XPK Start: Mon Apr 20 18:40:17 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-20 18:40:41.079221: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 18:40:41.257552 134495762577216 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 18:40:50.298643 134495762577216 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 18:40:50.300961 134495762577216 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-utiyj-slice-job-0-0.mt-15-ocdbt-false-rs-utiyj:8482 I0420 18:40:51.920066 134495762577216 max_utils.py:284] Jax distributed system initialized! I0420 18:40:56.910958 134495762577216 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 18:40:56.911058 134495762577216 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 18:40:56.911115 134495762577216 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 I0420 18:40:56.911153 134495762577216 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 18:40:57.954266 134495762577216 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 18:40:58.067508 134495762577216 checkpointing.py:688] Setting up checkpoint logger... I0420 18:40:58.067617 134495762577216 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0420 18:40:58.067666 134495762577216 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 18:40:58.067875 134495762577216 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 0x7a521b318110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 18:40:59.402564 134495762577216 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 18:40:59.402794 134495762577216 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 0x7a3e7448c5c0>}, handler_registry=None I0420 18:40:59.403036 134495762577216 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3e7448c5c0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 18:40:59.403095 134495762577216 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a50f80e31d0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 18:40:59.403131 134495762577216 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 0x7a3e7448c5c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3e7448c5c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a50f80e31d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a50f80e31d0>}). I0420 18:40:59.403437 134495762577216 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 18:40:59.403507 134495762577216 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a3e4c5c44a0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 18:41:00.941943 134375542867712 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776710331320682006, 'commit_timestamp_nsecs': 1776710345430790261, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/9/_CHECKPOINT_METADATA I0420 18:41:01.199630 134375542867712 checkpoint_manager.py:1871] Missing metrics for step 9 I0420 18:41:01.605344 134495762577216 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints I0420 18:41:02.343820 134495762577216 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a50f8341cd0> I0420 18:41:02.343986 134495762577216 checkpointing.py:302] Checkpoint manager created! I0420 18:41:03.288039 134495762577216 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 18:41:03.288167 134495762577216 checkpointing.py:582] restoring from this run's directory step 9 I0420 18:41:03.536295 134495762577216 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/9. I0420 18:41:03.790614 134495762577216 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/9. I0420 18:41:03.791041 134495762577216 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/9. I0420 18:41:06.865986 1472 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 18:41:17.281253 134495762577216 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 132.836 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.89648699760437 s) (per-host) I0420 18:41:17.292527 134495762577216 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 128.410 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.324580430984497 s) (per-host) I0420 18:41:17.985199 134495762577216 event_tracking.py:138] [process=0] [sync] Finished load in 14.45 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/9 I0420 18:41:17.985352 134495762577216 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776710463.5361965, 'checkpointer_duration_secs': 14.449125051498413, 'checkpoint_manager_start_time': 1776710463.2887666, 'checkpoint_manager_duration_secs': 14.696556806564331} fsdp: 32 I0420 18:41:18.060555 134495762577216 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 18:41:18.060667 134495762577216 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 18:41:18.060718 134495762577216 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 18:41:18.060760 134495762577216 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 18:41:18.060796 134495762577216 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 18:41:18.060828 134495762577216 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 18:41:18.060863 134495762577216 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:41:18.060897 134495762577216 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 18:41:18.060930 134495762577216 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:41:18.060963 134495762577216 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:41:18.060995 134495762577216 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 18:41:18.061027 134495762577216 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 18:41:18.461340 134495762577216 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 18:41:18.461434 134495762577216 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:41:18.466888 134495762577216 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 18:41:18.466942 134495762577216 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:41:18.483073 134495762577216 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 18:41:18.483148 134495762577216 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:41:18.498720 134495762577216 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 18:41:18.498779 134495762577216 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:41:18.522574 134495762577216 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:41:18.522642 134495762577216 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:41:18.538299 134495762577216 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:41:18.538361 134495762577216 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:41:18.553934 134495762577216 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:41:18.553995 134495762577216 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:41:18.583972 134495762577216 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 18:41:18.584041 134495762577216 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:41:18.603864 134495762577216 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 18:41:18.603925 134495762577216 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 18:41:29.925714 134495762577216 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0420 18:41:29.936819 134495762577216 metric_logger.py:301] number parameters: 1.104 billion I0420 18:41:41.379353 134495762577216 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 18:41:41.531712 134495762577216 checkpointing.py:798] Waited 0.1523456573486328 seconds for step 10 to finish before starting checkpointing. I0420 18:41:41.534021 134495762577216 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:41:41.536270 134495762577216 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0420 18:41:41.537976 134495762577216 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10. I0420 18:41:41.872662 134495762577216 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 18:41:41.889299 134495762577216 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:41:41.889418 134495762577216 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 18:41:42.157506 134371426289408 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10 I0420 18:41:42.159061 134495762577216 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271230s I0420 18:41:42.159962 134495762577216 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.512 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27986836433410645 s) (per-host) I0420 18:41:42.160026 134495762577216 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.279950s (batch_requests_ready=0.002573s, total_serialization_initiated=0.276578s, others=0.000798s) I0420 18:41:42.160117 134495762577216 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283785s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.283768) I0420 18:41:42.160853 134495762577216 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10. I0420 18:41:42.161171 134372522968832 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 19:01:42.161139 I0420 18:41:42.177122 134495762577216 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:41:42.177422 134370359187200 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:41:42.177577 134495762577216 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776710501.5340035, 'wait_for_prev_duration_secs': 5.91278076171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776710501.5363512, 'checkpointer_blocking_duration_secs': 0.6249661445617676, 'get_old_steps_start_time': 1776710502.1613412, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776710501.5320675, 'checkpoint_manager_blocking_duration_secs': 0.6454746723175049} I0420 18:41:42.177682 134495762577216 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 18:41:42.177744 134495762577216 max_utils.py:750] Memstats: After params initialized: I0420 18:41:42.177791 134495762577216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 18:41:42.177821 134495762577216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 18:41:42.177847 134495762577216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 18:41:42.177869 134495762577216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 18:41:42.547453 134495762577216 metric_logger.py:196] completed step: 10, seconds: 11.127, TFLOP/s/device: 1.221, Tokens/s/device: 184.063, total_weights: 65536, loss: 6.827, lm_loss: 6.827, perplexity: 922.537 I0420 18:41:42.688469 134495762577216 metric_logger.py:196] completed step: 11, seconds: 1.164, TFLOP/s/device: 11.668, Tokens/s/device: 1758.773, total_weights: 65536, loss: 6.522, lm_loss: 6.522, perplexity: 679.821 I0420 18:41:42.836064 134495762577216 metric_logger.py:196] completed step: 12, seconds: 0.014, TFLOP/s/device: 976.087, Tokens/s/device: 147126.437, total_weights: 65536, loss: 6.260, lm_loss: 6.260, perplexity: 523.013 I0420 18:41:42.984093 134495762577216 metric_logger.py:196] completed step: 13, seconds: 0.142, TFLOP/s/device: 95.445, Tokens/s/device: 14386.468, total_weights: 65536, loss: 6.039, lm_loss: 6.039, perplexity: 419.520 I0420 18:41:42.997167 134495762577216 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0420 18:41:43.278628 134495762577216 checkpointing.py:798] Waited 0.28142619132995605 seconds for step 15 to finish before starting checkpointing. I0420 18:41:43.281367 134495762577216 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 18:41:43.987197 134370905470720 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/items I0420 18:41:44.129541 134370350794496 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776710503731643244, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776710503731643244, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0420 18:41:46.927793 134370888685312 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_0 I0420 18:41:56.039624 134370880292608 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 10.936894s (commit=9.837126s, array_metadata_write=1.099768s) I0420 18:41:56.041031 134372522968832 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 111.547 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 14.160888195037842 s) (per-host) I0420 18:41:56.041214 134372522968832 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 13.879929s. W0420 18:42:02.650151 134372522968832 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/items, I0420 18:42:02.726993 134370350794496 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776710503731643244, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0420 18:42:03.465938 134372522968832 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000743s. I0420 18:42:03.466229 134372522968832 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.555852s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/items I0420 18:42:03.468508 134372522968832 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10/items I0420 18:42:04.013276 134372522968832 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10 I0420 18:42:05.409745 134372522968832 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10`. I0420 18:42:05.410932 134372522968832 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 23.87 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/10 I0420 18:42:05.412660 134372522968832 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 23.251371s. I0420 18:42:05.412864 134370359187200 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:42:05.412972 134370359187200 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:42:05.413070 134370359187200 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:42:05.415061 134370359187200 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 18:42:05.415380 134495762577216 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. W0420 18:42:05.415542 134495762577216 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.134192 seconds. If this number is high, consider checkpointing less frequently. I0420 18:42:05.417386 134495762577216 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0420 18:42:05.419906 134495762577216 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15. I0420 18:42:06.141002 134495762577216 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:42:06.141152 134495762577216 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 18:42:06.171560 134495762577216 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.033673s I0420 18:42:06.172508 134495762577216 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 29.139 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05293893814086914 s) (per-host) I0420 18:42:06.172622 134495762577216 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.053075s (batch_requests_ready=0.006260s, total_serialization_initiated=0.045969s, others=0.000845s) I0420 18:42:06.172754 134495762577216 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.057936s (all_items=0.000022s, per_item={'items': '0.00002241'}, temp_paths=0.057913) I0420 18:42:06.173815 134495762577216 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15. I0420 18:42:06.174262 134370359187200 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 19:02:06.174217 I0420 18:42:06.176736 134495762577216 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:42:06.177032 134370880292608 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:42:06.177230 134495762577216 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776710503.281323, 'wait_for_prev_duration_secs': 22.13419222831726, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776710525.4174526, 'checkpointer_blocking_duration_secs': 0.7569448947906494, 'get_old_steps_start_time': 1776710526.174427, 'get_old_steps_duration_secs': 4.315376281738281e-05, 'checkpoint_manager_blocking_start_time': 1776710503.2791448, 'checkpoint_manager_blocking_duration_secs': 22.89803457260132} I0420 18:42:06.177378 134495762577216 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0420 18:42:06.360383 134372522968832 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15 I0420 18:42:08.148455 134370888685312 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15/items I0420 18:42:11.164662 134370937534208 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_0 I0420 18:42:19.010323 134370897078016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 9.684701s (commit=8.485672s, array_metadata_write=1.199029s) I0420 18:42:19.011448 134370359187200 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 122.527 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.89186692237854 s) (per-host) I0420 18:42:19.011561 134370359187200 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 12.836488s. W0420 18:42:20.529814 134370359187200 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15/items, I0420 18:42:21.365512 134370359187200 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000404s. I0420 18:42:21.365741 134370359187200 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.570280s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15/items I0420 18:42:21.367146 134370359187200 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15/items I0420 18:42:21.877132 134370359187200 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15 I0420 18:42:23.263130 134370359187200 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15`. I0420 18:42:23.263960 134370359187200 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 17.85 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/15 I0420 18:42:23.265066 134370359187200 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 17.089992s. I0420 18:42:23.265211 134370880292608 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:42:23.265289 134370880292608 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:42:23.265355 134370880292608 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:42:23.267706 134370880292608 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0420 18:42:27.835074 134495762577216 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.237, Tokens/s/device: 13902.937, total_weights: 65536, loss: 5.858, lm_loss: 5.858, perplexity: 350.067 I0420 18:42:27.852052 134495762577216 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.691, Tokens/s/device: 13820.655, total_weights: 65536, loss: 5.713, lm_loss: 5.713, perplexity: 302.780 I0420 18:42:27.990643 134495762577216 metric_logger.py:196] completed step: 16, seconds: 44.854, TFLOP/s/device: 0.303, Tokens/s/device: 45.659, total_weights: 65536, loss: 5.599, lm_loss: 5.599, perplexity: 270.201 I0420 18:42:28.137900 134495762577216 metric_logger.py:196] completed step: 17, seconds: 0.009, TFLOP/s/device: 1463.499, Tokens/s/device: 220594.571, total_weights: 65536, loss: 5.511, lm_loss: 5.511, perplexity: 247.443 I0420 18:42:28.285628 134495762577216 metric_logger.py:196] completed step: 18, seconds: 0.140, TFLOP/s/device: 97.213, Tokens/s/device: 14652.958, total_weights: 65536, loss: 5.442, lm_loss: 5.442, perplexity: 230.909 I0420 18:42:28.433775 134495762577216 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 18:42:28.434491 134495762577216 checkpointing.py:798] Waited 0.0007295608520507812 seconds for step 19 to finish before starting checkpointing. I0420 18:42:28.436544 134495762577216 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:42:28.438241 134495762577216 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0420 18:42:28.439679 134495762577216 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19. I0420 18:42:28.725852 134495762577216 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:42:28.725946 134495762577216 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 18:42:28.753351 134495762577216 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029185s I0420 18:42:28.754160 134495762577216 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.995 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0376286506652832 s) (per-host) I0420 18:42:28.754238 134495762577216 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037728s (batch_requests_ready=0.002505s, total_serialization_initiated=0.034485s, others=0.000738s) I0420 18:42:28.754322 134495762577216 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041726s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.041708) I0420 18:42:28.755108 134495762577216 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.32 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19. I0420 18:42:28.755371 134370880292608 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 19:02:28.755347 I0420 18:42:28.757274 134495762577216 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:42:28.757496 134370897078016 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:42:28.757594 134495762577216 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776710548.4365118, 'wait_for_prev_duration_secs': 7.033348083496094e-05, 'time_between_consecutive_saves_sec': 5.168765544891357, 'checkpointer_blocking_start_time': 1776710548.4382796, 'checkpointer_blocking_duration_secs': 0.3171885013580322, 'get_old_steps_start_time': 1776710548.7554872, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776710548.43473, 'checkpoint_manager_blocking_duration_secs': 0.32283496856689453} I0420 18:42:28.757777 134495762577216 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 18:42:28.757826 134495762577216 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 18:42:28.980428 134370359187200 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19 I0420 18:42:31.070031 134367664404224 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19/items I0420 18:42:34.030597 134370937534208 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_0 I0420 18:42:41.424071 134370905470720 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 9.151416s (commit=7.751587s, array_metadata_write=1.399829s) I0420 18:42:41.425212 134370880292608 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 124.293 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.708652257919312 s) (per-host) I0420 18:42:41.425333 134370880292608 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 12.669886s. W0420 18:42:42.948379 134370880292608 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19/items, I0420 18:42:43.406980 134370880292608 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000450s. I0420 18:42:43.407226 134370880292608 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.186035s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19/items I0420 18:42:43.408659 134370880292608 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19/items I0420 18:42:43.926321 134370880292608 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19 I0420 18:42:45.300716 134370880292608 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19`. I0420 18:42:45.301557 134370880292608 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 16.86 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_false/checkpoints/19 I0420 18:42:45.302842 134370880292608 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 16.547391s. I0420 18:42:45.303052 134370897078016 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:42:45.303141 134370897078016 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:42:45.303200 134370897078016 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:42:45.305938 134370897078016 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 18:42:45.306145 134495762577216 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. I0420 18:42:45.306252 134495762577216 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:42:45.307276 134495762577216 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.337, Tokens/s/device: 13918.055, total_weights: 65536, loss: 5.385, lm_loss: 5.385, perplexity: 218.045 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 18:42:58 UTC 2026 EXIT_CODE=0