XPK Start: Sat Apr 25 17:39:46 UTC 2026 2026-04-25 17:39:50.221792: 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 00:00:1777138790.234927 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777138790.238703 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777138790.250412 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138790.250433 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138790.250435 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138790.250437 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:40:09.441969: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0425 17:40:09.973632 139654831384384 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 17:40:19,014:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 17:40:19.014439 139654831384384 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 17:40:19,016:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-rs-dkior-slice-job-0-0.mt-15-ocdbt-false-rs-dkior:8482 I0425 17:40:19.016800 139654831384384 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-dkior-slice-job-0-0.mt-15-ocdbt-false-rs-dkior:8482 I0425 17:40:20.265006 139654831384384 max_utils.py:284] Jax distributed system initialized! I0425 17:40:27.126486 139654831384384 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 17:40:27.126591 139654831384384 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 17:40:27.126631 139654831384384 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0425 17:40:27.126665 139654831384384 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 17:40:27.748123 139654831384384 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 17:40:27.865025 139654831384384 checkpointing.py:688] Setting up checkpoint logger... I0425 17:40:27.865178 139654831384384 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0425 17:40:27.865226 139654831384384 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 17:40:27.865434 139654831384384 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7f03683708f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 17:40:30.086020 139654831384384 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 17:40:30.086430 139654831384384 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 0x7ef04a5dcd10>}, handler_registry=None I0425 17:40:30.086669 139654831384384 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ef04a5dcd10>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0425 17:40:30.086717 139654831384384 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ef04a5dc170>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0425 17:40:30.086752 139654831384384 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 0x7ef04a5dcd10>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ef04a5dcd10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ef04a5dc170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ef04a5dc170>}). I0425 17:40:30.087077 139654831384384 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 17:40:30.087169 139654831384384 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ef04a5a0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 17:40:34.107794 139536474552064 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777138699000384761, 'commit_timestamp_nsecs': 1777138712902394602, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/9/_CHECKPOINT_METADATA I0425 17:40:34.801430 139536474552064 checkpoint_manager.py:1877] Missing metrics for step 9 I0425 17:40:36.035847 139654831384384 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints I0425 17:40:36.914603 139654831384384 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ef04a630290> I0425 17:40:36.914780 139654831384384 checkpointing.py:302] Checkpoint manager created! I0425 17:40:37.854767 139654831384384 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 17:40:37.854887 139654831384384 checkpointing.py:582] restoring from this run's directory step 9 I0425 17:40:38.393025 139654831384384 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/9. I0425 17:40:42.042018 1443 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 17:40:51.559901 139654831384384 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 144.090 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.967283248901367 s) (per-host) I0425 17:40:51.581896 139654831384384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 139.144 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.373836517333984 s) (per-host) I0425 17:40:51.583923 139654831384384 checkpointer.py:318] Finished restoring checkpoint in 13.45 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/9. I0425 17:40:51.584017 139654831384384 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1777138838.132322, 'checkpointer_duration_secs': 13.451667070388794, 'checkpoint_manager_start_time': 1777138837.8555372, 'checkpoint_manager_duration_secs': 13.728453636169434} fsdp: 32 I0425 17:40:51.665264 139654831384384 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 17:40:51.665379 139654831384384 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:40:51.665431 139654831384384 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:40:51.665474 139654831384384 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 17:40:51.665512 139654831384384 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:40:51.665547 139654831384384 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:40:51.665590 139654831384384 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:40:51.665630 139654831384384 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 17:40:51.665667 139654831384384 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:40:51.665701 139654831384384 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:40:51.665735 139654831384384 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 17:40:51.665769 139654831384384 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 17:40:52.061259 139654831384384 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:40:52.061417 139654831384384 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:40:52.066834 139654831384384 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:40:52.066929 139654831384384 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:40:52.223437 139654831384384 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:40:52.223580 139654831384384 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:40:52.376883 139654831384384 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:40:52.377017 139654831384384 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:40:52.541567 139654831384384 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:40:52.541704 139654831384384 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:40:52.696969 139654831384384 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:40:52.697113 139654831384384 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:40:52.857416 139654831384384 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:40:52.857576 139654831384384 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:40:53.025850 139654831384384 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 17:40:53.025981 139654831384384 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:40:53.048935 139654831384384 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 17:40:53.049033 139654831384384 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 17:41:04.411442 139654831384384 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0425 17:41:04.413988 139654831384384 metric_logger.py:289] number parameters: 1.104 billion I0425 17:41:15.717227 139654831384384 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0425 17:41:15.867833 139654831384384 checkpointing.py:798] Waited 0.15059256553649902 seconds for step 10 to finish before starting checkpointing. I0425 17:41:15.870236 139654831384384 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:41:15.872280 139654831384384 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0425 17:41:15.874910 139654831384384 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/10. I0425 17:41:17.340808 139654831384384 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 17:41:17.341787 139654831384384 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:41:17.341845 139654831384384 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0425 17:41:17.596800 139654831384384 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.257139s I0425 17:41:17.596969 139654831384384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.865 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.538327693939209 s) (per-host) I0425 17:41:17.597024 139654831384384 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.538393s (batch_requests_ready=0.276544s, total_serialization_initiated=0.261775s, others=0.000074s) I0425 17:41:17.597140 139654831384384 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.542478s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.542459) I0425 17:41:17.598047 139533798590208 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:41:17.598258 139654831384384 async_checkpointer.py:561] Finished blocking save. Time taken: 1.725913s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/10. I0425 17:41:18.419172 139654831384384 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:41:18.419545 139533286897408 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:41:18.421295 139654831384384 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138875.870216, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138875.872324, 'checkpointer_blocking_duration_secs': 1.7260072231292725, 'get_old_steps_start_time': 1777138877.5983496, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1777138875.8682249, 'checkpoint_manager_blocking_duration_secs': 2.553032159805298} I0425 17:41:18.421412 139654831384384 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0425 17:41:18.421487 139654831384384 max_utils.py:750] Memstats: After params initialized: I0425 17:41:18.421545 139654831384384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 17:41:18.421577 139654831384384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 17:41:18.421603 139654831384384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 17:41:18.421633 139654831384384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 17:41:18.695185 139654831384384 metric_logger.py:185] completed step: 10, seconds: 11.302, TFLOP/s/device: 1.202, Tokens/s/device: 181.204, total_weights: 65536, loss: 6.827 I0425 17:41:18.862081 139654831384384 metric_logger.py:185] completed step: 11, seconds: 2.976, TFLOP/s/device: 4.565, Tokens/s/device: 688.110, total_weights: 65536, loss: 6.522 I0425 17:41:19.009618 139654831384384 metric_logger.py:185] completed step: 12, seconds: 0.011, TFLOP/s/device: 1182.930, Tokens/s/device: 178304.022, total_weights: 65536, loss: 6.260 I0425 17:41:19.157053 139654831384384 metric_logger.py:185] completed step: 13, seconds: 0.163, TFLOP/s/device: 83.172, Tokens/s/device: 12536.575, total_weights: 65536, loss: 6.039 I0425 17:41:19.163688 139654831384384 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0425 17:41:19.452129 139654831384384 checkpointing.py:798] Waited 0.2883932590484619 seconds for step 15 to finish before starting checkpointing. I0425 17:41:19.454892 139654831384384 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:41:22.198355 139531676284672 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6 I0425 17:41:28.346435 139533798590208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 139.939 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.287750005722046 s) (per-host) I0425 17:41:28.346552 139533798590208 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.748367s. I0425 17:41:34.530970 139533798590208 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.932761s. I0425 17:41:34.531299 139533286897408 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:41:34.531435 139533286897408 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:41:34.531483 139533286897408 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:41:34.532976 139533286897408 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 17:41:34.533107 139654831384384 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0425 17:41:34.533188 139654831384384 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.078303 seconds. If this number is high, consider checkpointing less frequently. I0425 17:41:34.535376 139654831384384 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0425 17:41:34.539457 139654831384384 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/15. I0425 17:41:35.387437 139654831384384 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:41:35.387536 139654831384384 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0425 17:41:35.436009 139654831384384 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.050498s I0425 17:41:35.436196 139654831384384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.848 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3181583881378174 s) (per-host) I0425 17:41:35.436249 139654831384384 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.318221s (batch_requests_ready=0.262666s, total_serialization_initiated=0.055485s, others=0.000071s) I0425 17:41:35.436347 139654831384384 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.322985s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.322973) I0425 17:41:35.437209 139529495230208 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:41:35.437369 139654831384384 async_checkpointer.py:561] Finished blocking save. Time taken: 0.901920s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/15. I0425 17:41:35.439293 139654831384384 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:41:35.439605 139533286897408 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:41:35.439741 139654831384384 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138879.4548624, 'wait_for_prev_duration_secs': 15.078303098678589, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138894.535416, 'checkpointer_blocking_duration_secs': 0.9020636081695557, 'get_old_steps_start_time': 1777138895.4374993, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1777138879.4524956, 'checkpoint_manager_blocking_duration_secs': 15.98721170425415} I0425 17:41:35.439899 139654831384384 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0425 17:41:35.440730 139654831384384 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.340, Tokens/s/device: 13918.433, total_weights: 65536, loss: 5.858 I0425 17:41:35.448382 139654831384384 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.835, Tokens/s/device: 13842.327, total_weights: 65536, loss: 5.713 I0425 17:41:54.422869 139654831384384 metric_logger.py:185] completed step: 16, seconds: 16.283, TFLOP/s/device: 0.834, Tokens/s/device: 125.775, total_weights: 65536, loss: 5.599 I0425 17:41:54.570338 139654831384384 metric_logger.py:185] completed step: 17, seconds: 18.829, TFLOP/s/device: 0.722, Tokens/s/device: 108.767, total_weights: 65536, loss: 5.511 I0425 17:41:54.717803 139654831384384 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.567, Tokens/s/device: 13349.760, total_weights: 65536, loss: 5.442 I0425 17:41:54.866305 139654831384384 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0425 17:41:54.866958 139654831384384 checkpointing.py:798] Waited 0.0006651878356933594 seconds for step 19 to finish before starting checkpointing. I0425 17:41:54.868788 139654831384384 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:41:59.151149 139531676284672 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6 I0425 17:42:05.524284 139529495230208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 51.950 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 30.40620183944702 s) (per-host) I0425 17:42:05.524399 139529495230208 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.087079s. I0425 17:42:10.514514 139529495230208 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 35.077180s. I0425 17:42:10.514837 139533286897408 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:42:10.514953 139533286897408 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:42:10.515003 139533286897408 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:42:10.516467 139533286897408 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 17:42:10.516649 139654831384384 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0425 17:42:10.516767 139654831384384 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.647992 seconds. If this number is high, consider checkpointing less frequently. I0425 17:42:10.518598 139654831384384 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0425 17:42:10.521958 139654831384384 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/19. I0425 17:42:11.058362 139654831384384 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:42:11.058458 139654831384384 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0425 17:42:11.109221 139654831384384 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.053141s I0425 17:42:11.109391 139654831384384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.193 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29703307151794434 s) (per-host) I0425 17:42:11.109441 139654831384384 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.297092s (batch_requests_ready=0.239146s, total_serialization_initiated=0.057882s, others=0.000064s) I0425 17:42:11.109534 139654831384384 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.301132s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.301121) I0425 17:42:11.110416 139529520408320 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:42:11.110545 139654831384384 async_checkpointer.py:561] Finished blocking save. Time taken: 0.591878s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/19. I0425 17:42:11.127088 139654831384384 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:42:11.127400 139533286897408 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:42:11.127573 139654831384384 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138914.8687587, 'wait_for_prev_duration_secs': 15.647992134094238, 'time_between_consecutive_saves_sec': 20.335744619369507, 'checkpointer_blocking_start_time': 1777138930.5186365, 'checkpointer_blocking_duration_secs': 0.5920846462249756, 'get_old_steps_start_time': 1777138931.1107333, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1777138914.867193, 'checkpoint_manager_blocking_duration_secs': 16.26034903526306} I0425 17:42:11.127690 139654831384384 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0425 17:42:11.127732 139654831384384 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:42:16.169231 139531676284672 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6 I0425 17:42:21.882983 139529520408320 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 142.684 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.07058334350586 s) (per-host) I0425 17:42:21.883118 139529520408320 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.772474s. I0425 17:42:28.382878 139529520408320 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 17.272236s. I0425 17:42:28.383189 139533286897408 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:42:28.383313 139533286897408 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:42:28.383358 139533286897408 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:42:28.384665 139533286897408 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 17:42:28.384844 139654831384384 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0425 17:42:28.385015 139654831384384 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:42:28.385697 139654831384384 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.372, Tokens/s/device: 13923.259, total_weights: 65536, loss: 5.385 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 17:42:41 UTC 2026 EXIT_CODE=0