XPK Start: Fri Apr 24 17:55:42 UTC 2026 2026-04-24 17:55:46.494547: 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 00:00:1777053346.507510 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777053346.511644 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777053346.523714 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777053346.523735 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777053346.523737 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777053346.523739 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 17:56:05.670653: 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) I0424 17:56:06.192964 139226277926720 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 17:56:15,233:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 17:56:15.233422 139226277926720 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 17:56:15,235:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-rs-qzezq-slice-job-0-0.mt-16-zarr3-false-rs-qzezq:8482 I0424 17:56:15.235742 139226277926720 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-rs-qzezq-slice-job-0-0.mt-16-zarr3-false-rs-qzezq:8482 I0424 17:56:16.373069 139226277926720 max_utils.py:284] Jax distributed system initialized! I0424 17:56:23.237449 139226277926720 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 17:56:23.237554 139226277926720 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 17:56:23.237598 139226277926720 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 I0424 17:56:23.237633 139226277926720 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 17:56:23.868796 139226277926720 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 17:56:23.981770 139226277926720 checkpointing.py:688] Setting up checkpoint logger... I0424 17:56:23.981886 139226277926720 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0424 17:56:23.981932 139226277926720 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 17:56:23.982179 139226277926720 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e9f96778fe0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 17:56:25.322084 139226277926720 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 17:56:25.322450 139226277926720 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 0x7e8afc350ad0>}, handler_registry=None I0424 17:56:25.322688 139226277926720 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e8afc350ad0>` 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`. I0424 17:56:25.322735 139226277926720 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e8afc45fe60>` 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`. I0424 17:56:25.322772 139226277926720 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 0x7e8afc350ad0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e8afc350ad0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e8afc45fe60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e8afc45fe60>}). I0424 17:56:25.323126 139226277926720 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 17:56:25.323206 139226277926720 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e8afc400f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 17:56:26.843657 139106029905664 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777053089901210310, 'commit_timestamp_nsecs': 1777053136979428976, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0424 17:56:27.103240 139106029905664 checkpoint_manager.py:1877] Missing metrics for step 0 I0424 17:56:27.450624 139226277926720 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints I0424 17:56:27.513965 139226277926720 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e8afc48e330> I0424 17:56:27.514152 139226277926720 checkpointing.py:302] Checkpoint manager created! I0424 17:56:28.449693 139226277926720 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 17:56:28.449811 139226277926720 checkpointing.py:582] restoring from this run's directory step 9 I0424 17:56:28.953088 139226277926720 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/9. I0424 17:56:30.513042 1419 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 17:56:38.328659 139226277926720 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 202.047 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.821343660354614 s) (per-host) I0424 17:56:38.337883 139226277926720 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 191.929 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 8.245779037475586 s) (per-host) I0424 17:56:39.384404 139226277926720 checkpointer.py:318] Finished restoring checkpoint in 10.68 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/9. I0424 17:56:39.384571 139226277926720 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1777053388.7063801, 'checkpointer_duration_secs': 10.678161144256592, 'checkpoint_manager_start_time': 1777053388.4504266, 'checkpoint_manager_duration_secs': 10.93411660194397} fsdp: 32 I0424 17:56:39.464701 139226277926720 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 17:56:39.464818 139226277926720 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:56:39.464870 139226277926720 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:56:39.464912 139226277926720 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 17:56:39.464950 139226277926720 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:56:39.464984 139226277926720 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:56:39.465020 139226277926720 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:56:39.465057 139226277926720 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 17:56:39.465105 139226277926720 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:56:39.465142 139226277926720 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:56:39.465175 139226277926720 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 17:56:39.465208 139226277926720 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 17:56:39.861015 139226277926720 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:56:39.861172 139226277926720 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:56:39.866519 139226277926720 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:56:39.866612 139226277926720 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:56:40.013241 139226277926720 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:56:40.013370 139226277926720 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:56:40.151144 139226277926720 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:56:40.151278 139226277926720 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:56:40.298684 139226277926720 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:56:40.298820 139226277926720 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:56:40.436880 139226277926720 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:56:40.437006 139226277926720 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:56:40.574939 139226277926720 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:56:40.575066 139226277926720 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:56:40.724979 139226277926720 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 17:56:40.725148 139226277926720 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:56:40.747957 139226277926720 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 17:56:40.748052 139226277926720 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 17:56:51.967489 139226277926720 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. I0424 17:56:51.969918 139226277926720 metric_logger.py:289] number parameters: 1.104 billion I0424 17:57:02.930073 139226277926720 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0424 17:57:03.081791 139226277926720 checkpointing.py:798] Waited 0.15170574188232422 seconds for step 10 to finish before starting checkpointing. I0424 17:57:03.083981 139226277926720 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:57:03.086004 139226277926720 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0424 17:57:03.088766 139226277926720 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/10. I0424 17:57:03.704186 139226277926720 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 17:57:03.705242 139226277926720 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:57:03.705308 139226277926720 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 I0424 17:57:03.951860 139226277926720 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.248868s I0424 17:57:03.952031 139226277926720 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.009 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5127377510070801 s) (per-host) I0424 17:57:03.952081 139226277926720 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.512799s (batch_requests_ready=0.259157s, total_serialization_initiated=0.253573s, others=0.000069s) I0424 17:57:03.952189 139226277926720 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.516887s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.516868) I0424 17:57:03.953056 139104394147584 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:57:03.953233 139226277926720 async_checkpointer.py:561] Finished blocking save. Time taken: 0.867159s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/10. I0424 17:57:04.096420 139226277926720 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:57:04.096800 139100350830336 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:57:04.096959 139226277926720 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777053423.0839624, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777053423.0860436, 'checkpointer_blocking_duration_secs': 0.8673031330108643, 'get_old_steps_start_time': 1777053423.9533668, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1777053423.0821755, 'checkpoint_manager_blocking_duration_secs': 1.014744758605957} I0424 17:57:04.097074 139226277926720 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0424 17:57:04.097168 139226277926720 max_utils.py:750] Memstats: After params initialized: I0424 17:57:04.097228 139226277926720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 17:57:04.097262 139226277926720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 17:57:04.097295 139226277926720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 17:57:04.097318 139226277926720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 17:57:04.375307 139226277926720 metric_logger.py:185] completed step: 10, seconds: 10.959, TFLOP/s/device: 1.240, Tokens/s/device: 186.877, total_weights: 65536, loss: 6.827 I0424 17:57:04.559138 139226277926720 metric_logger.py:185] completed step: 11, seconds: 1.444, TFLOP/s/device: 9.413, Tokens/s/device: 1418.767, total_weights: 65536, loss: 6.522 I0424 17:57:04.706787 139226277926720 metric_logger.py:185] completed step: 12, seconds: 0.028, TFLOP/s/device: 478.353, Tokens/s/device: 72102.521, total_weights: 65536, loss: 6.260 I0424 17:57:04.854529 139226277926720 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 83.096, Tokens/s/device: 12525.075, total_weights: 65536, loss: 6.039 I0424 17:57:04.862740 139226277926720 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0424 17:57:05.149411 139226277926720 checkpointing.py:798] Waited 0.28665614128112793 seconds for step 15 to finish before starting checkpointing. I0424 17:57:05.151952 139226277926720 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:57:08.792033 139100367615744 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/10/items/array_metadatas/process_6 I0424 17:57:44.790294 139104394147584 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.200 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.3509624004364 s) (per-host) I0424 17:57:44.790414 139104394147584 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.837236s. I0424 17:57:51.283682 139104394147584 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.330487s. I0424 17:57:51.283987 139100350830336 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:57:51.284135 139100350830336 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:57:51.284186 139100350830336 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:57:51.286810 139100350830336 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0424 17:57:51.286996 139226277926720 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. W0424 17:57:51.287137 139226277926720 checkpoint_manager.py:1458] Waiting for previous save to complete took 46.135190 seconds. If this number is high, consider checkpointing less frequently. I0424 17:57:51.288902 139226277926720 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0424 17:57:51.293003 139226277926720 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/15. I0424 17:57:51.880741 139226277926720 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:57:51.880844 139226277926720 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 I0424 17:57:51.926156 139226277926720 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047570s I0424 17:57:51.926322 139226277926720 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.966 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3106343746185303 s) (per-host) I0424 17:57:51.926371 139226277926720 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.310692s (batch_requests_ready=0.258498s, total_serialization_initiated=0.052128s, others=0.000065s) I0424 17:57:51.926459 139226277926720 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.315415s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.315403) I0424 17:57:51.927329 139100367615744 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:57:51.927458 139226277926720 async_checkpointer.py:561] Finished blocking save. Time taken: 0.638484s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/15. I0424 17:57:51.938124 139226277926720 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:57:51.938417 139100350830336 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:57:51.938576 139226277926720 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777053425.1519222, 'wait_for_prev_duration_secs': 46.13518977165222, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777053471.2889419, 'checkpointer_blocking_duration_secs': 0.6387279033660889, 'get_old_steps_start_time': 1777053471.9276903, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1777053425.1497838, 'checkpoint_manager_blocking_duration_secs': 46.788758993148804} I0424 17:57:51.938682 139226277926720 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0424 17:57:51.939465 139226277926720 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.254, Tokens/s/device: 13905.581, total_weights: 65536, loss: 5.858 I0424 17:57:51.946913 139226277926720 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 90.926, Tokens/s/device: 13705.414, total_weights: 65536, loss: 5.713 I0424 17:58:10.091845 139226277926720 metric_logger.py:185] completed step: 16, seconds: 47.083, TFLOP/s/device: 0.289, Tokens/s/device: 43.498, total_weights: 65536, loss: 5.599 I0424 17:58:10.239045 139226277926720 metric_logger.py:185] completed step: 17, seconds: 18.000, TFLOP/s/device: 0.755, Tokens/s/device: 113.780, total_weights: 65536, loss: 5.511 I0424 17:58:10.386487 139226277926720 metric_logger.py:185] completed step: 18, seconds: 0.154, TFLOP/s/device: 88.465, Tokens/s/device: 13334.462, total_weights: 65536, loss: 5.442 I0424 17:58:10.534991 139226277926720 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0424 17:58:10.535647 139226277926720 checkpointing.py:798] Waited 0.0006687641143798828 seconds for step 19 to finish before starting checkpointing. I0424 17:58:10.537657 139226277926720 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:58:14.502878 139101122565888 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/15/items/array_metadatas/process_6 I0424 17:58:52.222276 139100367615744 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.063 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.606547832489014 s) (per-host) I0424 17:58:52.222410 139100367615744 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.294871s. I0424 17:58:57.112119 139100367615744 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 65.184544s. I0424 17:58:57.112442 139100350830336 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:58:57.112568 139100350830336 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:58:57.112630 139100350830336 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:58:57.113988 139100350830336 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0424 17:58:57.114209 139226277926720 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. W0424 17:58:57.114330 139226277926720 checkpoint_manager.py:1458] Waiting for previous save to complete took 46.576685 seconds. If this number is high, consider checkpointing less frequently. I0424 17:58:57.118532 139226277926720 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0424 17:58:57.122078 139226277926720 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/19. I0424 17:58:57.652210 139226277926720 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:58:57.652310 139226277926720 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 I0424 17:58:57.696849 139226277926720 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046700s I0424 17:58:57.697016 139226277926720 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.194 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2970132827758789 s) (per-host) I0424 17:58:57.697063 139226277926720 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.297071s (batch_requests_ready=0.245784s, total_serialization_initiated=0.051224s, others=0.000062s) I0424 17:58:57.697167 139226277926720 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.300940s (all_items=0.000010s, per_item={'items': '0.00000954'}, temp_paths=0.300931) I0424 17:58:57.698081 139104394147584 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:58:57.698256 139226277926720 async_checkpointer.py:561] Finished blocking save. Time taken: 0.579652s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/19. I0424 17:58:57.712336 139226277926720 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:58:57.712598 139100350830336 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:58:57.712755 139226277926720 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777053490.5376294, 'wait_for_prev_duration_secs': 46.57668495178223, 'time_between_consecutive_saves_sec': 19.25077795982361, 'checkpointer_blocking_start_time': 1777053537.1185722, 'checkpointer_blocking_duration_secs': 0.5797913074493408, 'get_old_steps_start_time': 1777053537.6983774, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1777053490.5358777, 'checkpoint_manager_blocking_duration_secs': 47.176846742630005} I0424 17:58:57.712862 139226277926720 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0424 17:58:57.712903 139226277926720 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:59:02.491525 139105434318592 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/19/items/array_metadatas/process_6 I0424 17:59:39.651752 139104394147584 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.385 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.251715421676636 s) (per-host) I0424 17:59:39.651857 139104394147584 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.953650s. I0424 17:59:45.512202 139104394147584 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.813979s. I0424 17:59:45.512505 139100350830336 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:59:45.512640 139100350830336 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:59:45.512684 139100350830336 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:59:45.513831 139100350830336 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0424 17:59:45.514005 139226277926720 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. I0424 17:59:45.514181 139226277926720 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:59:45.514829 139226277926720 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.613, Tokens/s/device: 13959.702, 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: Fri Apr 24 17:59:57 UTC 2026 EXIT_CODE=0