XPK Start: Sun Apr 19 06:01:05 UTC 2026 2026-04-19 06:01:09.616070: 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 00:00:1776578469.628610 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776578469.632300 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776578469.643497 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776578469.643516 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776578469.643518 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776578469.643521 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 06:01:28.824117: 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) I0419 06:01:29.359551 139552266962752 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 06:01:38,401:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 06:01:38.401301 139552266962752 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 06:01:38,403:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-3o1b1-slice-job-0-0.mt-15-ocdbt-true-rs-3o1b1:8482 I0419 06:01:38.403602 139552266962752 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-3o1b1-slice-job-0-0.mt-15-ocdbt-true-rs-3o1b1:8482 I0419 06:01:40.293313 139552266962752 max_utils.py:284] Jax distributed system initialized! I0419 06:01:47.381959 139552266962752 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 06:01:47.382069 139552266962752 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 06:01:47.382111 139552266962752 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 I0419 06:01:47.382144 139552266962752 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 06:01:48.007344 139552266962752 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 06:01:48.122122 139552266962752 checkpointing.py:688] Setting up checkpoint logger... I0419 06:01:48.122258 139552266962752 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 06:01:48.122304 139552266962752 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 06:01:48.122508 139552266962752 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 0x7eeb84775190>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 06:01:49.831289 139552266962752 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 06:01:49.831642 139552266962752 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 0x7eea45ff67b0>}, handler_registry=None I0419 06:01:49.831883 139552266962752 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eea45ff67b0>` 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`. I0419 06:01:49.831933 139552266962752 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ed6f4368d70>` 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`. I0419 06:01:49.831969 139552266962752 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 0x7eea45ff67b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eea45ff67b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ed6f4368d70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ed6f4368d70>}). I0419 06:01:49.832317 139552266962752 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 06:01:49.832395 139552266962752 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ed6f42c4f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 06:01:52.273219 139431944120064 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776578288124266416, 'commit_timestamp_nsecs': 1776578337603388129, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/5/_CHECKPOINT_METADATA I0419 06:01:52.940520 139431944120064 checkpoint_manager.py:1877] Missing metrics for step 5 I0419 06:01:54.225786 139552266962752 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints I0419 06:01:54.542279 139552266962752 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ed6f4369c40> I0419 06:01:54.542453 139552266962752 checkpointing.py:302] Checkpoint manager created! I0419 06:01:55.470805 139552266962752 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 06:01:55.470927 139552266962752 checkpointing.py:582] restoring from this run's directory step 9 I0419 06:01:57.236887 139552266962752 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/9. I0419 06:01:59.209660 1416 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 06:02:06.700774 139552266962752 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 210.798 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.496649742126465 s) (per-host) I0419 06:02:06.709214 139552266962752 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 189.802 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 8.338207244873047 s) (per-host) I0419 06:02:06.711369 139552266962752 checkpointer.py:318] Finished restoring checkpoint in 10.56 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/9. I0419 06:02:06.711460 139552266962752 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776578516.149471, 'checkpointer_duration_secs': 10.56196141242981, 'checkpoint_manager_start_time': 1776578515.4715772, 'checkpoint_manager_duration_secs': 11.239856958389282} fsdp: 32 I0419 06:02:06.792680 139552266962752 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 06:02:06.792796 139552266962752 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:02:06.792850 139552266962752 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:02:06.792894 139552266962752 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 06:02:06.792931 139552266962752 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:02:06.792966 139552266962752 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:02:06.793003 139552266962752 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:02:06.793039 139552266962752 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 06:02:06.793077 139552266962752 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:02:06.793113 139552266962752 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:02:06.793147 139552266962752 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 06:02:06.793180 139552266962752 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 06:02:07.194015 139552266962752 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:02:07.194163 139552266962752 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:02:07.199683 139552266962752 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:02:07.199776 139552266962752 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:02:07.355025 139552266962752 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:02:07.355159 139552266962752 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:02:07.509011 139552266962752 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:02:07.509145 139552266962752 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:02:07.671764 139552266962752 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:02:07.671901 139552266962752 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:02:07.824933 139552266962752 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:02:07.825069 139552266962752 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:02:07.977052 139552266962752 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:02:07.977193 139552266962752 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:02:08.138950 139552266962752 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 06:02:08.139084 139552266962752 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:02:08.161956 139552266962752 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 06:02:08.162053 139552266962752 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 06:02:19.397225 139552266962752 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. I0419 06:02:19.399622 139552266962752 metric_logger.py:289] number parameters: 1.104 billion I0419 06:02:30.593558 139552266962752 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 06:02:30.744062 139552266962752 checkpointing.py:798] Waited 0.15048980712890625 seconds for step 10 to finish before starting checkpointing. I0419 06:02:30.746461 139552266962752 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:02:30.748030 139552266962752 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0419 06:02:30.750716 139552266962752 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/10. I0419 06:02:32.197315 139552266962752 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 06:02:32.198485 139552266962752 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:02:32.198548 139552266962752 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 I0419 06:02:32.454971 139552266962752 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.258797s I0419 06:02:32.455138 139552266962752 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.629 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.9472246170043945 s) (per-host) I0419 06:02:32.455191 139552266962752 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.947287s (batch_requests_ready=0.683384s, total_serialization_initiated=0.263833s, others=0.000070s) I0419 06:02:32.455313 139552266962752 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.951330s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.951311) I0419 06:02:32.456196 139430694196992 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:02:32.456380 139552266962752 async_checkpointer.py:561] Finished blocking save. Time taken: 1.708280s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/10. I0419 06:02:32.864449 139552266962752 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:02:32.864833 139426959730432 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:02:32.865000 139552266962752 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776578550.746444, 'wait_for_prev_duration_secs': 7.43865966796875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776578550.748068, 'checkpointer_blocking_duration_secs': 1.708421230316162, 'get_old_steps_start_time': 1776578552.456506, 'get_old_steps_duration_secs': 3.266334533691406e-05, 'checkpoint_manager_blocking_start_time': 1776578550.744467, 'checkpoint_manager_blocking_duration_secs': 2.1204946041107178} I0419 06:02:32.865116 139552266962752 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 06:02:32.865190 139552266962752 max_utils.py:750] Memstats: After params initialized: I0419 06:02:32.865257 139552266962752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 06:02:32.865293 139552266962752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 06:02:32.865322 139552266962752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 06:02:32.865346 139552266962752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 06:02:33.135702 139552266962752 metric_logger.py:185] completed step: 10, seconds: 11.193, TFLOP/s/device: 1.214, Tokens/s/device: 182.973, total_weights: 65536, loss: 6.827 I0419 06:02:33.308542 139552266962752 metric_logger.py:185] completed step: 11, seconds: 2.540, TFLOP/s/device: 5.348, Tokens/s/device: 806.176, total_weights: 65536, loss: 6.522 I0419 06:02:33.456417 139552266962752 metric_logger.py:185] completed step: 12, seconds: 0.017, TFLOP/s/device: 783.436, Tokens/s/device: 118087.989, total_weights: 65536, loss: 6.260 I0419 06:02:33.603997 139552266962752 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 83.020, Tokens/s/device: 12513.672, total_weights: 65536, loss: 6.039 I0419 06:02:33.610422 139552266962752 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 06:02:33.899164 139552266962752 checkpointing.py:798] Waited 0.2887108325958252 seconds for step 15 to finish before starting checkpointing. I0419 06:02:33.901769 139552266962752 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:02:40.256181 139427520272128 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6 I0419 06:03:16.131923 139430694196992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 35.398 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.62397027015686 s) (per-host) I0419 06:03:16.132047 139430694196992 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.675722s. I0419 06:03:23.954002 139430694196992 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.497663s. I0419 06:03:23.954316 139426959730432 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:03:23.954435 139426959730432 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:03:23.954482 139426959730432 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:03:23.955962 139426959730432 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 06:03:23.956131 139552266962752 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. W0419 06:03:23.956282 139552266962752 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.054516 seconds. If this number is high, consider checkpointing less frequently. I0419 06:03:23.958170 139552266962752 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0419 06:03:23.962209 139552266962752 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/15. I0419 06:03:25.358461 139552266962752 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:03:25.358559 139552266962752 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 I0419 06:03:25.407847 139552266962752 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051601s I0419 06:03:25.408007 139552266962752 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.073 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7441329956054688 s) (per-host) I0419 06:03:25.408060 139552266962752 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.744195s (batch_requests_ready=0.688025s, total_serialization_initiated=0.056102s, others=0.000068s) I0419 06:03:25.408154 139552266962752 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.749030s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.749017) I0419 06:03:25.409042 139429100369664 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:03:25.409200 139552266962752 async_checkpointer.py:561] Finished blocking save. Time taken: 1.450957s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/15. I0419 06:03:25.432836 139552266962752 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:03:25.433112 139426959730432 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:03:25.433274 139552266962752 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776578553.901741, 'wait_for_prev_duration_secs': 50.05451560020447, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776578603.9582105, 'checkpointer_blocking_duration_secs': 1.451111078262329, 'get_old_steps_start_time': 1776578605.4093409, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776578553.8995616, 'checkpoint_manager_blocking_duration_secs': 51.53365898132324} I0419 06:03:25.433390 139552266962752 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 06:03:25.434196 139552266962752 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.192, Tokens/s/device: 13896.240, total_weights: 65536, loss: 5.858 I0419 06:03:25.441975 139552266962752 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.932, Tokens/s/device: 13857.032, total_weights: 65536, loss: 5.713 I0419 06:03:44.316976 139552266962752 metric_logger.py:185] completed step: 16, seconds: 51.830, TFLOP/s/device: 0.262, Tokens/s/device: 39.514, total_weights: 65536, loss: 5.599 I0419 06:03:44.464504 139552266962752 metric_logger.py:185] completed step: 17, seconds: 18.730, TFLOP/s/device: 0.725, Tokens/s/device: 109.344, total_weights: 65536, loss: 5.511 I0419 06:03:44.612007 139552266962752 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.621, Tokens/s/device: 13357.945, total_weights: 65536, loss: 5.442 I0419 06:03:44.760354 139552266962752 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 06:03:44.761010 139552266962752 checkpointing.py:798] Waited 0.0006673336029052734 seconds for step 19 to finish before starting checkpointing. I0419 06:03:44.763583 139552266962752 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:03:48.100739 139425887536896 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6 I0419 06:04:24.117420 139429100369664 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.569 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.45350217819214 s) (per-host) I0419 06:04:24.117535 139429100369664 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 58.708381s. I0419 06:04:31.834831 139429100369664 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 66.425665s. I0419 06:04:31.835134 139426959730432 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:04:31.835292 139426959730432 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:04:31.835338 139426959730432 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:04:31.838268 139426959730432 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 06:04:31.838400 139552266962752 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. W0419 06:04:31.838461 139552266962752 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.074891 seconds. If this number is high, consider checkpointing less frequently. I0419 06:04:31.840071 139552266962752 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0419 06:04:31.843946 139552266962752 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/19. I0419 06:04:32.826780 139552266962752 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:04:32.826878 139552266962752 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 I0419 06:04:32.878720 139552266962752 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.053891s I0419 06:04:32.878887 139552266962752 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.885 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3157978057861328 s) (per-host) I0419 06:04:32.878936 139552266962752 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.315857s (batch_requests_ready=0.257375s, total_serialization_initiated=0.058417s, others=0.000065s) I0419 06:04:32.879035 139552266962752 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.319885s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.319873) I0419 06:04:32.879859 139430727767808 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:04:32.880011 139552266962752 async_checkpointer.py:561] Finished blocking save. Time taken: 1.039872s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/19. I0419 06:04:33.300309 139552266962752 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:04:33.300676 139426959730432 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:04:33.300823 139552266962752 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776578624.763558, 'wait_for_prev_duration_secs': 47.07489061355591, 'time_between_consecutive_saves_sec': 20.80755591392517, 'checkpointer_blocking_start_time': 1776578671.8401096, 'checkpointer_blocking_duration_secs': 1.0400090217590332, 'get_old_steps_start_time': 1776578672.880131, 'get_old_steps_duration_secs': 2.0742416381835938e-05, 'checkpoint_manager_blocking_start_time': 1776578624.7612538, 'checkpoint_manager_blocking_duration_secs': 48.539536237716675} I0419 06:04:33.300995 139552266962752 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 06:04:33.301041 139552266962752 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:04:40.094297 139431902156544 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6 I0419 06:05:16.193400 139430727767808 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.204 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.6302707195282 s) (per-host) I0419 06:05:16.193516 139430727767808 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.313550s. I0419 06:05:22.144602 139430727767808 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.264622s. I0419 06:05:22.144806 139426959730432 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:05:22.144870 139426959730432 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:05:22.144917 139426959730432 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:05:22.147785 139426959730432 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 06:05:22.147941 139552266962752 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. I0419 06:05:22.148111 139552266962752 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:05:22.148748 139552266962752 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.396, Tokens/s/device: 13926.952, 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: Sun Apr 19 06:05:31 UTC 2026 EXIT_CODE=0