XPK Start: Wed Apr 22 18:28:22 UTC 2026 2026-04-22 18:28:27.037256: 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 00:00:1776882507.050079 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776882507.054085 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776882507.065525 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882507.065544 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882507.065546 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882507.065548 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 18:28:46.343946: 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) I0422 18:28:46.871216 135922959333184 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 18:28:55,913:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 18:28:55.913435 135922959333184 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 18:28:55,915:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-h4tzd-slice-job-0-0.mt-15-ocdbt-true-sv-h4tzd:8482 I0422 18:28:55.915884 135922959333184 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-h4tzd-slice-job-0-0.mt-15-ocdbt-true-sv-h4tzd:8482 I0422 18:28:56.949839 135922959333184 max_utils.py:284] Jax distributed system initialized! I0422 18:29:03.955890 135922959333184 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:29:03.956001 135922959333184 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:29:03.956043 135922959333184 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 I0422 18:29:03.956078 135922959333184 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:29:04.580130 135922959333184 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:29:04.696227 135922959333184 checkpointing.py:688] Setting up checkpoint logger... I0422 18:29:04.696349 135922959333184 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 18:29:04.696394 135922959333184 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 18:29:04.696597 135922959333184 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 0x7b9e7e5792b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:29:08.519023 135922959333184 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 18:29:08.519404 135922959333184 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 0x7b8a08328200>}, handler_registry=None I0422 18:29:08.519642 135922959333184 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8a08328200>` 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`. I0422 18:29:08.519693 135922959333184 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8a084b4170>` 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`. I0422 18:29:08.519729 135922959333184 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 0x7b8a08328200>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8a08328200>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8a084b4170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8a084b4170>}). I0422 18:29:08.520061 135922959333184 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 18:29:08.520151 135922959333184 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b8a0841d260> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:29:09.671360 135922959333184 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints I0422 18:29:10.084620 135922959333184 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b8a084b3a70> I0422 18:29:10.084795 135922959333184 checkpointing.py:302] Checkpoint manager created! I0422 18:29:11.031206 135922959333184 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:29:11.031322 135922959333184 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 18:29:12.468669 135922959333184 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 18:29:12.468792 135922959333184 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:29:12.468848 135922959333184 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:29:12.468892 135922959333184 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 18:29:12.468932 135922959333184 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:29:12.468967 135922959333184 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:29:12.469005 135922959333184 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:29:12.469042 135922959333184 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 18:29:12.469079 135922959333184 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:29:12.469140 135922959333184 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:29:12.469182 135922959333184 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 18:29:12.469213 135922959333184 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 18:29:12.865681 135922959333184 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:29:12.865829 135922959333184 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:29:12.871293 135922959333184 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:29:12.871395 135922959333184 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:29:13.023661 135922959333184 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:29:13.023797 135922959333184 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:29:13.173758 135922959333184 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:29:13.173895 135922959333184 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:29:13.333947 135922959333184 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:29:13.334085 135922959333184 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:29:13.483030 135922959333184 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:29:13.483180 135922959333184 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:29:13.634815 135922959333184 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:29:13.634953 135922959333184 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:29:13.797485 135922959333184 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:29:13.797629 135922959333184 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:29:13.820328 135922959333184 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:29:13.820428 135922959333184 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 18:29:24.786485 135922959333184 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. I0422 18:29:24.788896 135922959333184 metric_logger.py:289] number parameters: 1.104 billion I0422 18:29:35.897742 135922959333184 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 18:29:36.049840 135922959333184 checkpointing.py:798] Waited 0.1520836353302002 seconds for step 0 to finish before starting checkpointing. I0422 18:29:36.052915 135922959333184 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:29:36.054568 135922959333184 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0422 18:29:36.057337 135922959333184 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/0. I0422 18:29:37.049202 135922959333184 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 18:29:37.050207 135922959333184 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:29:37.050272 135922959333184 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 I0422 18:29:37.300557 135922959333184 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252616s I0422 18:29:37.300725 135922959333184 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.987 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5164422988891602 s) (per-host) I0422 18:29:37.300784 135922959333184 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.516508s (batch_requests_ready=0.257005s, total_serialization_initiated=0.259429s, others=0.000073s) I0422 18:29:37.300899 135922959333184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.520572s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.520552) I0422 18:29:37.301821 135800020793088 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:29:37.301994 135922959333184 async_checkpointer.py:561] Finished blocking save. Time taken: 1.247360s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/0. I0422 18:29:37.718391 135922959333184 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:29:37.718772 135795709040384 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:29:37.718936 135922959333184 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882576.0528886, 'wait_for_prev_duration_secs': 0.00011515617370605469, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776882576.0546072, 'checkpointer_blocking_duration_secs': 1.247502088546753, 'get_old_steps_start_time': 1776882577.3021288, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776882576.0502558, 'checkpoint_manager_blocking_duration_secs': 1.6686394214630127} I0422 18:29:37.719046 135922959333184 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 18:29:37.719133 135922959333184 max_utils.py:750] Memstats: After params initialized: I0422 18:29:37.719193 135922959333184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 18:29:37.719228 135922959333184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 18:29:37.719256 135922959333184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 18:29:37.719281 135922959333184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 18:29:37.987298 135922959333184 metric_logger.py:185] completed step: 0, seconds: 11.108, TFLOP/s/device: 1.223, Tokens/s/device: 184.375, total_weights: 65536, loss: 10.874 I0422 18:29:38.166694 135922959333184 metric_logger.py:185] completed step: 1, seconds: 2.088, TFLOP/s/device: 6.508, Tokens/s/device: 980.941, total_weights: 65536, loss: 10.874 I0422 18:29:38.595561 135922959333184 metric_logger.py:185] completed step: 2, seconds: 0.034, TFLOP/s/device: 399.927, Tokens/s/device: 60281.392, total_weights: 65536, loss: 10.560 I0422 18:29:38.742861 135922959333184 metric_logger.py:185] completed step: 3, seconds: 0.420, TFLOP/s/device: 32.376, Tokens/s/device: 4880.025, total_weights: 65536, loss: 9.979 I0422 18:29:38.748984 135922959333184 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0422 18:29:39.037791 135922959333184 checkpointing.py:798] Waited 0.28877925872802734 seconds for step 5 to finish before starting checkpointing. I0422 18:29:39.040880 135922959333184 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:29:40.119208 2596 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 18:29:42.429284 135800507307776 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0422 18:30:14.772005 135800020793088 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 41.582 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.98768162727356 s) (per-host) I0422 18:30:14.772142 135800020793088 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.470205s. I0422 18:30:22.324830 135800020793088 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 45.022877s. I0422 18:30:22.325137 135795709040384 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:30:22.325290 135795709040384 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:30:22.325341 135795709040384 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:30:22.326854 135795709040384 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 18:30:22.327036 135922959333184 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. W0422 18:30:22.327195 135922959333184 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.286321 seconds. If this number is high, consider checkpointing less frequently. I0422 18:30:22.329185 135922959333184 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0422 18:30:22.333215 135922959333184 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5. I0422 18:30:23.731981 135922959333184 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:30:23.732161 135922959333184 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 I0422 18:30:23.773699 135922959333184 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043951s I0422 18:30:23.773871 135922959333184 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.109 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7315003871917725 s) (per-host) I0422 18:30:23.773923 135922959333184 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.731562s (batch_requests_ready=0.682465s, total_serialization_initiated=0.049028s, others=0.000069s) I0422 18:30:23.774045 135922959333184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.736532s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.736516) I0422 18:30:23.774904 135795161818880 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:30:23.775060 135922959333184 async_checkpointer.py:561] Finished blocking save. Time taken: 1.445805s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5. I0422 18:30:23.800400 135922959333184 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:30:23.800693 135795709040384 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:30:23.800832 135922959333184 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882579.0408497, 'wait_for_prev_duration_secs': 43.28632116317749, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776882622.3292239, 'checkpointer_blocking_duration_secs': 1.445946216583252, 'get_old_steps_start_time': 1776882623.7751887, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776882579.0381467, 'checkpoint_manager_blocking_duration_secs': 44.76265072822571} I0422 18:30:23.800950 135922959333184 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0422 18:30:23.801743 135922959333184 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.400, Tokens/s/device: 12570.973, total_weights: 65536, loss: 9.457 I0422 18:30:23.809126 135922959333184 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.521, Tokens/s/device: 13945.729, total_weights: 65536, loss: 8.955 I0422 18:30:43.476645 135922959333184 metric_logger.py:185] completed step: 6, seconds: 45.058, TFLOP/s/device: 0.302, Tokens/s/device: 45.452, total_weights: 65536, loss: 8.465 I0422 18:30:43.624036 135922959333184 metric_logger.py:185] completed step: 7, seconds: 19.522, TFLOP/s/device: 0.696, Tokens/s/device: 104.907, total_weights: 65536, loss: 7.998 I0422 18:30:43.771486 135922959333184 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.496, Tokens/s/device: 13339.152, total_weights: 65536, loss: 7.567 I0422 18:30:43.919918 135922959333184 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 18:30:43.920551 135922959333184 checkpointing.py:798] Waited 0.000644683837890625 seconds for step 9 to finish before starting checkpointing. I0422 18:30:43.923341 135922959333184 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:30:47.312119 135795170211584 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0422 18:31:24.184809 135795161818880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.835 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.14239764213562 s) (per-host) I0422 18:31:24.184938 135795161818880 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.409921s. I0422 18:31:30.676863 135795161818880 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 66.901833s. I0422 18:31:30.677186 135795709040384 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:31:30.677319 135795709040384 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:31:30.677366 135795709040384 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:31:30.678971 135795709040384 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 18:31:30.679166 135922959333184 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. W0422 18:31:30.679289 135922959333184 checkpoint_manager.py:1458] Waiting for previous save to complete took 46.755960 seconds. If this number is high, consider checkpointing less frequently. I0422 18:31:30.680879 135922959333184 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0422 18:31:30.684312 135922959333184 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9. I0422 18:31:31.667531 135922959333184 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:31:31.667634 135922959333184 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 I0422 18:31:31.709069 135922959333184 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043759s I0422 18:31:31.709254 135922959333184 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.112 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7304830551147461 s) (per-host) I0422 18:31:31.709304 135922959333184 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.730543s (batch_requests_ready=0.682236s, total_serialization_initiated=0.048242s, others=0.000065s) I0422 18:31:31.709399 135922959333184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.734654s (all_items=0.000013s, per_item={'items': '0.00001264'}, temp_paths=0.734642) I0422 18:31:31.710335 135796222625536 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:31:31.710491 135922959333184 async_checkpointer.py:561] Finished blocking save. Time taken: 1.029544s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9. I0422 18:31:31.712390 135922959333184 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:31:31.712612 135795709040384 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:31:31.712735 135922959333184 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882643.9233124, 'wait_for_prev_duration_secs': 46.75596022605896, 'time_between_consecutive_saves_sec': 21.59641981124878, 'checkpointer_blocking_start_time': 1776882690.6809163, 'checkpointer_blocking_duration_secs': 1.0296921730041504, 'get_old_steps_start_time': 1776882691.7106225, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1776882643.9207726, 'checkpoint_manager_blocking_duration_secs': 47.79193353652954} I0422 18:31:31.712887 135922959333184 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 18:31:31.712928 135922959333184 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:31:37.015306 135800020793088 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0422 18:32:13.552770 135796222625536 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.102 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.57395601272583 s) (per-host) I0422 18:32:13.552872 135796222625536 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.842427s. I0422 18:32:19.477655 135796222625536 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.767196s. I0422 18:32:19.477940 135795709040384 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:32:19.478084 135795709040384 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:32:19.478147 135795709040384 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:32:19.479595 135795709040384 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 18:32:19.479782 135922959333184 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. I0422 18:32:19.479947 135922959333184 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:32:19.480590 135922959333184 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.504, Tokens/s/device: 13943.260, 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: Wed Apr 22 18:32:33 UTC 2026 EXIT_CODE=0