XPK Start: Sat Apr 18 13:32:34 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 13:32:58.772175: 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 13:32:58.952540 134391244810048 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 13:33:07.993699 134391244810048 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 13:33:07.995987 134391244810048 distributed.py:172] Connecting to JAX distributed service on mt-09-pdb-lt-1-cjcqy-slice-job-0-0.mt-09-pdb-lt-1-cjcqy:8482
I0418 13:33:09.226254 134391244810048 max_utils.py:284] Jax distributed system initialized!
I0418 13:33:15.411947 134391244810048 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 13:33:15.412046 134391244810048 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 13:33:15.412087 134391244810048 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 13:33:15.412130 134391244810048 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 13:33:16.100085 134391244810048 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 8, 1, 1, 1, 1, 4, 1, 1, 1, 1)
I0418 13:33:16.100361 134391244810048 checkpointing.py:677] Setting up checkpoint logger...
I0418 13:33:16.100413 134391244810048 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 13:33:16.100457 134391244810048 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 13:33:16.100789 134391244810048 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 0x7a39e9bde6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 13:33:19.501873 134391244810048 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 13:33:19.502106 134391244810048 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 0x7a25843bce30>}, handler_registry=None
I0418 13:33:19.502343 134391244810048 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a25843bce30>` 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 13:33:19.502392 134391244810048 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2584230cb0>` 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 13:33:19.502427 134391244810048 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 0x7a25843bce30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a25843bce30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2584230cb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2584230cb0>}).
I0418 13:33:19.502748 134391244810048 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 13:33:19.502819 134391244810048 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a2484584220> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 13:33:21.448406 134391244810048 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints
I0418 13:33:22.312879 134391244810048 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_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a258422e870>
I0418 13:33:22.313048 134391244810048 checkpointing.py:301] Checkpoint manager created!
I0418 13:33:23.218027 134391244810048 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 13:33:23.218142 134391244810048 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor').
I0418 13:33:23.597335 134391244810048 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 13:33:23.597425 134391244810048 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor').
I0418 13:33:23.613944 134391244810048 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 13:33:23.614003 134391244810048 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor').
I0418 13:33:23.637558 134391244810048 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 13:33:23.637622 134391244810048 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None).
I0418 13:33:23.654004 134391244810048 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 13:33:23.654061 134391244810048 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None).
I0418 13:33:23.670543 134391244810048 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 13:33:23.670604 134391244810048 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None).
I0418 13:33:23.695326 134391244810048 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 13:33:23.695402 134391244810048 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None).
I0418 13:33:23.716395 134391244810048 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 13:33:23.716455 134391244810048 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, 'tensor').
I0418 13:33:23.936257 134391244810048 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 13:33:23.936365 134391244810048 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 8
tensor: 4
I0418 13:33:25.376742 134391244810048 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: ('tensor',)
I0418 13:33:25.376886 134391244810048 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 13:33:25.376954 134391244810048 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 13:33:25.377032 134391244810048 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 13:33:25.377097 134391244810048 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 13:33:25.377136 134391244810048 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 13:33:25.377185 134391244810048 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 13:33:25.377235 134391244810048 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 13:33:25.377275 134391244810048 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 13:33:25.377312 134391244810048 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 13:33:25.377359 134391244810048 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', 'tensor')
I0418 13:33:25.377405 134391244810048 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: ('tensor', 'fsdp')
I0418 13:33:25.875066 134391244810048 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0418 13:33:25.875166 134391244810048 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0418 13:33:25.890645 134391244810048 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0418 13:33:25.890701 134391244810048 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
I0418 13:33:40.919357 134391244810048 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 13:33:40.927388 134391244810048 metric_logger.py:301] number parameters: 1.104 billion
I0418 13:33:42.412984 134391244810048 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 13:33:56.498024 134391244810048 checkpointing.py:776] Waited 14.08501672744751 seconds for step 0 to finish before starting checkpointing.
I0418 13:33:56.500422 134391244810048 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 13:33:56.502481 134391244810048 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 13:33:56.504000 134391244810048 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0.
I0418 13:33:57.201547 134391244810048 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 13:33:57.215515 134391244810048 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 13:33:57.215581 134391244810048 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 13:33:57.458499 134262019847936 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0
I0418 13:33:57.476804 134391244810048 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.262006s
I0418 13:33:57.477637 134391244810048 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.744 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.268543004989624 s) (per-host)
I0418 13:33:57.477710 134391244810048 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.268635s (batch_requests_ready=0.002315s, total_serialization_initiated=0.265572s, others=0.000747s)
I0418 13:33:57.477789 134391244810048 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.272640s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.272624)
I0418 13:33:57.478502 134391244810048 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.98 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0.
I0418 13:33:57.478849 134262506362624 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:53:57.478796
I0418 13:33:57.487791 134391244810048 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 13:33:57.488127 134261482977024 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 13:33:57.488288 134391244810048 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776519236.5004036, 'wait_for_prev_duration_secs': 5.936622619628906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776519236.5025191, 'checkpointer_blocking_duration_secs': 0.9764254093170166, 'get_old_steps_start_time': 1776519237.4789686, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776519236.498541, 'checkpoint_manager_blocking_duration_secs': 0.9897055625915527}
I0418 13:33:57.488401 134391244810048 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 13:33:57.488452 134391244810048 max_utils.py:750]
Memstats: After params initialized:
I0418 13:33:57.488500 134391244810048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 13:33:57.488533 134391244810048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 13:33:57.488557 134391244810048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 13:33:57.488582 134391244810048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 13:33:57.823848 134391244810048 metric_logger.py:196] completed step: 0, seconds: 1.147, TFLOP/s/device: 2.961, Tokens/s/device: 446.351, total_weights: 16384, loss: 10.862, lm_loss: 10.862, perplexity: 52153.723
I0418 13:33:57.827351 134391244810048 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/tensorboard/'
I0418 13:33:58.221623 134391244810048 metric_logger.py:196] completed step: 1, seconds: 15.409, TFLOP/s/device: 0.220, Tokens/s/device: 33.227, total_weights: 16384, loss: 10.862, lm_loss: 10.862, perplexity: 52153.723
I0418 13:33:58.278603 134391244810048 metric_logger.py:196] completed step: 2, seconds: 0.398, TFLOP/s/device: 8.524, Tokens/s/device: 1284.857, total_weights: 16384, loss: 9.763, lm_loss: 9.763, perplexity: 17373.129
I0418 13:33:58.336894 134391244810048 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 391.831, Tokens/s/device: 59061.022, total_weights: 16384, loss: 8.823, lm_loss: 8.823, perplexity: 6791.817
I0418 13:34:00.918087 134262003062528 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/items
I0418 13:34:01.118685 134261468473088 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776519240671026374, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776519240671026374, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/_CHECKPOINT_METADATA
I0418 13:34:03.313865 2588 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 13:34:05.495027 134261986277120 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/items/array_metadatas/process_0
I0418 13:34:13.250793 134391244810048 metric_logger.py:196] completed step: 4, seconds: 0.058, TFLOP/s/device: 59.049, Tokens/s/device: 8900.478, total_weights: 16384, loss: 7.947, lm_loss: 7.947, perplexity: 2828.417
I0418 13:34:13.259516 134391244810048 metric_logger.py:196] completed step: 5, seconds: 0.058, TFLOP/s/device: 58.123, Tokens/s/device: 8760.973, total_weights: 16384, loss: 7.212, lm_loss: 7.212, perplexity: 1355.339
I0418 13:34:13.316550 134391244810048 metric_logger.py:196] completed step: 6, seconds: 14.913, TFLOP/s/device: 0.228, Tokens/s/device: 34.332, total_weights: 16384, loss: 6.652, lm_loss: 6.652, perplexity: 774.659
I0418 13:34:13.374951 134391244810048 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 527.696, Tokens/s/device: 79540.158, total_weights: 16384, loss: 6.269, lm_loss: 6.269, perplexity: 527.696
I0418 13:34:13.433240 134391244810048 metric_logger.py:196] completed step: 8, seconds: 0.058, TFLOP/s/device: 58.689, Tokens/s/device: 8846.194, total_weights: 16384, loss: 6.031, lm_loss: 6.031, perplexity: 416.165
I0418 13:34:13.490814 134391244810048 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 13:34:13.491434 134391244810048 checkpointing.py:776] Waited 0.0006301403045654297 seconds for step 9 to finish before starting checkpointing.
I0418 13:34:13.493519 134391244810048 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 13:34:39.435375 134261969491712 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.121348s (commit=35.019305s, array_metadata_write=1.102044s)
I0418 13:34:39.436407 134262506362624 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.407 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.227309226989746 s) (per-host)
I0418 13:34:39.436476 134262506362624 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.957016s.
I0418 13:34:40.979851 134262506362624 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776519240671026374, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/_CHECKPOINT_METADATA
I0418 13:34:42.783948 134262506362624 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 13:34:42.943777 134262506362624 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000248s.
I0418 13:34:43.014626 134261468473088 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776519240671026374, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/_CHECKPOINT_METADATA
I0418 13:34:43.718145 134262506362624 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.183639s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/items
I0418 13:34:43.719923 134262506362624 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0/items
I0418 13:34:44.679686 134262506362624 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0
I0418 13:34:47.424961 134262506362624 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0`.
I0418 13:34:47.425879 134262506362624 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.92 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/0
I0418 13:34:47.427200 134262506362624 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.947738s.
I0418 13:34:47.427433 134261482977024 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 13:34:47.427556 134261482977024 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 13:34:47.427618 134261482977024 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 13:34:47.429450 134261482977024 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 13:34:47.429581 134391244810048 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 13:34:47.429666 134391244810048 checkpoint_manager.py:1452] Waiting for previous save to complete took 33.936149 seconds. If this number is high, consider checkpointing less frequently.
I0418 13:34:47.431559 134391244810048 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 13:34:47.433446 134391244810048 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9.
I0418 13:34:48.139605 134391244810048 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 13:34:48.139697 134391244810048 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 13:34:48.166339 134391244810048 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027471s
I0418 13:34:48.167074 134391244810048 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.231 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031983137130737305 s) (per-host)
I0418 13:34:48.167148 134391244810048 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032073s (batch_requests_ready=0.001919s, total_serialization_initiated=0.029464s, others=0.000690s)
I0418 13:34:48.167220 134391244810048 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036583s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.036566)
I0418 13:34:48.167890 134391244810048 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9.
I0418 13:34:48.168206 134261482977024 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 13:54:48.168171
I0418 13:34:48.170502 134391244810048 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 13:34:48.170765 134260400842496 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 13:34:48.170920 134391244810048 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776519253.493488, 'wait_for_prev_duration_secs': 33.93614935874939, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776519287.4315987, 'checkpointer_blocking_duration_secs': 0.7367384433746338, 'get_old_steps_start_time': 1776519288.1683614, 'get_old_steps_duration_secs': 3.409385681152344e-05, 'checkpoint_manager_blocking_start_time': 1776519253.4916544, 'checkpoint_manager_blocking_duration_secs': 34.67923069000244}
I0418 13:34:48.171096 134391244810048 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 13:34:48.171146 134391244810048 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 13:34:48.391663 134262506362624 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9
I0418 13:34:50.970212 134261986277120 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9/items
I0418 13:34:55.178596 134260946106112 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9/items/array_metadatas/process_0
I0418 13:35:36.143209 134260409235200 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.592706s (commit=41.192633s, array_metadata_write=2.400073s)
I0418 13:35:36.144261 134261482977024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 32.902 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 48.00916361808777 s) (per-host)
I0418 13:35:36.144326 134261482977024 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.976015s.
I0418 13:35:38.648734 134261482977024 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 13:35:38.825868 134261482977024 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000241s.
I0418 13:35:39.593059 134261482977024 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.633520s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9/items
I0418 13:35:39.594805 134261482977024 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9/items
I0418 13:35:40.599289 134261482977024 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9
I0418 13:35:42.448448 134261482977024 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9`.
I0418 13:35:42.449281 134261482977024 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 55.02 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_09_pdb_lt_1/checkpoints/9
I0418 13:35:42.450940 134261482977024 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 54.282625s.
I0418 13:35:42.451158 134260400842496 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 13:35:42.451281 134260400842496 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 13:35:42.451343 134260400842496 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 13:35:42.455414 134260400842496 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 13:35:42.455538 134391244810048 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 13:35:42.455633 134391244810048 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 13:35:42.456670 134391244810048 metric_logger.py:196] completed step: 9, seconds: 0.058, TFLOP/s/device: 58.149, Tokens/s/device: 8764.872, 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 13:35:54 UTC 2026
EXIT_CODE=0