XPK Start: Fri Apr 17 02:10:09 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-17 02:10:33.633967: 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) I0417 02:10:33.812485 136123904083776 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 02:10:42.852732 136123904083776 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 02:10:42.855067 136123904083776 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-k669n-slice-job-0-0.mt-15-ocdbt-false-rs-k669n:8482 I0417 02:10:44.010236 136123904083776 max_utils.py:284] Jax distributed system initialized! I0417 02:10:50.187259 136123904083776 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 02:10:50.187363 136123904083776 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 02:10:50.187403 136123904083776 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0417 02:10:50.187436 136123904083776 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 02:10:50.852257 136123904083776 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 02:10:50.966324 136123904083776 checkpointing.py:688] Setting up checkpoint logger... I0417 02:10:50.966441 136123904083776 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0417 02:10:50.966485 136123904083776 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 02:10:50.966678 136123904083776 base_pytree_checkpoint_handler.py:441] 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 0x7bcd1924d010>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 02:10:52.696565 136123904083776 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 02:10:52.696800 136123904083776 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bb8cc1c4f20>}, handler_registry=None I0417 02:10:52.697052 136123904083776 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bb8cc1c4f20>` 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`. I0417 02:10:52.697105 136123904083776 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bb8cc6eee70>` 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`. I0417 02:10:52.697141 136123904083776 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 0x7bb8cc1c4f20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bb8cc1c4f20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bb8cc6eee70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bb8cc6eee70>}). I0417 02:10:52.697464 136123904083776 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 02:10:52.697535 136123904083776 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bb8cc632020> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 02:10:54.580118 136002261731072 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776391714990288741, 'commit_timestamp_nsecs': 1776391729295421279, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/9/_CHECKPOINT_METADATA I0417 02:10:54.828863 136002270123776 checkpoint_manager.py:1871] Missing metrics for step 5 I0417 02:10:55.244937 136123904083776 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints I0417 02:10:56.398929 136123904083776 checkpoint_manager.py:929] [process=0][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_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bb8cc6adf70> I0417 02:10:56.399107 136123904083776 checkpointing.py:302] Checkpoint manager created! I0417 02:10:57.346830 136123904083776 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 02:10:57.346955 136123904083776 checkpointing.py:582] restoring from this run's directory step 9 I0417 02:10:57.594253 136123904083776 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/9. I0417 02:10:57.856775 136123904083776 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/9. I0417 02:10:57.857220 136123904083776 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.26 seconds. Continuing load @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/9. I0417 02:11:00.821039 1484 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 02:11:10.142281 136123904083776 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 146.993 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.750688076019287 s) (per-host) I0417 02:11:10.152918 136123904083776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 141.529 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.182211875915527 s) (per-host) I0417 02:11:13.110667 136123904083776 event_tracking.py:138] [process=0] [sync] Finished load in 15.52 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/9 I0417 02:11:13.110851 136123904083776 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776391857.5941534, 'checkpointer_duration_secs': 15.516640663146973, 'checkpoint_manager_start_time': 1776391857.3475714, 'checkpoint_manager_duration_secs': 15.763224601745605} fsdp: 32 I0417 02:11:13.185985 136123904083776 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0417 02:11:13.186097 136123904083776 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 02:11:13.186148 136123904083776 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 02:11:13.186189 136123904083776 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0417 02:11:13.186225 136123904083776 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 02:11:13.186258 136123904083776 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 02:11:13.186293 136123904083776 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 02:11:13.186327 136123904083776 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0417 02:11:13.186360 136123904083776 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 02:11:13.186392 136123904083776 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 02:11:13.186425 136123904083776 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0417 02:11:13.186456 136123904083776 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0417 02:11:13.587316 136123904083776 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 02:11:13.587412 136123904083776 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 02:11:13.592900 136123904083776 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 02:11:13.592955 136123904083776 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 02:11:13.609653 136123904083776 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 02:11:13.609713 136123904083776 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 02:11:13.625302 136123904083776 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 02:11:13.625360 136123904083776 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 02:11:13.649283 136123904083776 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 02:11:13.649349 136123904083776 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 02:11:13.665077 136123904083776 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 02:11:13.665143 136123904083776 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 02:11:13.680737 136123904083776 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 02:11:13.680794 136123904083776 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 02:11:13.710868 136123904083776 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 02:11:13.710940 136123904083776 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 02:11:13.730744 136123904083776 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 02:11:13.730806 136123904083776 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 02:11:24.775114 136123904083776 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. I0417 02:11:24.784351 136123904083776 metric_logger.py:289] number parameters: 1.104 billion I0417 02:11:36.184495 136123904083776 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0417 02:11:36.337130 136123904083776 checkpointing.py:798] Waited 0.15262389183044434 seconds for step 10 to finish before starting checkpointing. I0417 02:11:36.339684 136123904083776 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 02:11:36.341222 136123904083776 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0417 02:11:36.342690 136123904083776 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10. I0417 02:11:36.740519 136123904083776 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 02:11:36.756853 136123904083776 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 02:11:36.756978 136123904083776 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 I0417 02:11:36.986621 135997945792256 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10 I0417 02:11:37.026379 136123904083776 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.270921s I0417 02:11:37.027226 136123904083776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.522 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2793433666229248 s) (per-host) I0417 02:11:37.027311 136123904083776 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.279444s (batch_requests_ready=0.002501s, total_serialization_initiated=0.276174s, others=0.000769s) I0417 02:11:37.027419 136123904083776 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283410s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.283394) I0417 02:11:37.028122 136123904083776 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.69 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10. I0417 02:11:37.028522 135998479156992 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 02:31:37.028484 I0417 02:11:37.046585 136123904083776 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0417 02:11:37.046929 135997421500160 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 02:11:37.047085 136123904083776 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776391896.339665, 'wait_for_prev_duration_secs': 6.175041198730469e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776391896.3412597, 'checkpointer_blocking_duration_secs': 0.6874127388000488, 'get_old_steps_start_time': 1776391897.0286934, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776391896.3374877, 'checkpoint_manager_blocking_duration_secs': 0.7095634937286377} I0417 02:11:37.047197 136123904083776 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0417 02:11:37.047263 136123904083776 max_utils.py:750] Memstats: After params initialized: I0417 02:11:37.047312 136123904083776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 02:11:37.047345 136123904083776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 02:11:37.047371 136123904083776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 02:11:37.047394 136123904083776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 02:11:37.399615 136123904083776 metric_logger.py:185] completed step: 10, seconds: 11.098, TFLOP/s/device: 1.224, Tokens/s/device: 184.546, total_weights: 65536, loss: 6.827 I0417 02:11:37.540564 136123904083776 metric_logger.py:185] completed step: 11, seconds: 1.212, TFLOP/s/device: 11.213, Tokens/s/device: 1690.116, total_weights: 65536, loss: 6.522 I0417 02:11:37.688589 136123904083776 metric_logger.py:185] completed step: 12, seconds: 0.013, TFLOP/s/device: 1007.872, Tokens/s/device: 151917.514, total_weights: 65536, loss: 6.260 I0417 02:11:37.835978 136123904083776 metric_logger.py:185] completed step: 13, seconds: 0.142, TFLOP/s/device: 95.671, Tokens/s/device: 14420.504, total_weights: 65536, loss: 6.039 I0417 02:11:37.849183 136123904083776 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0417 02:11:38.131486 136123904083776 checkpointing.py:798] Waited 0.2822716236114502 seconds for step 15 to finish before starting checkpointing. I0417 02:11:38.134235 136123904083776 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 02:11:38.918742 135997455070976 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/items I0417 02:11:39.085163 135997413107456 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776391898662631719, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776391898662631719, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0417 02:11:41.966037 135997438285568 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_0 I0417 02:11:50.675732 135997429892864 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 10.660523s (commit=9.152601s, array_metadata_write=1.507922s) I0417 02:11:50.677151 135998479156992 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 113.402 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.92922830581665 s) (per-host) I0417 02:11:50.677321 135998479156992 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 13.648684s. I0417 02:11:52.099518 135997413107456 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776391898662631719, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA W0417 02:11:52.184638 135998479156992 ocdbt_utils.py:168] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/items, I0417 02:11:53.033311 135998479156992 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000771s. I0417 02:11:53.033593 135998479156992 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.559971s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/items I0417 02:11:53.035888 135998479156992 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10/items I0417 02:11:53.537360 135998479156992 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10 I0417 02:11:54.940342 135998479156992 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10`. I0417 02:11:54.941518 135998479156992 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 18.60 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/10 I0417 02:11:54.942999 135998479156992 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 17.914360s. I0417 02:11:54.943299 135997421500160 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 02:11:54.943463 135997421500160 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 02:11:54.943568 135997421500160 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0417 02:11:54.945200 135997421500160 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0417 02:11:54.945490 136123904083776 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0417 02:11:54.945658 136123904083776 checkpoint_manager.py:1452] Waiting for previous save to complete took 16.811439 seconds. If this number is high, consider checkpointing less frequently. I0417 02:11:54.947396 136123904083776 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0417 02:11:54.949691 136123904083776 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15. I0417 02:11:55.250907 136123904083776 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 02:11:55.251038 136123904083776 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 I0417 02:11:55.282577 136123904083776 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034682s I0417 02:11:55.283571 136123904083776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 28.703 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.053743600845336914 s) (per-host) I0417 02:11:55.283680 136123904083776 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.053876s (batch_requests_ready=0.006302s, total_serialization_initiated=0.046679s, others=0.000895s) I0417 02:11:55.283815 136123904083776 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.058864s (all_items=0.000023s, per_item={'items': '0.00002289'}, temp_paths=0.058841) I0417 02:11:55.284830 136123904083776 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15. I0417 02:11:55.285293 135997421500160 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 02:31:55.285250 I0417 02:11:55.287689 136123904083776 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0417 02:11:55.288033 135997429892864 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 02:11:55.288216 136123904083776 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776391898.1341877, 'wait_for_prev_duration_secs': 16.81143856048584, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776391914.9474602, 'checkpointer_blocking_duration_secs': 0.3380167484283447, 'get_old_steps_start_time': 1776391915.2855043, 'get_old_steps_duration_secs': 4.601478576660156e-05, 'checkpoint_manager_blocking_start_time': 1776391898.131994, 'checkpoint_manager_blocking_duration_secs': 17.156168699264526} I0417 02:11:55.288371 136123904083776 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0417 02:11:55.460200 135998479156992 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15 I0417 02:11:57.686910 135997438285568 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15/items I0417 02:12:00.987349 135997455070976 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_0 I0417 02:12:09.184806 135997446678272 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 9.974527s (commit=8.881325s, array_metadata_write=1.093202s) I0417 02:12:09.186050 135997421500160 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 113.183 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.95620059967041 s) (per-host) I0417 02:12:09.186185 135997421500160 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 13.900748s. W0417 02:12:11.137647 135997421500160 ocdbt_utils.py:168] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15/items, I0417 02:12:11.570703 135997421500160 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000454s. I0417 02:12:11.570965 135997421500160 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.182472s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15/items I0417 02:12:11.572342 135997421500160 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15/items I0417 02:12:12.101759 135997421500160 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15 I0417 02:12:13.941049 135997421500160 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15`. I0417 02:12:13.941929 135997421500160 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 18.99 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/15 I0417 02:12:13.943591 135997421500160 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 18.658153s. I0417 02:12:13.943742 135997429892864 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 02:12:13.943801 135997429892864 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 02:12:13.943876 135997429892864 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0417 02:12:13.945590 135997429892864 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0417 02:12:16.611763 136123904083776 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.048, Tokens/s/device: 13874.493, total_weights: 65536, loss: 5.858 I0417 02:12:16.622825 136123904083776 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.761, Tokens/s/device: 13831.203, total_weights: 65536, loss: 5.713 I0417 02:12:16.766520 136123904083776 metric_logger.py:185] completed step: 16, seconds: 38.773, TFLOP/s/device: 0.350, Tokens/s/device: 52.820, total_weights: 65536, loss: 5.599 I0417 02:12:16.914273 136123904083776 metric_logger.py:185] completed step: 17, seconds: 0.009, TFLOP/s/device: 1541.190, Tokens/s/device: 232304.900, total_weights: 65536, loss: 5.511 I0417 02:12:17.062125 136123904083776 metric_logger.py:185] completed step: 18, seconds: 0.145, TFLOP/s/device: 93.578, Tokens/s/device: 14105.072, total_weights: 65536, loss: 5.442 I0417 02:12:17.210631 136123904083776 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0417 02:12:17.211433 136123904083776 checkpointing.py:798] Waited 0.0008111000061035156 seconds for step 19 to finish before starting checkpointing. I0417 02:12:17.213795 136123904083776 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 02:12:17.215516 136123904083776 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0417 02:12:17.217014 136123904083776 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19. I0417 02:12:17.980098 136123904083776 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 02:12:17.980198 136123904083776 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 I0417 02:12:18.006280 136123904083776 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027637s I0417 02:12:18.006956 136123904083776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 42.991 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03588104248046875 s) (per-host) I0417 02:12:18.007019 136123904083776 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.035961s (batch_requests_ready=0.002532s, total_serialization_initiated=0.032800s, others=0.000628s) I0417 02:12:18.007088 136123904083776 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.039999s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.039982) I0417 02:12:18.007776 136123904083776 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.79 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19. I0417 02:12:18.008121 135997429892864 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 02:32:18.008085 I0417 02:12:18.010051 136123904083776 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0417 02:12:18.010263 135997438285568 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 02:12:18.010352 136123904083776 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776391937.2137628, 'wait_for_prev_duration_secs': 9.202957153320312e-05, 'time_between_consecutive_saves_sec': 3.268131971359253, 'checkpointer_blocking_start_time': 1776391937.2155573, 'checkpointer_blocking_duration_secs': 0.7927019596099854, 'get_old_steps_start_time': 1776391938.0082781, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776391937.2116854, 'checkpoint_manager_blocking_duration_secs': 0.7986366748809814} I0417 02:12:18.010496 136123904083776 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0417 02:12:18.010540 136123904083776 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 02:12:18.208911 135997421500160 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19 I0417 02:12:19.919299 135998479156992 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19/items I0417 02:12:22.793703 135997455070976 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_0 I0417 02:12:29.624826 135997446678272 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 8.590031s (commit=7.512436s, array_metadata_write=1.077595s) I0417 02:12:29.625982 135997429892864 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 135.531 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.654870510101318 s) (per-host) I0417 02:12:29.626112 135997429892864 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.617875s. W0417 02:12:31.481303 135997429892864 ocdbt_utils.py:168] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19/items, I0417 02:12:31.936614 135997429892864 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000437s. I0417 02:12:31.936858 135997429892864 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.572813s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19/items I0417 02:12:31.938278 135997429892864 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19/items I0417 02:12:32.453695 135997429892864 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19 I0417 02:12:33.868015 135997429892864 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19`. I0417 02:12:33.868823 135997429892864 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 16.65 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_15_ocdbt_false/checkpoints/19 I0417 02:12:33.870187 135997429892864 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 15.861948s. I0417 02:12:33.870378 135997438285568 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 02:12:33.870490 135997438285568 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 02:12:33.870541 135997438285568 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0417 02:12:33.872486 135997438285568 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0417 02:12:33.872690 136123904083776 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0417 02:12:33.872796 136123904083776 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 02:12:33.873495 136123904083776 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.294, Tokens/s/device: 13911.626, 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: Fri Apr 17 02:12:45 UTC 2026 EXIT_CODE=0