XPK Start: Sat Apr 25 16:35:41 UTC 2026 2026-04-25 16:35:45.654816: 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 00:00:1777134945.667579 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777134945.671362 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777134945.682888 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134945.682907 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134945.682909 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134945.682911 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:36:04.855057: 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) I0425 16:36:05.384624 137997029226304 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:36:14,425:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:36:14.425895 137997029226304 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:36:14,428:jax._src.distributed:157: Connecting to JAX distributed service on mt-01-base-b9enw-slice-job-0-0.mt-01-base-b9enw:8482 I0425 16:36:14.428224 137997029226304 distributed.py:157] Connecting to JAX distributed service on mt-01-base-b9enw-slice-job-0-0.mt-01-base-b9enw:8482 I0425 16:36:15.595835 137997029226304 max_utils.py:284] Jax distributed system initialized! I0425 16:36:22.343703 137997029226304 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:36:22.343810 137997029226304 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:36:22.343851 137997029226304 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 I0425 16:36:22.343885 137997029226304 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:36:22.976947 137997029226304 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:36:23.092388 137997029226304 checkpointing.py:688] Setting up checkpoint logger... I0425 16:36:23.092512 137997029226304 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 16:36:23.092559 137997029226304 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:36:23.092769 137997029226304 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 0x7d8170bbaa50>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:36:26.792768 137997029226304 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:36:26.793024 137997029226304 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 0x7d6e6cf119d0>}, handler_registry=None I0425 16:36:26.793260 137997029226304 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d6e6cf119d0>` 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`. I0425 16:36:26.793319 137997029226304 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d6e6cf12600>` 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`. I0425 16:36:26.793357 137997029226304 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 0x7d6e6cf119d0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d6e6cf119d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d6e6cf12600>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d6e6cf12600>}). I0425 16:36:26.793687 137997029226304 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:36:26.793767 137997029226304 async_checkpointer.py:177] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d6e6c5f8f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:36:28.073385 137997029226304 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints I0425 16:36:28.098517 137997029226304 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=10), 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d6e6cf15c40> I0425 16:36:28.098655 137997029226304 checkpointing.py:302] Checkpoint manager created! I0425 16:36:29.032841 137997029226304 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:36:29.032975 137997029226304 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 16:36:30.458066 137997029226304 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 16:36:30.458190 137997029226304 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 16:36:30.458247 137997029226304 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 16:36:30.458292 137997029226304 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 16:36:30.458330 137997029226304 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 16:36:30.458366 137997029226304 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 16:36:30.458402 137997029226304 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:36:30.458439 137997029226304 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 16:36:30.458475 137997029226304 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:36:30.458513 137997029226304 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:36:30.458549 137997029226304 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 16:36:30.458584 137997029226304 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 16:36:30.851596 137997029226304 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:36:30.851759 137997029226304 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:36:30.857193 137997029226304 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:36:30.857291 137997029226304 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:36:31.012146 137997029226304 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:36:31.012284 137997029226304 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:36:31.167859 137997029226304 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:36:31.168061 137997029226304 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:36:31.330206 137997029226304 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:36:31.330346 137997029226304 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:36:31.481532 137997029226304 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:36:31.481662 137997029226304 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:36:31.634399 137997029226304 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:36:31.634541 137997029226304 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:36:31.800089 137997029226304 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:36:31.800235 137997029226304 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:36:31.822818 137997029226304 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:36:31.822917 137997029226304 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 16:36:42.871382 137997029226304 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. I0425 16:36:42.873785 137997029226304 metric_logger.py:289] number parameters: 1.104 billion I0425 16:36:54.096406 137997029226304 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 16:36:54.246652 137997029226304 checkpointing.py:798] Waited 0.15023374557495117 seconds for step 0 to finish before starting checkpointing. I0425 16:36:54.249075 137997029226304 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:36:54.250978 137997029226304 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 0 I0425 16:36:54.253676 137997029226304 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints/0. I0425 16:36:54.892554 137997029226304 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:36:54.893467 137997029226304 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 16:36:54.893525 137997029226304 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 I0425 16:36:55.140269 137997029226304 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.248841s I0425 16:36:55.140447 137997029226304 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.777 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5554471015930176 s) (per-host) I0425 16:36:55.140507 137997029226304 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.555518s (batch_requests_ready=0.300152s, total_serialization_initiated=0.255287s, others=0.000080s) I0425 16:36:55.140608 137997029226304 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.559901s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.559881) I0425 16:36:55.141542 137870897899264 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0425 16:36:55.141708 137997029226304 async_checkpointer.py:561] Finished blocking save. Time taken: 0.890659s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints/0. I0425 16:36:55.153339 137997029226304 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:36:55.153591 137869834307328 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:36:55.153719 137997029226304 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777135014.249057, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777135014.2510166, 'checkpointer_blocking_duration_secs': 0.8907978534698486, 'get_old_steps_start_time': 1777135015.141836, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1777135014.2470808, 'checkpoint_manager_blocking_duration_secs': 0.9066016674041748} I0425 16:36:55.153825 137997029226304 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 16:36:55.153901 137997029226304 max_utils.py:750] Memstats: After params initialized: I0425 16:36:55.153965 137997029226304 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0425 16:36:55.153999 137997029226304 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0425 16:36:55.154026 137997029226304 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0425 16:36:55.154053 137997029226304 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0425 16:36:55.424069 137997029226304 metric_logger.py:185] completed step: 0, seconds: 11.222, TFLOP/s/device: 1.211, Tokens/s/device: 182.506, total_weights: 65536, loss: 10.874 I0425 16:36:55.612549 137997029226304 metric_logger.py:185] completed step: 1, seconds: 1.326, TFLOP/s/device: 10.247, Tokens/s/device: 1544.540, total_weights: 65536, loss: 10.874 I0425 16:36:56.047523 137997029226304 metric_logger.py:185] completed step: 2, seconds: 0.043, TFLOP/s/device: 316.768, Tokens/s/device: 47746.719, total_weights: 65536, loss: 10.262 I0425 16:36:56.195103 137997029226304 metric_logger.py:185] completed step: 3, seconds: 0.426, TFLOP/s/device: 31.918, Tokens/s/device: 4810.990, total_weights: 65536, loss: 9.732 I0425 16:36:56.490541 137997029226304 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.329, Tokens/s/device: 12560.179, total_weights: 65536, loss: 9.274 I0425 16:36:56.497947 137997029226304 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.370, Tokens/s/device: 13923.070, total_weights: 65536, loss: 8.888 I0425 16:37:15.310581 137997029226304 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.923, Tokens/s/device: 6921.959, total_weights: 65536, loss: 8.589 I0425 16:37:15.457853 137997029226304 metric_logger.py:185] completed step: 7, seconds: 18.667, TFLOP/s/device: 0.728, Tokens/s/device: 109.713, total_weights: 65536, loss: 8.380 I0425 16:37:15.605596 137997029226304 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.588, Tokens/s/device: 13352.980, total_weights: 65536, loss: 8.251 I0425 16:37:15.753847 137997029226304 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 16:37:15.754456 137997029226304 checkpointing.py:798] Waited 0.000621795654296875 seconds for step 9 to finish before starting checkpointing. I0425 16:37:15.756397 137997029226304 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:37:16.446897 2644 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:37:18.355839 137870906291968 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints/0/items/array_metadatas/process_3 I0425 16:37:50.147162 137870897899264 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 28.429 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 55.5621235370636 s) (per-host) I0425 16:37:50.147280 137870897899264 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.005627s. I0425 16:37:59.547304 137870897899264 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 64.405634s. I0425 16:37:59.547592 137869834307328 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:37:59.547709 137869834307328 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:37:59.547759 137869834307328 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:37:59.549390 137869834307328 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 16:37:59.549575 137997029226304 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0425 16:37:59.549705 137997029226304 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.793313 seconds. If this number is high, consider checkpointing less frequently. I0425 16:37:59.551626 137997029226304 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 9 I0425 16:37:59.555024 137997029226304 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints/9. I0425 16:38:00.086439 137997029226304 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 16:38:00.086593 137997029226304 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 I0425 16:38:00.129289 137997029226304 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045087s I0425 16:38:00.129462 137997029226304 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.183 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2976243495941162 s) (per-host) I0425 16:38:00.129513 137997029226304 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.297686s (batch_requests_ready=0.247777s, total_serialization_initiated=0.049839s, others=0.000070s) I0425 16:38:00.129608 137997029226304 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.301561s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.301544) I0425 16:38:00.130460 137870346807040 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0425 16:38:00.130592 137997029226304 async_checkpointer.py:561] Finished blocking save. Time taken: 0.578894s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints/9. I0425 16:38:00.148184 137997029226304 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:38:00.148448 137869834307328 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:38:00.148583 137997029226304 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777135035.7563682, 'wait_for_prev_duration_secs': 43.7933132648468, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777135079.551666, 'checkpointer_blocking_duration_secs': 0.5791118144989014, 'get_old_steps_start_time': 1777135080.1307976, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1777135035.7546837, 'checkpoint_manager_blocking_duration_secs': 44.39386439323425} I0425 16:38:00.148699 137997029226304 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 16:38:00.148745 137997029226304 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:38:05.036582 137870906291968 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_01_base/checkpoints/9/items/array_metadatas/process_3 I0425 16:38:42.180299 137870346807040 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 37.300 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.34843587875366 s) (per-host) I0425 16:38:42.180403 137870346807040 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.049705s. I0425 16:38:49.628022 137870346807040 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 49.497306s. I0425 16:38:49.628307 137869834307328 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:38:49.628438 137869834307328 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:38:49.628484 137869834307328 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:38:49.630908 137869834307328 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 16:38:49.631061 137997029226304 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0425 16:38:49.631175 137997029226304 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:38:49.631763 137997029226304 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.639, Tokens/s/device: 13963.509, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 16:39:02 UTC 2026 EXIT_CODE=0