XPK Start: Wed Apr 22 18:42:41 UTC 2026 2026-04-22 18:42:45.440648: 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 00:00:1776883365.453647 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776883365.457367 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776883365.468927 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883365.468949 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883365.468952 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883365.468954 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 18:43:04.706570: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0422 18:43:05.236056 140372721342272 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 18:43:14,278:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 18:43:14.278624 140372721342272 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 18:43:14,280:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-rs-1yw3l-slice-job-0-0.mt-15-ocdbt-false-rs-1yw3l:8482 I0422 18:43:14.280896 140372721342272 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-1yw3l-slice-job-0-0.mt-15-ocdbt-false-rs-1yw3l:8482 I0422 18:43:15.264885 140372721342272 max_utils.py:284] Jax distributed system initialized! I0422 18:43:22.440263 140372721342272 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:43:22.440373 140372721342272 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:43:22.440414 140372721342272 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0422 18:43:22.440448 140372721342272 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:43:23.065854 140372721342272 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:43:23.181189 140372721342272 checkpointing.py:688] Setting up checkpoint logger... I0422 18:43:23.181313 140372721342272 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0422 18:43:23.181358 140372721342272 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 18:43:23.181569 140372721342272 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 0x7faa9ad79160>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:43:24.481770 140372721342272 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 18:43:24.482087 140372721342272 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 0x7f95cc515340>}, handler_registry=None I0422 18:43:24.482347 140372721342272 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f95cc515340>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0422 18:43:24.482396 140372721342272 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f95cc59f470>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0422 18:43:24.482433 140372721342272 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 0x7f95cc515340>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f95cc515340>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f95cc59f470>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f95cc59f470>}). I0422 18:43:24.482785 140372721342272 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 18:43:24.482861 140372721342272 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f95cc518f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:43:27.495169 140251653715712 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776883217127337995, 'commit_timestamp_nsecs': 1776883232263516265, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0422 18:43:27.741193 140251653715712 checkpoint_manager.py:1877] Missing metrics for step 0 I0422 18:43:27.937287 140372721342272 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints I0422 18:43:27.939855 140372721342272 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f95cc5d4350> I0422 18:43:27.939970 140372721342272 checkpointing.py:302] Checkpoint manager created! I0422 18:43:28.883136 140372721342272 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:43:28.883254 140372721342272 checkpointing.py:582] restoring from this run's directory step 9 I0422 18:43:29.398549 140372721342272 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/9. I0422 18:43:32.949921 1466 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 18:43:43.044147 140372721342272 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 135.557 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.65763235092163 s) (per-host) I0422 18:43:43.065296 140372721342272 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 130.998 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.081120491027832 s) (per-host) I0422 18:43:43.119953 140372721342272 checkpointer.py:318] Finished restoring checkpoint in 13.97 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/9. I0422 18:43:43.120123 140372721342272 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776883409.1479416, 'checkpointer_duration_secs': 13.972121477127075, 'checkpoint_manager_start_time': 1776883408.8838506, 'checkpoint_manager_duration_secs': 14.236214399337769} fsdp: 32 I0422 18:43:43.200816 140372721342272 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 18:43:43.200933 140372721342272 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:43:43.200986 140372721342272 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:43:43.201031 140372721342272 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 18:43:43.201068 140372721342272 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:43:43.201116 140372721342272 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:43:43.201156 140372721342272 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:43:43.201191 140372721342272 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 18:43:43.201225 140372721342272 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:43:43.201260 140372721342272 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:43:43.201292 140372721342272 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 18:43:43.201330 140372721342272 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 18:43:43.598353 140372721342272 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:43:43.598501 140372721342272 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:43:43.603927 140372721342272 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:43:43.604022 140372721342272 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:43:43.758272 140372721342272 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:43:43.758413 140372721342272 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:43:43.913183 140372721342272 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:43:43.913321 140372721342272 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:43:44.079217 140372721342272 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:43:44.079371 140372721342272 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:43:44.233804 140372721342272 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:43:44.233937 140372721342272 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:43:44.389891 140372721342272 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:43:44.390025 140372721342272 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:43:44.555610 140372721342272 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:43:44.555742 140372721342272 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:43:44.580662 140372721342272 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:43:44.580765 140372721342272 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 18:43:55.388807 140372721342272 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0422 18:43:55.391398 140372721342272 metric_logger.py:289] number parameters: 1.104 billion I0422 18:44:06.578442 140372721342272 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0422 18:44:06.731163 140372721342272 checkpointing.py:798] Waited 0.1526622772216797 seconds for step 10 to finish before starting checkpointing. I0422 18:44:06.733628 140372721342272 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:44:06.735352 140372721342272 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0422 18:44:06.738233 140372721342272 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/10. I0422 18:44:07.328604 140372721342272 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 18:44:07.329686 140372721342272 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:44:07.329744 140372721342272 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0422 18:44:07.578351 140372721342272 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250883s I0422 18:44:07.578525 140372721342272 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.979 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5178227424621582 s) (per-host) I0422 18:44:07.578581 140372721342272 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.517895s (batch_requests_ready=0.262288s, total_serialization_initiated=0.255527s, others=0.000080s) I0422 18:44:07.578678 140372721342272 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.522014s (all_items=0.000020s, per_item={'items': '0.00002027'}, temp_paths=0.521994) I0422 18:44:07.579478 140249472694016 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:44:07.579614 140372721342272 async_checkpointer.py:561] Finished blocking save. Time taken: 0.844191s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/10. I0422 18:44:08.033581 140372721342272 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:44:08.033966 140247811725056 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:44:08.034143 140372721342272 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883446.7336116, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883446.7353983, 'checkpointer_blocking_duration_secs': 0.844367265701294, 'get_old_steps_start_time': 1776883447.579784, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776883446.7315376, 'checkpoint_manager_blocking_duration_secs': 1.3025517463684082} I0422 18:44:08.034258 140372721342272 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0422 18:44:08.034334 140372721342272 max_utils.py:750] Memstats: After params initialized: I0422 18:44:08.034390 140372721342272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 18:44:08.034423 140372721342272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 18:44:08.034450 140372721342272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 18:44:08.034474 140372721342272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 18:44:08.299885 140372721342272 metric_logger.py:185] completed step: 10, seconds: 11.186, TFLOP/s/device: 1.215, Tokens/s/device: 183.086, total_weights: 65536, loss: 6.827 I0422 18:44:08.476497 140372721342272 metric_logger.py:185] completed step: 11, seconds: 1.720, TFLOP/s/device: 7.901, Tokens/s/device: 1190.872, total_weights: 65536, loss: 6.522 I0422 18:44:08.623908 140372721342272 metric_logger.py:185] completed step: 12, seconds: 0.021, TFLOP/s/device: 641.326, Tokens/s/device: 96667.611, total_weights: 65536, loss: 6.260 I0422 18:44:08.771699 140372721342272 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.152, Tokens/s/device: 12533.583, total_weights: 65536, loss: 6.039 I0422 18:44:08.778152 140372721342272 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0422 18:44:09.066594 140372721342272 checkpointing.py:798] Waited 0.28841423988342285 seconds for step 15 to finish before starting checkpointing. I0422 18:44:09.069410 140372721342272 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:44:12.041629 140247828510464 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6 I0422 18:44:18.522806 140249472694016 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 137.811 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.462067127227783 s) (per-host) I0422 18:44:18.522913 140249472694016 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.943229s. I0422 18:44:25.339015 140249472694016 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 17.759316s. I0422 18:44:25.339256 140247811725056 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:44:25.339323 140247811725056 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:44:25.339367 140247811725056 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:44:25.341181 140247811725056 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0422 18:44:25.341372 140372721342272 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0422 18:44:25.341502 140372721342272 checkpoint_manager.py:1458] Waiting for previous save to complete took 16.272096 seconds. If this number is high, consider checkpointing less frequently. I0422 18:44:25.343512 140372721342272 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0422 18:44:25.347604 140372721342272 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/15. I0422 18:44:26.306270 140372721342272 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:44:26.306376 140372721342272 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0422 18:44:26.351164 140372721342272 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047084s I0422 18:44:26.351350 140372721342272 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.147 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29970312118530273 s) (per-host) I0422 18:44:26.351417 140372721342272 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299780s (batch_requests_ready=0.248020s, total_serialization_initiated=0.051673s, others=0.000087s) I0422 18:44:26.351541 140372721342272 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.304507s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.304494) I0422 18:44:26.352473 140247820117760 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:44:26.352646 140372721342272 async_checkpointer.py:561] Finished blocking save. Time taken: 1.009061s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/15. I0422 18:44:26.362953 140372721342272 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:44:26.363212 140247811725056 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:44:26.363348 140372721342272 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883449.0693824, 'wait_for_prev_duration_secs': 16.272095918655396, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883465.3435502, 'checkpointer_blocking_duration_secs': 1.0092127323150635, 'get_old_steps_start_time': 1776883466.352782, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776883449.0669434, 'checkpoint_manager_blocking_duration_secs': 17.29637098312378} I0422 18:44:26.363457 140372721342272 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0422 18:44:26.364258 140372721342272 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.418, Tokens/s/device: 13930.267, total_weights: 65536, loss: 5.858 I0422 18:44:26.371974 140372721342272 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.765, Tokens/s/device: 13831.763, total_weights: 65536, loss: 5.713 I0422 18:44:44.456081 140372721342272 metric_logger.py:185] completed step: 16, seconds: 17.592, TFLOP/s/device: 0.772, Tokens/s/device: 116.415, total_weights: 65536, loss: 5.599 I0422 18:44:44.603366 140372721342272 metric_logger.py:185] completed step: 17, seconds: 17.939, TFLOP/s/device: 0.757, Tokens/s/device: 114.165, total_weights: 65536, loss: 5.511 I0422 18:44:44.750771 140372721342272 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.569, Tokens/s/device: 13350.021, total_weights: 65536, loss: 5.442 I0422 18:44:44.899148 140372721342272 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0422 18:44:44.899786 140372721342272 checkpointing.py:798] Waited 0.0006494522094726562 seconds for step 19 to finish before starting checkpointing. I0422 18:44:44.901653 140372721342272 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:44:49.920035 140247828510464 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6 I0422 18:44:55.127323 140247820117760 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 54.327 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 29.075619220733643 s) (per-host) I0422 18:44:55.127440 140247820117760 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.774858s. I0422 18:45:00.639875 140247820117760 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 34.287278s. I0422 18:45:00.640193 140247811725056 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:45:00.640334 140247811725056 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:45:00.640381 140247811725056 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:45:00.642023 140247811725056 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0422 18:45:00.642222 140372721342272 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0422 18:45:00.642343 140372721342272 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.740702 seconds. If this number is high, consider checkpointing less frequently. I0422 18:45:00.644036 140372721342272 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0422 18:45:00.647547 140372721342272 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/19. I0422 18:45:01.177469 140372721342272 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:45:01.177570 140372721342272 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0422 18:45:01.224278 140372721342272 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049072s I0422 18:45:01.224447 140372721342272 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.145 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.299832820892334 s) (per-host) I0422 18:45:01.224497 140372721342272 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299892s (batch_requests_ready=0.246215s, total_serialization_initiated=0.053613s, others=0.000064s) I0422 18:45:01.224600 140372721342272 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303864s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.303852) I0422 18:45:01.225487 140244056643328 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:45:01.225657 140372721342272 async_checkpointer.py:561] Finished blocking save. Time taken: 0.581551s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/19. I0422 18:45:01.250004 140372721342272 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:45:01.250294 140247811725056 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:45:01.250438 140372721342272 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883484.9016259, 'wait_for_prev_duration_secs': 15.740702390670776, 'time_between_consecutive_saves_sec': 19.56040668487549, 'checkpointer_blocking_start_time': 1776883500.6440752, 'checkpointer_blocking_duration_secs': 0.5816879272460938, 'get_old_steps_start_time': 1776883501.2257764, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1776883484.9000049, 'checkpoint_manager_blocking_duration_secs': 16.350401163101196} I0422 18:45:01.250558 140372721342272 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0422 18:45:01.250601 140372721342272 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:45:05.947633 140247828510464 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6 I0422 18:45:11.543968 140244056643328 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 148.747 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.619316816329956 s) (per-host) I0422 18:45:11.544088 140244056643328 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.318477s. I0422 18:45:16.875478 140244056643328 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 15.649852s. I0422 18:45:16.875687 140247811725056 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:45:16.875741 140247811725056 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:45:16.875782 140247811725056 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:45:16.877952 140247811725056 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0422 18:45:16.878124 140372721342272 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0422 18:45:16.878305 140372721342272 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:45:16.879023 140372721342272 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.488, Tokens/s/device: 13940.792, total_weights: 65536, loss: 5.385 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Wed Apr 22 18:45:27 UTC 2026 EXIT_CODE=0