71.837408s. I0420 21:59:57.923706 132310284363520 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:59:57.923836 132310284363520 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:59:57.923889 132310284363520 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:59:57.925324 132310284363520 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0420 21:59:57.925508 132436679788352 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. W0420 21:59:57.925650 132436679788352 checkpoint_manager.py:1458] Waiting for previous save to complete took 52.134897 seconds. If this number is high, consider checkpointing less frequently. I0420 21:59:57.927300 132436679788352 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0420 21:59:57.930868 132436679788352 async_checkpointer.py:452] [process=6] 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_15_ocdbt_true/checkpoints/19. I0420 21:59:58.464318 132436679788352 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:59:58.464427 132436679788352 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 21:59:58.507995 132436679788352 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045796s I0420 21:59:58.508157 132436679788352 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.139 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3001434803009033 s) (per-host) I0420 21:59:58.508205 132436679788352 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.300201s (batch_requests_ready=0.249902s, total_serialization_initiated=0.050236s, others=0.000063s) I0420 21:59:58.508297 132436679788352 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303976s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.303964) I0420 21:59:58.509206 132314008889088 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0420 21:59:58.509332 132436679788352 async_checkpointer.py:561] Finished blocking save. Time taken: 0.581964s. 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_15_ocdbt_true/checkpoints/19. I0420 21:59:59.362127 132436679788352 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:59:59.362509 132310284363520 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:59:59.362680 132436679788352 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722345.7907357, 'wait_for_prev_duration_secs': 52.134897232055664, 'time_between_consecutive_saves_sec': 20.75648045539856, 'checkpointer_blocking_start_time': 1776722397.9273386, 'checkpointer_blocking_duration_secs': 0.582183837890625, 'get_old_steps_start_time': 1776722398.5095365, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1776722345.789003, 'checkpoint_manager_blocking_duration_secs': 53.57363748550415} I0420 21:59:59.362819 132436679788352 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 21:59:59.362866 132436679788352 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 22:00:06.081132 132314017281792 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6 I0420 22:00:40.896077 132314008889088 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.003 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.688023805618286 s) (per-host) I0420 22:00:40.896198 132314008889088 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.386791s. I0420 22:00:49.640213 132314008889088 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.130792s. I0420 22:00:49.640529 132310284363520 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 22:00:49.640656 132310284363520 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0420 22:00:49.640708 132310284363520 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 22:00:49.642702 132310284363520 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 22:00:49.642869 132436679788352 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. I0420 22:00:49.643041 132436679788352 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 22:00:49.643641 132436679788352 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.125, Tokens/s/device: 13886.064, 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 22:01:01 UTC 2026 EXIT_CODE=0