XPK Start: Fri Apr 24 17:35:50 UTC 2026 2026-04-24 17:35:54.570683: 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 00:00:1777052154.583508 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777052154.587495 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777052154.598743 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052154.598761 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052154.598763 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777052154.598765 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 17:36:13.873112: 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 17:36:14.394326 132175109138240 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 17:36:23,435:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 17:36:23.435502 132175109138240 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 17:36:23,437:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-rs-92ivs-slice-job-0-0.mt-15-ocdbt-false-rs-92ivs:8482 I0424 17:36:23.437912 132175109138240 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-92ivs-slice-job-0-0.mt-15-ocdbt-false-rs-92ivs:8482 I0424 17:36:24.379842 132175109138240 max_utils.py:284] Jax distributed system initialized! I0424 17:36:31.394850 132175109138240 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 17:36:31.394973 132175109138240 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 17:36:31.395016 132175109138240 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 17:36:31.395053 132175109138240 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 17:36:32.016536 132175109138240 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 17:36:32.131659 132175109138240 checkpointing.py:688] Setting up checkpoint logger... I0424 17:36:32.131779 132175109138240 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0424 17:36:32.131825 132175109138240 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 17:36:32.132062 132175109138240 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 0x7835e396d130>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 17:36:33.374404 132175109138240 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 17:36:33.374729 132175109138240 checkpoint_manager.py:709] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7821683aa180>}, handler_registry=None I0424 17:36:33.374981 132175109138240 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7821683aa180>` 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 17:36:33.375033 132175109138240 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7821404cf7a0>` 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 17:36:33.375069 132175109138240 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 0x7821683aa180>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7821683aa180>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7821404cf7a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7821404cf7a0>}). I0424 17:36:33.375405 132175109138240 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 17:36:33.375481 132175109138240 async_checkpointer.py:177] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7821405ccf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 17:36:34.967706 132054582806272 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777052044130559011, 'commit_timestamp_nsecs': 1777052058401503894, '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_15_ocdbt_false/checkpoints/5/_CHECKPOINT_METADATA I0424 17:36:35.221423 132054574413568 checkpoint_manager.py:1877] Missing metrics for step 9 I0424 17:36:35.491220 132175109138240 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_15_ocdbt_false/checkpoints I0424 17:36:35.647493 132175109138240 checkpoint_manager.py:929] [process=3][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_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7821401b6a80> I0424 17:36:35.647665 132175109138240 checkpointing.py:302] Checkpoint manager created! I0424 17:36:36.589082 132175109138240 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 17:36:36.589199 132175109138240 checkpointing.py:582] restoring from this run's directory step 9 I0424 17:36:37.099832 132175109138240 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_15_ocdbt_false/checkpoints/9. I0424 17:36:40.122670 1452 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 17:36:48.563068 132175109138240 jax_array_handlers.py:837] [process=3] /jax/orbax/read/worker/io/requested throughput: 163.354 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 9.673946619033813 s) (per-host) I0424 17:36:48.583957 132175109138240 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/read/gbytes_per_sec: 156.658 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.10227084159851 s) (per-host) I0424 17:36:49.843758 132175109138240 checkpointer.py:318] Finished restoring checkpoint in 13.00 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/9. I0424 17:36:49.843918 132175109138240 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_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1777052196.8438897, 'checkpointer_duration_secs': 12.99999737739563, 'checkpoint_manager_start_time': 1777052196.589801, 'checkpoint_manager_duration_secs': 13.254088163375854} fsdp: 32 I0424 17:36:49.925410 132175109138240 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 17:36:49.925529 132175109138240 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:36:49.925584 132175109138240 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:36:49.925628 132175109138240 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 17:36:49.925667 132175109138240 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:36:49.925700 132175109138240 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:36:49.925745 132175109138240 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:36:49.925783 132175109138240 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 17:36:49.925819 132175109138240 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:36:49.925856 132175109138240 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:36:49.925889 132175109138240 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 17:36:49.925935 132175109138240 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 17:36:50.324791 132175109138240 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:36:50.324955 132175109138240 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:36:50.330477 132175109138240 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:36:50.330572 132175109138240 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:36:50.483577 132175109138240 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:36:50.483717 132175109138240 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:36:50.635683 132175109138240 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:36:50.635818 132175109138240 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:36:50.798327 132175109138240 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 17:36:50.798466 132175109138240 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:36:50.948111 132175109138240 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 17:36:50.948242 132175109138240 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:36:51.096440 132175109138240 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 17:36:51.096569 132175109138240 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:36:51.258732 132175109138240 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 17:36:51.258864 132175109138240 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:36:51.282847 132175109138240 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 17:36:51.282958 132175109138240 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 17:37:02.499589 132175109138240 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 17:37:02.502178 132175109138240 metric_logger.py:289] number parameters: 1.104 billion I0424 17:37:13.825036 132175109138240 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0424 17:37:13.976334 132175109138240 checkpointing.py:798] Waited 0.15128636360168457 seconds for step 10 to finish before starting checkpointing. I0424 17:37:13.978594 132175109138240 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:37:13.980618 132175109138240 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 10 I0424 17:37:13.983276 132175109138240 async_checkpointer.py:452] [process=3] 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_15_ocdbt_false/checkpoints/10. I0424 17:37:14.956208 132175109138240 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 17:37:14.957255 132175109138240 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:37:14.957313 132175109138240 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 17:37:15.209639 132175109138240 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254573s I0424 17:37:15.209815 132175109138240 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.066 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5030503273010254 s) (per-host) I0424 17:37:15.209868 132175109138240 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.503114s (batch_requests_ready=0.243790s, total_serialization_initiated=0.259252s, others=0.000072s) I0424 17:37:15.209980 132175109138240 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.507088s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.507070) I0424 17:37:15.210831 132051772626688 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0424 17:37:15.210973 132175109138240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.230282s. 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_15_ocdbt_false/checkpoints/10. I0424 17:37:15.251271 132175109138240 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:37:15.251593 132049625143040 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:37:15.251747 132175109138240 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052233.9785762, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052233.9806578, 'checkpointer_blocking_duration_secs': 1.2305052280426025, 'get_old_steps_start_time': 1777052235.2111828, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1777052233.976716, 'checkpoint_manager_blocking_duration_secs': 1.2749948501586914} I0424 17:37:15.251857 132175109138240 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0424 17:37:15.251945 132175109138240 max_utils.py:750] Memstats: After params initialized: I0424 17:37:15.251998 132175109138240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0424 17:37:15.252031 132175109138240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0424 17:37:15.252058 132175109138240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0424 17:37:15.252082 132175109138240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0424 17:37:15.524843 132175109138240 metric_logger.py:185] completed step: 10, seconds: 11.322, TFLOP/s/device: 1.200, Tokens/s/device: 180.889, total_weights: 65536, loss: 6.827 I0424 17:37:15.696179 132175109138240 metric_logger.py:185] completed step: 11, seconds: 1.698, TFLOP/s/device: 8.001, Tokens/s/device: 1206.040, total_weights: 65536, loss: 6.522 I0424 17:37:15.843803 132175109138240 metric_logger.py:185] completed step: 12, seconds: 0.016, TFLOP/s/device: 847.395, Tokens/s/device: 127728.577, total_weights: 65536, loss: 6.260 I0424 17:37:15.991169 132175109138240 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.340, Tokens/s/device: 12561.951, total_weights: 65536, loss: 6.039 I0424 17:37:15.997333 132175109138240 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0424 17:37:16.286343 132175109138240 checkpointing.py:798] Waited 0.28898143768310547 seconds for step 15 to finish before starting checkpointing. I0424 17:37:16.288959 132175109138240 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:37:19.637345 132049641928448 array_metadata_store.py:203] [process=3][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_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_3 I0424 17:37:25.528612 132051772626688 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 145.964 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.821813583374023 s) (per-host) I0424 17:37:25.528741 132051772626688 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.317678s. I0424 17:37:31.237590 132051772626688 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 16.026512s. I0424 17:37:31.237889 132049625143040 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:37:31.238023 132049625143040 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:37:31.238074 132049625143040 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:37:31.239661 132049625143040 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0424 17:37:31.239844 132175109138240 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0424 17:37:31.239995 132175109138240 checkpoint_manager.py:1458] Waiting for previous save to complete took 14.951058 seconds. If this number is high, consider checkpointing less frequently. I0424 17:37:31.241913 132175109138240 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 15 I0424 17:37:31.246051 132175109138240 async_checkpointer.py:452] [process=3] 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_15_ocdbt_false/checkpoints/15. I0424 17:37:31.792190 132175109138240 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:37:31.792291 132175109138240 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 17:37:31.836679 132175109138240 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046553s I0424 17:37:31.836856 132175109138240 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.066 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3045022487640381 s) (per-host) I0424 17:37:31.836909 132175109138240 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304565s (batch_requests_ready=0.253373s, total_serialization_initiated=0.051120s, others=0.000071s) I0424 17:37:31.837026 132175109138240 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.309501s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.309485) I0424 17:37:31.837965 132047460873984 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0424 17:37:31.838126 132175109138240 async_checkpointer.py:561] Finished blocking save. Time taken: 0.596126s. 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_15_ocdbt_false/checkpoints/15. I0424 17:37:31.850452 132175109138240 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:37:31.850720 132049625143040 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:37:31.850881 132175109138240 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052236.2889137, 'wait_for_prev_duration_secs': 14.95105791091919, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052251.2419674, 'checkpointer_blocking_duration_secs': 0.5962677001953125, 'get_old_steps_start_time': 1777052251.838255, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1777052236.2867146, 'checkpoint_manager_blocking_duration_secs': 15.56413197517395} I0424 17:37:31.851002 132175109138240 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0424 17:37:31.851761 132175109138240 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.292, Tokens/s/device: 13911.248, total_weights: 65536, loss: 5.858 I0424 17:37:31.859323 132175109138240 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.036, Tokens/s/device: 13872.707, total_weights: 65536, loss: 5.713 I0424 17:37:50.089368 132175109138240 metric_logger.py:185] completed step: 16, seconds: 15.860, TFLOP/s/device: 0.857, Tokens/s/device: 129.127, total_weights: 65536, loss: 5.599 I0424 17:37:50.236798 132175109138240 metric_logger.py:185] completed step: 17, seconds: 18.085, TFLOP/s/device: 0.751, Tokens/s/device: 113.243, total_weights: 65536, loss: 5.511 I0424 17:37:50.384269 132175109138240 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.563, Tokens/s/device: 13349.151, total_weights: 65536, loss: 5.442 I0424 17:37:50.532621 132175109138240 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0424 17:37:50.533247 132175109138240 checkpointing.py:798] Waited 0.0006384849548339844 seconds for step 19 to finish before starting checkpointing. I0424 17:37:50.535325 132175109138240 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:37:54.834460 132049641928448 array_metadata_store.py:203] [process=3][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_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_3 I0424 17:38:00.603864 132047460873984 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 54.335 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 29.07147216796875 s) (per-host) I0424 17:38:00.604002 132047460873984 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.765926s. I0424 17:38:06.091655 132047460873984 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 34.253564s. I0424 17:38:06.091965 132049625143040 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:38:06.092081 132049625143040 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:38:06.092132 132049625143040 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:38:06.094466 132049625143040 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0424 17:38:06.094632 132175109138240 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0424 17:38:06.094753 132175109138240 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.559441 seconds. If this number is high, consider checkpointing less frequently. I0424 17:38:06.096271 132175109138240 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 19 I0424 17:38:06.099628 132175109138240 async_checkpointer.py:452] [process=3] 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_15_ocdbt_false/checkpoints/19. I0424 17:38:06.651147 132175109138240 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:38:06.651245 132175109138240 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 17:38:06.696568 132175109138240 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047420s I0424 17:38:06.696747 132175109138240 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.835 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31902050971984863 s) (per-host) I0424 17:38:06.696800 132175109138240 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.319084s (batch_requests_ready=0.266814s, total_serialization_initiated=0.052201s, others=0.000068s) I0424 17:38:06.696897 132175109138240 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.323166s (all_items=0.000013s, per_item={'items': '0.00001264'}, temp_paths=0.323153) I0424 17:38:06.697774 132047494444800 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0424 17:38:06.697940 132175109138240 async_checkpointer.py:561] Finished blocking save. Time taken: 0.601588s. 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_15_ocdbt_false/checkpoints/19. I0424 17:38:06.708654 132175109138240 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:38:06.708943 132049625143040 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:38:06.709104 132175109138240 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052270.5352955, 'wait_for_prev_duration_secs': 15.559441328048706, 'time_between_consecutive_saves_sec': 19.295592308044434, 'checkpointer_blocking_start_time': 1777052286.0963116, 'checkpointer_blocking_duration_secs': 0.6017405986785889, 'get_old_steps_start_time': 1777052286.6980655, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1777052270.5334818, 'checkpoint_manager_blocking_duration_secs': 16.175589561462402} I0424 17:38:06.709213 132175109138240 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0424 17:38:06.709256 132175109138240 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:38:11.719694 132049641928448 array_metadata_store.py:203] [process=3][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_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_3 I0424 17:38:16.971860 132047494444800 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 149.102 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.59409475326538 s) (per-host) I0424 17:38:16.972012 132047494444800 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.274128s. I0424 17:38:23.400595 132047494444800 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 16.702697s. I0424 17:38:23.400880 132049625143040 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:38:23.401013 132049625143040 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:38:23.401057 132049625143040 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:38:23.402594 132049625143040 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0424 17:38:23.402758 132175109138240 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0424 17:38:23.402946 132175109138240 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:38:23.403617 132175109138240 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.468, Tokens/s/device: 13937.756, 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 17:38:36 UTC 2026 EXIT_CODE=0