XPK Start: Mon Apr 20 21:50:56 UTC 2026 2026-04-20 21:51:00.690464: 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 00:00:1776721860.703347 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776721860.707041 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776721860.718532 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776721860.718552 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776721860.718554 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776721860.718556 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 21:51:19.973383: 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) I0420 21:51:20.503090 134096028448576 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 21:51:29,545:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 21:51:29.545924 134096028448576 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 21:51:29,548:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-lgl0s-slice-job-0-0.mt-15-ocdbt-true-sv-lgl0s:8482 I0420 21:51:29.548249 134096028448576 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-lgl0s-slice-job-0-0.mt-15-ocdbt-true-sv-lgl0s:8482 I0420 21:51:32.016588 134096028448576 max_utils.py:284] Jax distributed system initialized! I0420 21:51:38.858981 134096028448576 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 21:51:38.859086 134096028448576 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 21:51:38.859127 134096028448576 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 I0420 21:51:38.859163 134096028448576 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 21:51:39.482189 134096028448576 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 21:51:39.597139 134096028448576 checkpointing.py:688] Setting up checkpoint logger... I0420 21:51:39.597260 134096028448576 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 21:51:39.597307 134096028448576 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 21:51:39.597521 134096028448576 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 0x79f520d79310>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 21:51:42.872570 134096028448576 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 21:51:42.872940 134096028448576 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e1b73af500>}, handler_registry=None I0420 21:51:42.873267 134096028448576 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e1b73af500>` 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`. I0420 21:51:42.873329 134096028448576 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e1b501a420>` 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`. I0420 21:51:42.873381 134096028448576 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 0x79e1b73af500>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e1b73af500>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e1b501a420>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e1b501a420>}). I0420 21:51:42.873766 134096028448576 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 21:51:42.873867 134096028448576 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79e06c390f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 21:51:43.761715 134096028448576 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints I0420 21:51:43.803476 134096028448576 checkpoint_manager.py:929] [process=7][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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79e1b5026600> I0420 21:51:43.803623 134096028448576 checkpointing.py:302] Checkpoint manager created! I0420 21:51:44.724323 134096028448576 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 21:51:44.724445 134096028448576 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 21:51:46.154070 134096028448576 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 21:51:46.154192 134096028448576 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:51:46.154252 134096028448576 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:51:46.154298 134096028448576 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 21:51:46.154338 134096028448576 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:51:46.154374 134096028448576 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:51:46.154412 134096028448576 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:51:46.154448 134096028448576 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 21:51:46.154483 134096028448576 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:51:46.154517 134096028448576 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:51:46.154550 134096028448576 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 21:51:46.154583 134096028448576 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 21:51:46.550109 134096028448576 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:51:46.550266 134096028448576 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:51:46.555789 134096028448576 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:51:46.555896 134096028448576 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:51:46.712391 134096028448576 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:51:46.712528 134096028448576 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:51:46.872577 134096028448576 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:51:46.872733 134096028448576 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:51:47.037722 134096028448576 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:51:47.037859 134096028448576 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:51:47.188502 134096028448576 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:51:47.188637 134096028448576 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:51:47.337277 134096028448576 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:51:47.337415 134096028448576 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:51:47.494697 134096028448576 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 21:51:47.494835 134096028448576 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:51:47.518777 134096028448576 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 21:51:47.518874 134096028448576 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 21:51:58.654727 134096028448576 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. I0420 21:51:58.657434 134096028448576 metric_logger.py:289] number parameters: 1.104 billion I0420 21:52:09.887835 134096028448576 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 21:52:10.044288 134096028448576 checkpointing.py:798] Waited 0.15644288063049316 seconds for step 0 to finish before starting checkpointing. I0420 21:52:10.046956 134096028448576 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:52:10.048719 134096028448576 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0420 21:52:10.051587 134096028448576 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/0. I0420 21:52:11.058202 134096028448576 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 21:52:11.059206 134096028448576 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:52:11.059265 134096028448576 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 I0420 21:52:11.308185 134096028448576 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251105s I0420 21:52:11.308368 134096028448576 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.927 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5269408226013184 s) (per-host) I0420 21:52:11.308423 134096028448576 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.527007s (batch_requests_ready=0.269354s, total_serialization_initiated=0.257578s, others=0.000075s) I0420 21:52:11.308521 134096028448576 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.531063s (all_items=0.000027s, per_item={'items': '0.00002694'}, temp_paths=0.531036) I0420 21:52:11.309469 133972872898304 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:52:11.309632 134096028448576 async_checkpointer.py:561] Finished blocking save. Time taken: 1.260842s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/0. I0420 21:52:11.341140 134096028448576 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:52:11.341487 133968577931008 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:52:11.341648 134096028448576 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776721930.0469244, 'wait_for_prev_duration_secs': 0.00013399124145507812, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776721930.0487583, 'checkpointer_blocking_duration_secs': 1.2610018253326416, 'get_old_steps_start_time': 1776721931.30978, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776721930.0450215, 'checkpoint_manager_blocking_duration_secs': 1.2965872287750244} I0420 21:52:11.341778 134096028448576 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 21:52:11.341856 134096028448576 max_utils.py:750] Memstats: After params initialized: I0420 21:52:11.341910 134096028448576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 21:52:11.341943 134096028448576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 21:52:11.341971 134096028448576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 21:52:11.341998 134096028448576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 21:52:11.612519 134096028448576 metric_logger.py:185] completed step: 0, seconds: 11.229, TFLOP/s/device: 1.210, Tokens/s/device: 182.379, total_weights: 65536, loss: 10.874 I0420 21:52:11.778448 134096028448576 metric_logger.py:185] completed step: 1, seconds: 1.723, TFLOP/s/device: 7.886, Tokens/s/device: 1188.725, total_weights: 65536, loss: 10.874 I0420 21:52:12.220921 134096028448576 metric_logger.py:185] completed step: 2, seconds: 0.020, TFLOP/s/device: 667.541, Tokens/s/device: 100619.043, total_weights: 65536, loss: 10.560 I0420 21:52:12.368431 134096028448576 metric_logger.py:185] completed step: 3, seconds: 0.433, TFLOP/s/device: 31.412, Tokens/s/device: 4734.702, total_weights: 65536, loss: 9.979 I0420 21:52:12.374350 134096028448576 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0420 21:52:12.663513 134096028448576 checkpointing.py:798] Waited 0.2891366481781006 seconds for step 5 to finish before starting checkpointing. I0420 21:52:12.666245 134096028448576 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:52:14.430065 2604 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 21:52:16.377104 133972881291008 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_7 I0420 21:52:48.106263 133972872898304 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 42.320 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.32477617263794 s) (per-host) I0420 21:52:48.106391 133972872898304 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.796812s. I0420 21:53:02.027626 133972872898304 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 50.718005s. I0420 21:53:02.027941 133968577931008 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:53:02.028083 133968577931008 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:53:02.028135 133968577931008 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:53:02.030805 133968577931008 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 21:53:02.031000 134096028448576 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 21:53:02.031146 134096028448576 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.364904 seconds. If this number is high, consider checkpointing less frequently. I0420 21:53:02.033265 134096028448576 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 5 I0420 21:53:02.037307 134096028448576 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/5. I0420 21:53:03.323136 134096028448576 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:53:03.323287 134096028448576 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 I0420 21:53:03.369904 134096028448576 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048785s I0420 21:53:03.370066 134096028448576 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.563 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.6019606590270996 s) (per-host) I0420 21:53:03.370115 134096028448576 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.602019s (batch_requests_ready=0.548080s, total_serialization_initiated=0.053873s, others=0.000066s) I0420 21:53:03.370206 134096028448576 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.607093s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.607077) I0420 21:53:03.371052 133969108883200 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:53:03.371166 134096028448576 async_checkpointer.py:561] Finished blocking save. Time taken: 1.337831s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/5. I0420 21:53:03.922313 134096028448576 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:53:03.922721 133968577931008 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:53:03.922875 134096028448576 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776721932.666216, 'wait_for_prev_duration_secs': 49.364903926849365, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776721982.033305, 'checkpointer_blocking_duration_secs': 1.33797025680542, 'get_old_steps_start_time': 1776721983.371294, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1776721932.663901, 'checkpoint_manager_blocking_duration_secs': 51.25893759727478} I0420 21:53:03.923058 134096028448576 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0420 21:53:03.923881 134096028448576 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 83.057, Tokens/s/device: 12519.256, total_weights: 65536, loss: 9.457 I0420 21:53:03.931417 134096028448576 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.372, Tokens/s/device: 13923.259, total_weights: 65536, loss: 8.955 I0420 21:53:26.529212 134096028448576 metric_logger.py:185] completed step: 6, seconds: 51.555, TFLOP/s/device: 0.264, Tokens/s/device: 39.724, total_weights: 65536, loss: 8.465 I0420 21:53:26.676760 134096028448576 metric_logger.py:185] completed step: 7, seconds: 22.453, TFLOP/s/device: 0.605, Tokens/s/device: 91.214, total_weights: 65536, loss: 7.998 I0420 21:53:26.824464 134096028448576 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.662, Tokens/s/device: 13364.046, total_weights: 65536, loss: 7.567 I0420 21:53:26.972866 134096028448576 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 21:53:26.973506 134096028448576 checkpointing.py:798] Waited 0.0006496906280517578 seconds for step 9 to finish before starting checkpointing. I0420 21:53:26.975388 134096028448576 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:53:32.048685 133972881291008 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_7 I0420 21:54:08.702644 133969108883200 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 23.957 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 65.9345064163208 s) (per-host) I0420 21:54:08.702791 133969108883200 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 65.331669s. I0420 21:54:16.857713 133969108883200 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 73.486576s. I0420 21:54:16.858018 133968577931008 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:54:16.858159 133968577931008 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:54:16.858226 133968577931008 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:54:16.859880 133968577931008 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0420 21:54:16.860074 134096028448576 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0420 21:54:16.860228 134096028448576 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.884844 seconds. If this number is high, consider checkpointing less frequently. I0420 21:54:16.861932 134096028448576 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0420 21:54:16.865528 134096028448576 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9. I0420 21:54:17.438332 134096028448576 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:54:17.438434 134096028448576 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 I0420 21:54:17.482740 134096028448576 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046452s I0420 21:54:17.482904 134096028448576 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.966 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3105964660644531 s) (per-host) I0420 21:54:17.482956 134096028448576 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.310658s (batch_requests_ready=0.259445s, total_serialization_initiated=0.051146s, others=0.000067s) I0420 21:54:17.483051 134096028448576 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.314826s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.314815) I0420 21:54:17.483958 133972872898304 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:54:17.484121 134096028448576 async_checkpointer.py:561] Finished blocking save. Time taken: 0.622120s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9. I0420 21:54:17.493828 134096028448576 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:54:17.494118 133968577931008 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:54:17.494280 134096028448576 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776722006.9753613, 'wait_for_prev_duration_secs': 49.884844064712524, 'time_between_consecutive_saves_sec': 24.94451642036438, 'checkpointer_blocking_start_time': 1776722056.8619714, 'checkpointer_blocking_duration_secs': 0.6222596168518066, 'get_old_steps_start_time': 1776722057.4842503, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1776722006.9737558, 'checkpoint_manager_blocking_duration_secs': 50.520493030548096} I0420 21:54:17.494397 134096028448576 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 21:54:17.494443 134096028448576 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:54:22.223955 133972881291008 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_7 I0420 21:55:00.718721 133972872898304 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 36.274 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.54635572433472 s) (per-host) I0420 21:55:00.718841 133972872898304 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.234767s. I0420 21:55:11.427488 133972872898304 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 53.943399s. I0420 21:55:11.427824 133968577931008 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:55:11.427880 133968577931008 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:55:11.427930 133968577931008 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:55:11.429349 133968577931008 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 21:55:11.429527 134096028448576 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 21:55:11.429709 134096028448576 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:55:11.430486 134096028448576 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.009, Tokens/s/device: 13868.574, 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: Mon Apr 20 21:55:22 UTC 2026 EXIT_CODE=0