XPK Start: Mon Apr 20 22:06:06 UTC 2026 2026-04-20 22:06:10.266330: 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 00:00:1776722770.279107 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776722770.282896 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776722770.294158 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776722770.294176 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776722770.294179 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776722770.294181 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 22:06:29.528042: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 22:06:30.057212 138547772659520 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 22:06:39,098:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 22:06:39.098940 138547772659520 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 22:06:39,101:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-rs-13jlj-slice-job-0-0.mt-15-ocdbt-false-rs-13jlj:8482 I0420 22:06:39.101521 138547772659520 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-13jlj-slice-job-0-0.mt-15-ocdbt-false-rs-13jlj:8482 I0420 22:06:39.933873 138547772659520 max_utils.py:284] Jax distributed system initialized! I0420 22:06:47.116455 138547772659520 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 22:06:47.116563 138547772659520 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 22:06:47.116604 138547772659520 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0420 22:06:47.116639 138547772659520 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 22:06:47.736036 138547772659520 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 22:06:47.850404 138547772659520 checkpointing.py:688] Setting up checkpoint logger... I0420 22:06:47.850525 138547772659520 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0420 22:06:47.850571 138547772659520 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 22:06:47.850788 138547772659520 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e01a3f799d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 22:06:49.150813 138547772659520 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 22:06:49.151137 138547772659520 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dec84459dc0>}, handler_registry=None I0420 22:06:49.151440 138547772659520 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dec84459dc0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 22:06:49.151494 138547772659520 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dec5869c6b0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 22:06:49.151530 138547772659520 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 0x7dec84459dc0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dec84459dc0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dec5869c6b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dec5869c6b0>}). I0420 22:06:49.151883 138547772659520 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 22:06:49.151961 138547772659520 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dec843c0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 22:06:51.118578 138427039205120 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776722599983493056, 'commit_timestamp_nsecs': 1776722614802205140, '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_false/checkpoints/0/_CHECKPOINT_METADATA I0420 22:06:51.366818 138427039205120 checkpoint_manager.py:1877] Missing metrics for step 0 I0420 22:06:51.441615 138547772659520 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_false/checkpoints I0420 22:06:52.271245 138547772659520 checkpoint_manager.py:929] [process=7][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_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7dec84458d70> I0420 22:06:52.271429 138547772659520 checkpointing.py:302] Checkpoint manager created! I0420 22:06:53.198814 138547772659520 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 22:06:53.198940 138547772659520 checkpointing.py:582] restoring from this run's directory step 9 I0420 22:06:54.151237 138547772659520 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_false/checkpoints/9. I0420 22:06:57.264550 1457 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 22:07:07.755859 138547772659520 jax_array_handlers.py:837] [process=7] /jax/orbax/read/worker/io/requested throughput: 131.034 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.060055494308472 s) (per-host) I0420 22:07:07.775619 138547772659520 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/read/gbytes_per_sec: 126.942 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.46710753440857 s) (per-host) I0420 22:07:08.645604 138547772659520 checkpointer.py:318] Finished restoring checkpoint in 14.77 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_false/checkpoints/9. I0420 22:07:08.645788 138547772659520 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_false/checkpoints', 'checkpointer_start_time': 1776722813.878859, 'checkpointer_duration_secs': 14.766894817352295, 'checkpoint_manager_start_time': 1776722813.1995418, 'checkpoint_manager_duration_secs': 15.446213722229004} fsdp: 32 I0420 22:07:08.726756 138547772659520 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 22:07:08.726875 138547772659520 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 22:07:08.726926 138547772659520 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 22:07:08.726970 138547772659520 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 22:07:08.727007 138547772659520 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 22:07:08.727041 138547772659520 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 22:07:08.727077 138547772659520 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 22:07:08.727113 138547772659520 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 22:07:08.727148 138547772659520 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 22:07:08.727181 138547772659520 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 22:07:08.727214 138547772659520 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 22:07:08.727246 138547772659520 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 22:07:09.123178 138547772659520 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 22:07:09.123335 138547772659520 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 22:07:09.128759 138547772659520 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 22:07:09.128855 138547772659520 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 22:07:09.286543 138547772659520 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 22:07:09.286690 138547772659520 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 22:07:09.445005 138547772659520 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 22:07:09.445138 138547772659520 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 22:07:09.612108 138547772659520 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 22:07:09.612255 138547772659520 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 22:07:09.769451 138547772659520 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 22:07:09.769587 138547772659520 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 22:07:09.921348 138547772659520 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 22:07:09.921480 138547772659520 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 22:07:10.086161 138547772659520 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 22:07:10.086295 138547772659520 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 22:07:10.110379 138547772659520 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 22:07:10.110476 138547772659520 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 22:07:21.323531 138547772659520 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0420 22:07:21.326251 138547772659520 metric_logger.py:289] number parameters: 1.104 billion I0420 22:07:32.582252 138547772659520 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 22:07:32.734004 138547772659520 checkpointing.py:798] Waited 0.15173006057739258 seconds for step 10 to finish before starting checkpointing. I0420 22:07:32.736203 138547772659520 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 22:07:32.737950 138547772659520 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 10 I0420 22:07:32.740577 138547772659520 async_checkpointer.py:452] [process=7] 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_false/checkpoints/10. I0420 22:07:34.588536 138547772659520 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 22:07:34.589724 138547772659520 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 22:07:34.589782 138547772659520 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 22:07:34.837594 138547772659520 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250186s I0420 22:07:34.837782 138547772659520 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.638 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.9416296482086182 s) (per-host) I0420 22:07:34.837836 138547772659520 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.941694s (batch_requests_ready=0.686788s, total_serialization_initiated=0.254833s, others=0.000073s) I0420 22:07:34.837931 138547772659520 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.945726s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.945709) I0420 22:07:34.838858 138423239177984 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 22:07:34.839030 138547772659520 async_checkpointer.py:561] Finished blocking save. Time taken: 2.101004s. 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_false/checkpoints/10. I0420 22:07:34.840951 138547772659520 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 22:07:34.841183 138421050464000 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 22:07:34.841283 138547772659520 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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722852.736185, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776722852.737994, 'checkpointer_blocking_duration_secs': 2.1011464595794678, 'get_old_steps_start_time': 1776722854.8391583, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776722852.7343748, 'checkpoint_manager_blocking_duration_secs': 2.1068692207336426} I0420 22:07:34.841445 138547772659520 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 22:07:34.841516 138547772659520 max_utils.py:750] Memstats: After params initialized: I0420 22:07:34.841565 138547772659520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 22:07:34.841597 138547772659520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 22:07:34.841623 138547772659520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 22:07:34.841646 138547772659520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 22:07:35.117016 138547772659520 metric_logger.py:185] completed step: 10, seconds: 11.255, TFLOP/s/device: 1.207, Tokens/s/device: 181.964, total_weights: 65536, loss: 6.827 I0420 22:07:35.285136 138547772659520 metric_logger.py:185] completed step: 11, seconds: 2.533, TFLOP/s/device: 5.364, Tokens/s/device: 808.507, total_weights: 65536, loss: 6.522 I0420 22:07:35.432485 138547772659520 metric_logger.py:185] completed step: 12, seconds: 0.012, TFLOP/s/device: 1103.927, Tokens/s/device: 166395.840, total_weights: 65536, loss: 6.260 I0420 22:07:35.580117 138547772659520 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 83.062, Tokens/s/device: 12520.021, total_weights: 65536, loss: 6.039 I0420 22:07:35.586410 138547772659520 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0420 22:07:35.875246 138547772659520 checkpointing.py:798] Waited 0.2888057231903076 seconds for step 15 to finish before starting checkpointing. I0420 22:07:35.877899 138547772659520 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 22:07:48.267105 138421067249408 array_metadata_store.py:203] [process=7][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_false/checkpoints/10/items/array_metadatas/process_7 I0420 22:07:53.229002 138423239177984 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 81.706 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 19.332801580429077 s) (per-host) I0420 22:07:53.229126 138423239177984 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 18.390158s. I0420 22:08:02.995837 138423239177984 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 28.156845s. I0420 22:08:02.996184 138421050464000 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 22:08:02.996303 138421050464000 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 22:08:02.996359 138421050464000 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 22:08:02.999180 138421050464000 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 22:08:02.999385 138547772659520 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0420 22:08:02.999514 138547772659520 checkpoint_manager.py:1458] Waiting for previous save to complete took 27.121620 seconds. If this number is high, consider checkpointing less frequently. I0420 22:08:03.001455 138547772659520 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 15 I0420 22:08:03.005584 138547772659520 async_checkpointer.py:452] [process=7] 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_false/checkpoints/15. I0420 22:08:04.001776 138547772659520 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 22:08:04.001875 138547772659520 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 22:08:04.045595 138547772659520 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045834s I0420 22:08:04.045755 138547772659520 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.942 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3121204376220703 s) (per-host) I0420 22:08:04.045805 138547772659520 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.312179s (batch_requests_ready=0.261685s, total_serialization_initiated=0.050428s, others=0.000067s) I0420 22:08:04.045886 138547772659520 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.317043s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.317031) I0420 22:08:04.046707 138418371446528 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 22:08:04.046866 138547772659520 async_checkpointer.py:561] Finished blocking save. Time taken: 1.045345s. 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_false/checkpoints/15. I0420 22:08:04.469273 138547772659520 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0420 22:08:04.469638 138421050464000 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 22:08:04.469820 138547772659520 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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722855.8778698, 'wait_for_prev_duration_secs': 27.121620178222656, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776722883.0014937, 'checkpointer_blocking_duration_secs': 1.0454802513122559, 'get_old_steps_start_time': 1776722884.046992, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776722855.875619, 'checkpoint_manager_blocking_duration_secs': 28.594165563583374} I0420 22:08:04.469946 138547772659520 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0420 22:08:04.470779 138547772659520 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.482, Tokens/s/device: 13939.844, total_weights: 65536, loss: 5.858 I0420 22:08:04.478180 138547772659520 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.795, Tokens/s/device: 13836.342, total_weights: 65536, loss: 5.713 I0420 22:08:23.201290 138547772659520 metric_logger.py:185] completed step: 16, seconds: 28.890, TFLOP/s/device: 0.470, Tokens/s/device: 70.889, total_weights: 65536, loss: 5.599 I0420 22:08:23.348978 138547772659520 metric_logger.py:185] completed step: 17, seconds: 18.578, TFLOP/s/device: 0.731, Tokens/s/device: 110.238, total_weights: 65536, loss: 5.511 I0420 22:08:23.496574 138547772659520 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.744, Tokens/s/device: 13376.528, total_weights: 65536, loss: 5.442 I0420 22:08:23.644923 138547772659520 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 22:08:23.645554 138547772659520 checkpointing.py:798] Waited 0.0006430149078369141 seconds for step 19 to finish before starting checkpointing. I0420 22:08:23.647610 138547772659520 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 22:08:28.010379 138421067249408 array_metadata_store.py:203] [process=7][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_false/checkpoints/15/items/array_metadatas/process_7 I0420 22:08:33.328598 138418371446528 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 53.374 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 29.594923734664917 s) (per-host) I0420 22:08:33.328739 138418371446528 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.281919s. I0420 22:08:39.542712 138418371446528 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 35.495877s. I0420 22:08:39.543004 138421050464000 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 22:08:39.543128 138421050464000 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 22:08:39.543176 138421050464000 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0420 22:08:39.544552 138421050464000 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0420 22:08:39.544728 138547772659520 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0420 22:08:39.544850 138547772659520 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.897252 seconds. If this number is high, consider checkpointing less frequently. I0420 22:08:39.546388 138547772659520 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 19 I0420 22:08:39.549969 138547772659520 async_checkpointer.py:452] [process=7] 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_false/checkpoints/19. I0420 22:08:40.508150 138547772659520 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 22:08:40.508259 138547772659520 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 22:08:40.552746 138547772659520 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046654s I0420 22:08:40.552906 138547772659520 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.172 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29827141761779785 s) (per-host) I0420 22:08:40.552958 138547772659520 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.298333s (batch_requests_ready=0.247145s, total_serialization_initiated=0.051121s, others=0.000067s) I0420 22:08:40.553051 138547772659520 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.302359s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.302347) I0420 22:08:40.553886 138424279348992 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 22:08:40.554040 138547772659520 async_checkpointer.py:561] Finished blocking save. Time taken: 1.007585s. 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_false/checkpoints/19. I0420 22:08:40.701065 138547772659520 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 22:08:40.701432 138421050464000 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 22:08:40.701598 138547772659520 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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722903.6475835, 'wait_for_prev_duration_secs': 15.897252082824707, 'time_between_consecutive_saves_sec': 20.648361921310425, 'checkpointer_blocking_start_time': 1776722919.546426, 'checkpointer_blocking_duration_secs': 1.007720708847046, 'get_old_steps_start_time': 1776722920.5541594, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1776722903.6457865, 'checkpoint_manager_blocking_duration_secs': 17.05577778816223} I0420 22:08:40.701719 138547772659520 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 22:08:40.701761 138547772659520 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 22:08:45.330739 138421067249408 array_metadata_store.py:203] [process=7][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_false/checkpoints/19/items/array_metadatas/process_7 I0420 22:08:50.750034 138424279348992 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 150.505 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.495345115661621 s) (per-host) I0420 22:08:50.750146 138424279348992 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.196151s. I0420 22:08:56.811357 138424279348992 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 16.257346s. I0420 22:08:56.811635 138421050464000 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 22:08:56.811780 138421050464000 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 22:08:56.811827 138421050464000 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 22:08:56.813156 138421050464000 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 22:08:56.813349 138547772659520 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0420 22:08:56.813527 138547772659520 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 22:08:56.814301 138547772659520 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.273, Tokens/s/device: 13908.319, 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 22:09:06 UTC 2026 EXIT_CODE=0