XPK Start: Fri Apr 24 17:21:48 UTC 2026 2026-04-24 17:21:52.732858: 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 00:00:1777051312.745568 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777051312.749576 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777051312.761306 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777051312.761325 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777051312.761327 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777051312.761329 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 17:22:12.002326: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0424 17:22:12.536821 134771456657216 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 17:22:21,578:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 17:22:21.578697 134771456657216 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 17:22:21,581:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-ohzie-slice-job-0-0.mt-15-ocdbt-true-sv-ohzie:8482 I0424 17:22:21.581148 134771456657216 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-ohzie-slice-job-0-0.mt-15-ocdbt-true-sv-ohzie:8482 I0424 17:22:22.510126 134771456657216 max_utils.py:284] Jax distributed system initialized! I0424 17:22:29.453004 134771456657216 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 17:22:29.453124 134771456657216 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 17:22:29.453171 134771456657216 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0424 17:22:29.453217 134771456657216 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 17:22:30.091412 134771456657216 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 17:22:30.206866 134771456657216 checkpointing.py:688] Setting up checkpoint logger... I0424 17:22:30.206992 134771456657216 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 17:22:30.207039 134771456657216 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 17:22:30.207277 134771456657216 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 0x7a9268f75490>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 17:22:33.160043 134771456657216 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 17:22:33.160312 134771456657216 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 0x7a7e0850e420>}, handler_registry=None I0424 17:22:33.160545 134771456657216 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a7e0850e420>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0424 17:22:33.160595 134771456657216 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a7e082c4680>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0424 17:22:33.160633 134771456657216 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 0x7a7e0850e420>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a7e0850e420>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a7e082c4680>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a7e082c4680>}). I0424 17:22:33.160955 134771456657216 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 17:22:33.161032 134771456657216 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a7e08660f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 17:22:33.883033 134771456657216 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints I0424 17:22:33.885303 134771456657216 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a7e086f0200> I0424 17:22:33.885417 134771456657216 checkpointing.py:302] Checkpoint manager created! I0424 17:22:34.815545 134771456657216 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 17:22:34.815663 134771456657216 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 17:22:36.229342 134771456657216 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 17:22:36.229464 134771456657216 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:22:36.229521 134771456657216 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:22:36.229571 134771456657216 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 17:22:36.229611 134771456657216 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:22:36.229648 134771456657216 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:22:36.229687 134771456657216 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:22:36.229724 134771456657216 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 17:22:36.229761 134771456657216 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:22:36.229797 134771456657216 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:22:36.229831 134771456657216 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 17:22:36.229865 134771456657216 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 17:22:36.624419 134771456657216 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:22:36.624578 134771456657216 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:22:36.629962 134771456657216 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:22:36.630055 134771456657216 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:22:36.786395 134771456657216 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:22:36.786532 134771456657216 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:22:36.943350 134771456657216 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:22:36.943486 134771456657216 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:22:37.107090 134771456657216 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:22:37.107240 134771456657216 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:22:37.261110 134771456657216 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:22:37.261245 134771456657216 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:22:37.411074 134771456657216 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 17:22:37.411226 134771456657216 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:22:37.572020 134771456657216 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 17:22:37.572182 134771456657216 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:22:37.596843 134771456657216 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 17:22:37.596939 134771456657216 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 17:22:48.485050 134771456657216 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0424 17:22:48.487653 134771456657216 metric_logger.py:289] number parameters: 1.104 billion I0424 17:22:59.777950 134771456657216 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 17:22:59.929137 134771456657216 checkpointing.py:798] Waited 0.15113544464111328 seconds for step 0 to finish before starting checkpointing. I0424 17:22:59.931713 134771456657216 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:22:59.933605 134771456657216 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 17:22:59.936761 134771456657216 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/0. I0424 17:23:00.859376 134771456657216 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 17:23:00.860589 134771456657216 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:23:00.860650 134771456657216 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0424 17:23:01.117389 134771456657216 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.259139s I0424 17:23:01.117557 134771456657216 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.004 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5135462284088135 s) (per-host) I0424 17:23:01.117607 134771456657216 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.513606s (batch_requests_ready=0.247771s, total_serialization_initiated=0.265769s, others=0.000067s) I0424 17:23:01.117700 134771456657216 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.517766s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.517747) I0424 17:23:01.118672 134649061811968 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:23:01.118852 134771456657216 async_checkpointer.py:561] Finished blocking save. Time taken: 1.185173s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/0. I0424 17:23:01.172937 134771456657216 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:23:01.173281 134643735066368 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:23:01.173443 134771456657216 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777051379.9316957, 'wait_for_prev_duration_secs': 6.556510925292969e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777051379.9336464, 'checkpointer_blocking_duration_secs': 1.1853158473968506, 'get_old_steps_start_time': 1777051381.1189833, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1777051379.9295442, 'checkpoint_manager_blocking_duration_secs': 1.2438592910766602} I0424 17:23:01.173556 134771456657216 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 17:23:01.173630 134771456657216 max_utils.py:750] Memstats: After params initialized: I0424 17:23:01.173683 134771456657216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 17:23:01.173716 134771456657216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 17:23:01.173743 134771456657216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 17:23:01.173768 134771456657216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 17:23:01.447005 134771456657216 metric_logger.py:185] completed step: 0, seconds: 11.289, TFLOP/s/device: 1.204, Tokens/s/device: 181.412, total_weights: 65536, loss: 10.874 I0424 17:23:01.612157 134771456657216 metric_logger.py:185] completed step: 1, seconds: 1.667, TFLOP/s/device: 8.149, Tokens/s/device: 1228.344, total_weights: 65536, loss: 10.874 I0424 17:23:02.054537 134771456657216 metric_logger.py:185] completed step: 2, seconds: 0.019, TFLOP/s/device: 698.747, Tokens/s/device: 105322.705, total_weights: 65536, loss: 10.560 I0424 17:23:02.201793 134771456657216 metric_logger.py:185] completed step: 3, seconds: 0.433, TFLOP/s/device: 31.359, Tokens/s/device: 4726.845, total_weights: 65536, loss: 9.979 I0424 17:23:02.207977 134771456657216 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 17:23:02.496577 134771456657216 checkpointing.py:798] Waited 0.28858304023742676 seconds for step 5 to finish before starting checkpointing. I0424 17:23:02.499202 134771456657216 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:23:04.032934 2654 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 17:23:05.960870 134649070204672 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0424 17:23:38.575753 134649061811968 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 41.599 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.97170090675354 s) (per-host) I0424 17:23:38.575875 134649061811968 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.457087s. I0424 17:23:44.500651 134649061811968 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 43.381847s. I0424 17:23:44.500891 134643735066368 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:23:44.500969 134643735066368 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:23:44.501016 134643735066368 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:23:44.502717 134643735066368 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 17:23:44.502914 134771456657216 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. W0424 17:23:44.503047 134771456657216 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.003850 seconds. If this number is high, consider checkpointing less frequently. I0424 17:23:44.505058 134771456657216 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0424 17:23:44.509128 134771456657216 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/5. I0424 17:23:45.415817 134771456657216 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:23:45.415975 134771456657216 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0424 17:23:45.466184 134771456657216 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.052404s I0424 17:23:45.466350 134771456657216 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.014 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30765676498413086 s) (per-host) I0424 17:23:45.466403 134771456657216 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307720s (batch_requests_ready=0.250606s, total_serialization_initiated=0.057045s, others=0.000070s) I0424 17:23:45.466531 134771456657216 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312608s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.312592) I0424 17:23:45.467472 134645300152064 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:23:45.467637 134771456657216 async_checkpointer.py:561] Finished blocking save. Time taken: 0.962492s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/5. I0424 17:23:45.486109 134771456657216 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:23:45.486387 134643735066368 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:23:45.486555 134771456657216 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777051382.4991748, 'wait_for_prev_duration_secs': 42.003849506378174, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777051424.505111, 'checkpointer_blocking_duration_secs': 0.9626362323760986, 'get_old_steps_start_time': 1777051425.4677675, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777051382.4969246, 'checkpoint_manager_blocking_duration_secs': 42.98959183692932} I0424 17:23:45.486666 134771456657216 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 17:23:45.487418 134771456657216 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.365, Tokens/s/device: 12565.651, total_weights: 65536, loss: 9.457 I0424 17:23:45.494528 134771456657216 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.453, Tokens/s/device: 13935.480, total_weights: 65536, loss: 8.955 I0424 17:24:05.335128 134771456657216 metric_logger.py:185] completed step: 6, seconds: 43.285, TFLOP/s/device: 0.314, Tokens/s/device: 47.314, total_weights: 65536, loss: 8.465 I0424 17:24:05.482641 134771456657216 metric_logger.py:185] completed step: 7, seconds: 19.695, TFLOP/s/device: 0.690, Tokens/s/device: 103.986, total_weights: 65536, loss: 7.998 I0424 17:24:05.629947 134771456657216 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.556, Tokens/s/device: 13348.194, total_weights: 65536, loss: 7.567 I0424 17:24:05.778401 134771456657216 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 17:24:05.778980 134771456657216 checkpointing.py:798] Waited 0.0005915164947509766 seconds for step 9 to finish before starting checkpointing. I0424 17:24:05.781127 134771456657216 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:24:09.790926 134649070204672 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0424 17:24:47.327858 134645300152064 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.408 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 62.16911506652832 s) (per-host) I0424 17:24:47.327981 134645300152064 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.860394s. I0424 17:24:52.541754 134645300152064 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 67.074151s. I0424 17:24:52.542048 134643735066368 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:24:52.542183 134643735066368 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:24:52.542230 134643735066368 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:24:52.543534 134643735066368 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 17:24:52.543716 134771456657216 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. W0424 17:24:52.543843 134771456657216 checkpoint_manager.py:1458] Waiting for previous save to complete took 46.762746 seconds. If this number is high, consider checkpointing less frequently. I0424 17:24:52.545453 134771456657216 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0424 17:24:52.548812 134771456657216 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/9. I0424 17:24:53.099722 134771456657216 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:24:53.099821 134771456657216 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0424 17:24:53.147372 134771456657216 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049604s I0424 17:24:53.147518 134771456657216 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.899 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3148462772369385 s) (per-host) I0424 17:24:53.147565 134771456657216 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314903s (batch_requests_ready=0.260821s, total_serialization_initiated=0.054020s, others=0.000061s) I0424 17:24:53.147655 134771456657216 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.319060s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.319048) I0424 17:24:53.148560 134649061811968 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:24:53.148699 134771456657216 async_checkpointer.py:561] Finished blocking save. Time taken: 0.603176s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/9. I0424 17:24:53.462677 134771456657216 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:24:53.463042 134643735066368 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:24:53.463233 134771456657216 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777051445.781082, 'wait_for_prev_duration_secs': 46.76274609565735, 'time_between_consecutive_saves_sec': 21.27832341194153, 'checkpointer_blocking_start_time': 1777051492.5454917, 'checkpointer_blocking_duration_secs': 0.6033847332000732, 'get_old_steps_start_time': 1777051493.1488903, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1777051445.7792144, 'checkpoint_manager_blocking_duration_secs': 47.68398594856262} I0424 17:24:53.463356 134771456657216 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 17:24:53.463401 134771456657216 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:24:58.347299 134649070204672 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_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0424 17:25:36.500487 134649061811968 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.173 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.66777539253235 s) (per-host) I0424 17:25:36.500607 134649061811968 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.351815s. I0424 17:25:41.349417 134649061811968 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.200622s. I0424 17:25:41.349753 134643735066368 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:25:41.349872 134643735066368 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:25:41.349916 134643735066368 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:25:41.351410 134643735066368 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 17:25:41.351600 134771456657216 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. I0424 17:25:41.351768 134771456657216 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:25:41.352404 134771456657216 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.261, Tokens/s/device: 13906.619, 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: Fri Apr 24 17:25:53 UTC 2026 EXIT_CODE=0