XPK Start: Sat Apr 18 00:41:41 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 00:42:05.431623: 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) I0418 00:42:05.967173 140660516149056 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 00:42:15.006942 140660516149056 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 00:42:15.009199 140660516149056 distributed.py:172] Connecting to JAX distributed service on mt-09-pdb-lt-1-th84e-slice-job-0-0.mt-09-pdb-lt-1-th84e:8482 I0418 00:42:16.406337 140660516149056 max_utils.py:284] Jax distributed system initialized! I0418 00:42:22.334671 140660516149056 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 00:42:22.334772 140660516149056 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 00:42:22.334813 140660516149056 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 I0418 00:42:22.334882 140660516149056 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 00:42:23.020544 140660516149056 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 8, 1, 1, 1, 1, 4, 1, 1, 1, 1) I0418 00:42:23.020850 140660516149056 checkpointing.py:677] Setting up checkpoint logger... I0418 00:42:23.020920 140660516149056 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 00:42:23.020986 140660516149056 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 00:42:23.021344 140660516149056 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7fed5c93c0b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 00:42:26.806202 140660516149056 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 00:42:26.806427 140660516149056 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 0x7fec507b3050>}, handler_registry=None I0418 00:42:26.806662 140660516149056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fec507b3050>` 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`. I0418 00:42:26.806709 140660516149056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd930245fa0>` 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`. I0418 00:42:26.806742 140660516149056 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 0x7fec507b3050>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fec507b3050>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd930245fa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd930245fa0>}). I0418 00:42:26.807088 140660516149056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 00:42:26.807156 140660516149056 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7fd92037fce0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 00:42:28.363485 140660516149056 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints I0418 00:42:28.797932 140660516149056 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=10), 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/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7fd93023b980> I0418 00:42:28.798101 140660516149056 checkpointing.py:301] Checkpoint manager created! I0418 00:42:29.720165 140660516149056 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 00:42:29.720276 140660516149056 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0418 00:42:30.099349 140660516149056 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 00:42:30.099443 140660516149056 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0418 00:42:30.115686 140660516149056 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 00:42:30.115746 140660516149056 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0418 00:42:30.139509 140660516149056 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 00:42:30.139581 140660516149056 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0418 00:42:30.155929 140660516149056 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 00:42:30.155988 140660516149056 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0418 00:42:30.172251 140660516149056 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 00:42:30.172308 140660516149056 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0418 00:42:30.197800 140660516149056 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 00:42:30.197888 140660516149056 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0418 00:42:30.218880 140660516149056 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 00:42:30.218946 140660516149056 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, 'tensor'). I0418 00:42:30.434271 140660516149056 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 00:42:30.434376 140660516149056 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 8 tensor: 4 I0418 00:42:31.866827 140660516149056 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: ('tensor',) I0418 00:42:31.867006 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, 'tensor') I0418 00:42:31.867063 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, 'tensor') I0418 00:42:31.867117 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: ('tensor', None, 'fsdp') I0418 00:42:31.867170 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: ('tensor', None) I0418 00:42:31.867209 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: ('tensor', None) I0418 00:42:31.867260 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, 'tensor', None) I0418 00:42:31.867311 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: ('tensor', None, None, 'fsdp') I0418 00:42:31.867349 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, 'tensor', None) I0418 00:42:31.867387 140660516149056 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, 'tensor', None) I0418 00:42:31.867434 140660516149056 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', 'tensor') I0418 00:42:31.867481 140660516149056 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: ('tensor', 'fsdp') I0418 00:42:32.357071 140660516149056 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0418 00:42:32.357168 140660516149056 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0418 00:42:32.373220 140660516149056 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0418 00:42:32.373283 140660516149056 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0418 00:42:47.361255 140660516149056 max_utils.py:791] Total memory size: 0.9 GB, Output size: 0.4 GB, Temp size: 0.5 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0418 00:42:47.369921 140660516149056 metric_logger.py:301] number parameters: 1.104 billion I0418 00:42:48.884906 140660516149056 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0418 00:43:03.007489 140660516149056 checkpointing.py:776] Waited 14.122562885284424 seconds for step 0 to finish before starting checkpointing. I0418 00:43:03.010115 140660516149056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 00:43:03.012069 140660516149056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0418 00:43:03.013520 140660516149056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0. I0418 00:43:03.753992 140660516149056 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 00:43:03.768248 140660516149056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 00:43:03.768318 140660516149056 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 I0418 00:43:04.030367 140660516149056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.262835s I0418 00:43:04.031170 140660516149056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.729 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2692532539367676 s) (per-host) I0418 00:43:04.031243 140660516149056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.269344s (batch_requests_ready=0.002203s, total_serialization_initiated=0.266424s, others=0.000717s) I0418 00:43:04.031322 140660516149056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.273352s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.273334) I0418 00:43:04.032134 140660516149056 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0. I0418 00:43:04.032466 140532713256704 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:03:04.032428 I0418 00:43:04.042885 140660516149056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 00:43:04.043206 140532160374528 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 00:43:04.043360 140660516149056 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776472983.0100975, 'wait_for_prev_duration_secs': 6.270408630371094e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776472983.0121088, 'checkpointer_blocking_duration_secs': 1.020500659942627, 'get_old_steps_start_time': 1776472984.0326347, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776472983.008082, 'checkpoint_manager_blocking_duration_secs': 1.0352404117584229} I0418 00:43:04.043484 140660516149056 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0418 00:43:04.043537 140660516149056 max_utils.py:750] Memstats: After params initialized: I0418 00:43:04.043586 140660516149056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0418 00:43:04.043619 140660516149056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0418 00:43:04.043646 140660516149056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0418 00:43:04.043670 140660516149056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0418 00:43:04.386957 140660516149056 metric_logger.py:196] completed step: 0, seconds: 1.159, TFLOP/s/device: 2.931, Tokens/s/device: 441.734, total_weights: 16384, loss: 10.862, lm_loss: 10.862, perplexity: 52153.723 I0418 00:43:04.390270 140660516149056 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/tensorboard/' I0418 00:43:04.478700 140532688852736 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0 I0418 00:43:05.237115 140660516149056 metric_logger.py:196] completed step: 1, seconds: 15.500, TFLOP/s/device: 0.219, Tokens/s/device: 33.032, total_weights: 16384, loss: 10.862, lm_loss: 10.862, perplexity: 52153.723 I0418 00:43:05.293997 140660516149056 metric_logger.py:196] completed step: 2, seconds: 0.851, TFLOP/s/device: 3.992, Tokens/s/device: 601.713, total_weights: 16384, loss: 9.763, lm_loss: 9.763, perplexity: 17373.129 I0418 00:43:05.352353 140660516149056 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 356.132, Tokens/s/device: 53680.017, total_weights: 16384, loss: 8.823, lm_loss: 8.823, perplexity: 6791.817 I0418 00:43:07.563505 140532193945344 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/items I0418 00:43:07.724115 140531661969152 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776472986882770695, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776472986882770695, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/_CHECKPOINT_METADATA I0418 00:43:09.089818 2568 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 00:43:11.284914 140532177159936 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/items/array_metadatas/process_0 I0418 00:43:20.358209 140660516149056 metric_logger.py:196] completed step: 4, seconds: 0.057, TFLOP/s/device: 59.449, Tokens/s/device: 8960.762, total_weights: 16384, loss: 7.947, lm_loss: 7.947, perplexity: 2828.417 I0418 00:43:20.366909 140660516149056 metric_logger.py:196] completed step: 5, seconds: 0.059, TFLOP/s/device: 57.873, Tokens/s/device: 8723.209, total_weights: 16384, loss: 7.212, lm_loss: 7.212, perplexity: 1355.339 I0418 00:43:20.424019 140660516149056 metric_logger.py:196] completed step: 6, seconds: 15.005, TFLOP/s/device: 0.226, Tokens/s/device: 34.123, total_weights: 16384, loss: 6.652, lm_loss: 6.652, perplexity: 774.659 I0418 00:43:20.482424 140660516149056 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 525.818, Tokens/s/device: 79256.966, total_weights: 16384, loss: 6.269, lm_loss: 6.269, perplexity: 527.696 I0418 00:43:20.540942 140660516149056 metric_logger.py:196] completed step: 8, seconds: 0.058, TFLOP/s/device: 58.580, Tokens/s/device: 8829.870, total_weights: 16384, loss: 6.031, lm_loss: 6.031, perplexity: 416.165 I0418 00:43:20.598634 140660516149056 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0418 00:43:20.599238 140660516149056 checkpointing.py:776] Waited 0.0006194114685058594 seconds for step 9 to finish before starting checkpointing. I0418 00:43:20.601304 140660516149056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 00:43:46.586417 140532168767232 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.505151s (commit=35.040705s, array_metadata_write=2.464445s) I0418 00:43:46.587571 140532713256704 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 36.884 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.8256139755249 s) (per-host) I0418 00:43:46.587721 140532713256704 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.555143s. I0418 00:43:47.806017 140532713256704 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776472986882770695, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/_CHECKPOINT_METADATA I0418 00:43:49.000264 140531661969152 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776472986882770695, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/_CHECKPOINT_METADATA I0418 00:43:49.216007 140532713256704 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 00:43:49.782738 140532713256704 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000245s. I0418 00:43:50.144161 140532713256704 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.216405s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/items I0418 00:43:50.147067 140532713256704 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0/items I0418 00:43:51.588744 140532713256704 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0 I0418 00:43:54.264778 140532713256704 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0`. I0418 00:43:54.265614 140532713256704 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.25 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/0 I0418 00:43:54.266959 140532713256704 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.234382s. I0418 00:43:54.267148 140532160374528 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 00:43:54.267259 140532160374528 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 00:43:54.267325 140532160374528 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 00:43:54.269268 140532160374528 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 00:43:54.269401 140660516149056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0418 00:43:54.269490 140660516149056 checkpoint_manager.py:1452] Waiting for previous save to complete took 33.668190 seconds. If this number is high, consider checkpointing less frequently. I0418 00:43:54.271444 140660516149056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0418 00:43:54.273347 140660516149056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9. I0418 00:43:54.989269 140660516149056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 00:43:54.989359 140660516149056 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 I0418 00:43:55.018195 140660516149056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029744s I0418 00:43:55.018929 140660516149056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 44.863 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03438401222229004 s) (per-host) I0418 00:43:55.018996 140660516149056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.034468s (batch_requests_ready=0.001986s, total_serialization_initiated=0.031802s, others=0.000680s) I0418 00:43:55.019071 140660516149056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.038734s (all_items=0.000025s, per_item={'items': '0.00002527'}, temp_paths=0.038709) I0418 00:43:55.019779 140660516149056 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9. I0418 00:43:55.020088 140532160374528 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:03:55.020057 I0418 00:43:55.022019 140660516149056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 00:43:55.022266 140531128596224 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 00:43:55.022397 140660516149056 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776473000.601273, 'wait_for_prev_duration_secs': 33.66818952560425, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776473034.2714834, 'checkpointer_blocking_duration_secs': 0.7487330436706543, 'get_old_steps_start_time': 1776473035.0202405, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776473000.5994592, 'checkpoint_manager_blocking_duration_secs': 34.42290425300598} I0418 00:43:55.022561 140660516149056 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0418 00:43:55.022608 140660516149056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 00:43:55.223633 140532713256704 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9 I0418 00:43:57.397452 140532697245440 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9/items I0418 00:44:01.127562 140531628398336 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9/items/array_metadatas/process_0 I0418 00:44:40.764251 140531611612928 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.253931s (commit=40.644851s, array_metadata_write=1.609079s) I0418 00:44:40.765315 140532160374528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.504 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.780757188797 s) (per-host) I0418 00:44:40.765389 140532160374528 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.745193s. I0418 00:44:43.097335 140532160374528 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000261s. I0418 00:44:43.425104 140532160374528 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 00:44:44.134181 140532160374528 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.642248s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9/items I0418 00:44:44.135807 140532160374528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9/items I0418 00:44:45.096637 140532160374528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9 I0418 00:44:47.365588 140532160374528 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9`. I0418 00:44:47.366459 140532160374528 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.09 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_09_pdb_lt_1/checkpoints/9 I0418 00:44:47.367908 140532160374528 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.347711s. I0418 00:44:47.368116 140531128596224 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 00:44:47.368228 140531128596224 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 00:44:47.368289 140531128596224 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 00:44:47.370249 140531128596224 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 00:44:47.370450 140660516149056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0418 00:44:47.370607 140660516149056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 00:44:47.371573 140660516149056 metric_logger.py:196] completed step: 9, seconds: 0.058, TFLOP/s/device: 58.191, Tokens/s/device: 8771.178, total_weights: 16384, loss: 5.892, lm_loss: 5.892, perplexity: 361.955 Per train step: Total TFLOPs: 3.40 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 00:44:57 UTC 2026 EXIT_CODE=0