XPK Start: Sat Apr 25 17:25:23 UTC 2026 2026-04-25 17:25:27.477917: 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 00:00:1777137927.490644 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777137927.494431 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777137927.506385 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777137927.506403 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777137927.506406 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777137927.506408 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:25:46.835625: 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 17:25:47.370599 140244012873536 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 17:25:56,412:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 17:25:56.412864 140244012873536 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 17:25:56,415:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-xvelr-slice-job-0-0.mt-15-ocdbt-true-sv-xvelr:8482 I0425 17:25:56.415262 140244012873536 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-xvelr-slice-job-0-0.mt-15-ocdbt-true-sv-xvelr:8482 I0425 17:25:57.217498 140244012873536 max_utils.py:284] Jax distributed system initialized! I0425 17:26:04.247863 140244012873536 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 17:26:04.247971 140244012873536 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 17:26:04.248011 140244012873536 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 17:26:04.248047 140244012873536 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 17:26:04.877027 140244012873536 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 17:26:04.991996 140244012873536 checkpointing.py:688] Setting up checkpoint logger... I0425 17:26:04.992135 140244012873536 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 17:26:04.992184 140244012873536 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 17:26:04.992397 140244012873536 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 0x7f8c93b79640>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 17:26:07.884628 140244012873536 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 17:26:07.884868 140244012873536 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 0x7f77d8328c50>}, handler_registry=None I0425 17:26:07.885118 140244012873536 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f77d8328c50>` 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 17:26:07.885168 140244012873536 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f77485f13d0>` 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 17:26:07.885205 140244012873536 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 0x7f77d8328c50>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f77d8328c50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f77485f13d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f77485f13d0>}). I0425 17:26:07.885600 140244012873536 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 17:26:07.885678 140244012873536 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f77486b8f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 17:26:09.010350 140244012873536 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_15_ocdbt_true/checkpoints I0425 17:26:09.012578 140244012873536 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f7748570830> I0425 17:26:09.012717 140244012873536 checkpointing.py:302] Checkpoint manager created! I0425 17:26:09.947953 140244012873536 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 17:26:09.948076 140244012873536 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 17:26:11.363679 140244012873536 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 17:26:11.363799 140244012873536 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:26:11.363854 140244012873536 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:26:11.363900 140244012873536 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 17:26:11.363941 140244012873536 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:26:11.363977 140244012873536 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:26:11.364016 140244012873536 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:26:11.364056 140244012873536 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 17:26:11.364106 140244012873536 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:26:11.364146 140244012873536 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:26:11.364182 140244012873536 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 17:26:11.364217 140244012873536 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 17:26:11.783885 140244012873536 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:26:11.784045 140244012873536 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:26:11.790018 140244012873536 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:26:11.790177 140244012873536 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:26:11.948273 140244012873536 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:26:11.948404 140244012873536 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:26:12.098447 140244012873536 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:26:12.098584 140244012873536 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:26:12.262459 140244012873536 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 17:26:12.262597 140244012873536 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:26:12.416580 140244012873536 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 17:26:12.416714 140244012873536 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:26:12.566066 140244012873536 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 17:26:12.566217 140244012873536 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:26:12.726675 140244012873536 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 17:26:12.726815 140244012873536 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:26:12.750668 140244012873536 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 17:26:12.750767 140244012873536 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 17:26:23.867883 140244012873536 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 17:26:23.870538 140244012873536 metric_logger.py:289] number parameters: 1.104 billion I0425 17:26:34.999250 140244012873536 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 17:26:35.152164 140244012873536 checkpointing.py:798] Waited 0.15287065505981445 seconds for step 0 to finish before starting checkpointing. I0425 17:26:35.154593 140244012873536 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:26:35.156223 140244012873536 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 17:26:35.159076 140244012873536 async_checkpointer.py:452] [process=6] 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_15_ocdbt_true/checkpoints/0. I0425 17:26:35.754542 140244012873536 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 17:26:35.755541 140244012873536 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:26:35.755601 140244012873536 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 17:26:35.998903 140244012873536 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245506s I0425 17:26:35.999077 140244012873536 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.850 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5412094593048096 s) (per-host) I0425 17:26:35.999145 140244012873536 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.541287s (batch_requests_ready=0.289145s, total_serialization_initiated=0.252056s, others=0.000085s) I0425 17:26:35.999241 140244012873536 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.545410s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.545391) I0425 17:26:36.000241 140117670885120 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:26:36.000409 140244012873536 async_checkpointer.py:561] Finished blocking save. Time taken: 0.844115s. 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_15_ocdbt_true/checkpoints/0. I0425 17:26:36.020385 140244012873536 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:26:36.020698 140118174185216 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:26:36.020856 140244012873536 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777137995.1545744, 'wait_for_prev_duration_secs': 6.365776062011719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777137995.156263, 'checkpointer_blocking_duration_secs': 0.8442525863647461, 'get_old_steps_start_time': 1777137996.0005376, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1777137995.152594, 'checkpoint_manager_blocking_duration_secs': 0.8682243824005127} I0425 17:26:36.020963 140244012873536 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 17:26:36.021038 140244012873536 max_utils.py:750] Memstats: After params initialized: I0425 17:26:36.021087 140244012873536 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 17:26:36.021133 140244012873536 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 17:26:36.021162 140244012873536 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 17:26:36.021186 140244012873536 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 17:26:36.287159 140244012873536 metric_logger.py:185] completed step: 0, seconds: 11.128, TFLOP/s/device: 1.221, Tokens/s/device: 184.047, total_weights: 65536, loss: 10.874 I0425 17:26:36.452607 140244012873536 metric_logger.py:185] completed step: 1, seconds: 1.286, TFLOP/s/device: 10.565, Tokens/s/device: 1592.469, total_weights: 65536, loss: 10.874 I0425 17:26:36.901950 140244012873536 metric_logger.py:185] completed step: 2, seconds: 0.020, TFLOP/s/device: 680.377, Tokens/s/device: 102553.831, total_weights: 65536, loss: 10.560 I0425 17:26:37.049344 140244012873536 metric_logger.py:185] completed step: 3, seconds: 0.440, TFLOP/s/device: 30.855, Tokens/s/device: 4650.867, total_weights: 65536, loss: 9.979 I0425 17:26:37.055449 140244012873536 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 17:26:37.344291 140244012873536 checkpointing.py:798] Waited 0.2888154983520508 seconds for step 5 to finish before starting checkpointing. I0425 17:26:37.347418 140244012873536 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:26:38.659281 2603 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 17:26:40.563457 140118157399808 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0425 17:27:12.703054 140117670885120 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 42.411 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.24513840675354 s) (per-host) I0425 17:27:12.703184 140117670885120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.702829s. I0425 17:27:20.711610 140117670885120 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.711240s. I0425 17:27:20.711906 140118174185216 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:27:20.712029 140118174185216 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:27:20.712080 140118174185216 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:27:20.713608 140118174185216 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 17:27:20.713783 140244012873536 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0425 17:27:20.713936 140244012873536 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.366523 seconds. If this number is high, consider checkpointing less frequently. I0425 17:27:20.715868 140244012873536 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0425 17:27:20.719879 140244012873536 async_checkpointer.py:452] [process=6] 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_15_ocdbt_true/checkpoints/5. I0425 17:27:21.294160 140244012873536 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:27:21.294325 140244012873536 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 17:27:21.338044 140244012873536 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045899s I0425 17:27:21.338211 140244012873536 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.905 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31451988220214844 s) (per-host) I0425 17:27:21.338260 140244012873536 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314579s (batch_requests_ready=0.263931s, total_serialization_initiated=0.050581s, others=0.000067s) I0425 17:27:21.338354 140244012873536 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.319625s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.319610) I0425 17:27:21.339195 140114944567040 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:27:21.339357 140244012873536 async_checkpointer.py:561] Finished blocking save. Time taken: 0.623419s. 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_15_ocdbt_true/checkpoints/5. I0425 17:27:21.341439 140244012873536 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:27:21.341672 140118174185216 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:27:21.341802 140244012873536 standard_logger.py:34] {'step': 5, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777137997.3473885, 'wait_for_prev_duration_secs': 43.366522550582886, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138040.7159061, 'checkpointer_blocking_duration_secs': 0.6235592365264893, 'get_old_steps_start_time': 1777138041.3394852, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1777137997.3447692, 'checkpoint_manager_blocking_duration_secs': 43.996997356414795} I0425 17:27:21.341917 140244012873536 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 17:27:21.342683 140244012873536 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.525, Tokens/s/device: 12589.751, total_weights: 65536, loss: 9.457 I0425 17:27:21.350384 140244012873536 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.329, Tokens/s/device: 13916.825, total_weights: 65536, loss: 8.955 I0425 17:27:40.971646 140244012873536 metric_logger.py:185] completed step: 6, seconds: 44.293, TFLOP/s/device: 0.307, Tokens/s/device: 46.237, total_weights: 65536, loss: 8.465 I0425 17:27:41.119025 140244012873536 metric_logger.py:185] completed step: 7, seconds: 19.476, TFLOP/s/device: 0.698, Tokens/s/device: 105.154, total_weights: 65536, loss: 7.998 I0425 17:27:41.266540 140244012873536 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.672, Tokens/s/device: 13365.616, total_weights: 65536, loss: 7.567 I0425 17:27:41.414964 140244012873536 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 17:27:41.415593 140244012873536 checkpointing.py:798] Waited 0.0006394386291503906 seconds for step 9 to finish before starting checkpointing. I0425 17:27:41.417550 140244012873536 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:27:45.860075 140118157399808 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0425 17:28:22.394349 140114944567040 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.739 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.370619773864746 s) (per-host) I0425 17:28:22.394460 140114944567040 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.055151s. I0425 17:28:31.891734 140114944567040 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 70.552409s. I0425 17:28:31.892013 140118174185216 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:28:31.892158 140118174185216 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:28:31.892209 140118174185216 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:28:31.893690 140118174185216 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 17:28:31.893868 140244012873536 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0425 17:28:31.893989 140244012873536 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.476451 seconds. If this number is high, consider checkpointing less frequently. I0425 17:28:31.895526 140244012873536 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 17:28:31.899026 140244012873536 async_checkpointer.py:452] [process=6] 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_15_ocdbt_true/checkpoints/9. I0425 17:28:32.454907 140244012873536 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:28:32.455004 140244012873536 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 17:28:32.501908 140244012873536 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048998s I0425 17:28:32.502077 140244012873536 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.001 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3084750175476074 s) (per-host) I0425 17:28:32.502150 140244012873536 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.308557s (batch_requests_ready=0.254989s, total_serialization_initiated=0.053482s, others=0.000087s) I0425 17:28:32.502250 140244012873536 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312563s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.312551) I0425 17:28:32.503150 140117670885120 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:28:32.503318 140244012873536 async_checkpointer.py:561] Finished blocking save. Time taken: 0.607721s. 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_15_ocdbt_true/checkpoints/9. I0425 17:28:32.508475 140244012873536 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:28:32.508756 140118174185216 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:28:32.508918 140244012873536 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138061.417522, 'wait_for_prev_duration_secs': 50.47645139694214, 'time_between_consecutive_saves_sec': 20.70387554168701, 'checkpointer_blocking_start_time': 1777138111.8955646, 'checkpointer_blocking_duration_secs': 0.6078650951385498, 'get_old_steps_start_time': 1777138112.5034444, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1777138061.4158316, 'checkpoint_manager_blocking_duration_secs': 51.0930552482605} I0425 17:28:32.509035 140244012873536 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 17:28:32.509079 140244012873536 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:28:37.155267 140118157399808 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0425 17:29:13.401453 140117670885120 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.332 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.20780920982361 s) (per-host) I0425 17:29:13.401577 140117670885120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.898311s. I0425 17:29:21.996247 140117670885120 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.492964s. I0425 17:29:21.996544 140118174185216 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:29:21.996666 140118174185216 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:29:21.996713 140118174185216 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:29:21.998277 140118174185216 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 17:29:21.998457 140244012873536 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0425 17:29:21.998646 140244012873536 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:29:21.999376 140244012873536 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.530, Tokens/s/device: 13947.058, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 17:29:34 UTC 2026 EXIT_CODE=0