XPK Start: Fri Apr 17 13:56:18 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-17 13:56:42.544948: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0417 13:56:42.723360 139397101459264 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 13:56:51.764577 139397101459264 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 13:56:51.767189 139397101459264 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-kfmdb-slice-job-0-0.mt-10-shardy-false-kfmdb:8482
I0417 13:56:53.654783 139397101459264 max_utils.py:284] Jax distributed system initialized!
I0417 13:56:58.760065 139397101459264 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 13:56:58.760166 139397101459264 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 13:56:58.760205 139397101459264 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0417 13:56:58.760240 139397101459264 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 13:56:59.807971 139397101459264 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 13:56:59.808249 139397101459264 checkpointing.py:677] Setting up checkpoint logger...
I0417 13:56:59.808303 139397101459264 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0417 13:56:59.808345 139397101459264 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 13:56:59.808689 139397101459264 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 0x7ec733140050>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 13:57:03.166905 139397101459264 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 13:57:03.167142 139397101459264 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eb2cc7054f0>}, handler_registry=None
I0417 13:57:03.167394 139397101459264 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eb2cc7054f0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0417 13:57:03.167453 139397101459264 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ec62740d250>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0417 13:57:03.167499 139397101459264 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 0x7eb2cc7054f0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eb2cc7054f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ec62740d250>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ec62740d250>}).
I0417 13:57:03.168056 139397101459264 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 13:57:03.168138 139397101459264 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7eb2ac760180> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 13:57:04.653903 139397101459264 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints
I0417 13:57:05.491134 139397101459264 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ec62740cce0>
I0417 13:57:05.491309 139397101459264 checkpointing.py:301] Checkpoint manager created!
I0417 13:57:06.414986 139397101459264 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 13:57:06.415108 139397101459264 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 13:57:06.797665 139397101459264 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 13:57:06.797755 139397101459264 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 13:57:06.814253 139397101459264 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 13:57:06.814321 139397101459264 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 13:57:06.838100 139397101459264 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 13:57:06.838169 139397101459264 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:57:06.854670 139397101459264 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 13:57:06.854729 139397101459264 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:57:06.871277 139397101459264 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 13:57:06.871340 139397101459264 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:57:06.897113 139397101459264 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 13:57:06.897183 139397101459264 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:57:06.918137 139397101459264 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 13:57:06.918202 139397101459264 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 13:57:07.135201 139397101459264 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 13:57:07.135311 139397101459264 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0417 13:57:08.511402 139397101459264 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0417 13:57:08.511530 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 13:57:08.511583 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 13:57:08.511646 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0417 13:57:08.511698 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 13:57:08.511736 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 13:57:08.511792 139397101459264 maxtext_utils.py:1651] 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, None, None)
I0417 13:57:08.511859 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/self_attention/out/kernel
Shape: float32[16,16,128,2048]
Logical: P('heads', 'layers', 'kv', 'embed')
Physical: (None, None, None, 'fsdp')
I0417 13:57:08.511903 139397101459264 maxtext_utils.py:1651] params/params/decoder/layers/self_attention/query/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'q_heads', 'kv')
Physical: ('fsdp', None, None, None)
I0417 13:57:08.511941 139397101459264 maxtext_utils.py:1651] 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, None, None)
I0417 13:57:08.511988 139397101459264 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0417 13:57:08.512037 139397101459264 maxtext_utils.py:1651] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0417 13:57:08.997801 139397101459264 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 13:57:08.997909 139397101459264 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 13:57:09.013378 139397101459264 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 13:57:09.013440 139397101459264 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 13:57:19.621839 139397101459264 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0417 13:57:19.622653 139397101459264 metric_logger.py:301] number parameters: 1.104 billion
I0417 13:57:30.834192 139397101459264 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0417 13:57:31.001487 139397101459264 checkpointing.py:776] Waited 0.16727423667907715 seconds for step 0 to finish before starting checkpointing.
I0417 13:57:31.003985 139397101459264 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 13:57:31.005733 139397101459264 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0417 13:57:31.007228 139397101459264 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints/0.
I0417 13:57:31.759785 139397101459264 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 13:57:31.760859 139397101459264 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 13:57:31.760915 139397101459264 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0417 13:57:32.039201 139397101459264 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.279490s
I0417 13:57:32.039375 139397101459264 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.413 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2849760055541992 s) (per-host)
I0417 13:57:32.039427 139397101459264 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285040s (batch_requests_ready=0.002320s, total_serialization_initiated=0.282648s, others=0.000072s)
I0417 13:57:32.039517 139397101459264 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289182s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.289164)
I0417 13:57:32.040347 139397101459264 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints/0.
I0417 13:57:32.040681 139271696545536 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:17:32.040644
I0417 13:57:32.044325 139397101459264 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 13:57:32.044644 139268957189888 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0417 13:57:32.044801 139397101459264 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776434251.0039659, 'wait_for_prev_duration_secs': 6.365776062011719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776434251.0057704, 'checkpointer_blocking_duration_secs': 1.035085916519165, 'get_old_steps_start_time': 1776434252.0408826, 'get_old_steps_duration_secs': 3.1948089599609375e-05, 'checkpoint_manager_blocking_start_time': 1776434251.0021126, 'checkpoint_manager_blocking_duration_secs': 1.042649745941162}
I0417 13:57:32.044919 139397101459264 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0417 13:57:32.044971 139397101459264 max_utils.py:750]
Memstats: After params initialized:
I0417 13:57:32.045021 139397101459264 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0417 13:57:32.045052 139397101459264 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0417 13:57:32.045077 139397101459264 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0417 13:57:32.045101 139397101459264 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0417 13:57:32.356862 139397101459264 metric_logger.py:196] completed step: 0, seconds: 11.211, TFLOP/s/device: 1.212, Tokens/s/device: 182.671, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0417 13:57:32.529223 139397101459264 metric_logger.py:196] completed step: 1, seconds: 1.521, TFLOP/s/device: 8.932, Tokens/s/device: 1346.375, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0417 13:57:32.934355 139397101459264 metric_logger.py:196] completed step: 2, seconds: 0.025, TFLOP/s/device: 543.398, Tokens/s/device: 81906.895, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129
I0417 13:57:33.081740 139397101459264 metric_logger.py:196] completed step: 3, seconds: 0.405, TFLOP/s/device: 33.510, Tokens/s/device: 5051.041, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203
I0417 13:57:33.376866 139397101459264 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.807, Tokens/s/device: 13385.883, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614
I0417 13:57:33.382901 139397101459264 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.242, Tokens/s/device: 13903.692, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717
I0417 13:57:36.582247 2751 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 13:57:40.460009 139268997601024 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints/0/items/array_metadatas/process_6
I0417 13:57:56.670098 139397101459264 metric_logger.py:196] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.938, Tokens/s/device: 6924.346, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759
I0417 13:57:56.817705 139397101459264 metric_logger.py:196] completed step: 7, seconds: 23.140, TFLOP/s/device: 0.587, Tokens/s/device: 88.503, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905
I0417 13:57:56.965280 139397101459264 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 89.068, Tokens/s/device: 13425.283, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406
I0417 13:57:57.112020 139397101459264 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0417 13:57:57.112698 139397101459264 checkpointing.py:776] Waited 0.0007002353668212891 seconds for step 9 to finish before starting checkpointing.
I0417 13:57:57.114728 139397101459264 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 13:58:09.880701 139271696545536 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 41.431 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.126264572143555 s) (per-host)
I0417 13:58:09.880855 139271696545536 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.840062s.
I0417 13:58:18.624451 139271696545536 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 46.583641s.
I0417 13:58:18.624741 139268957189888 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 13:58:18.624880 139268957189888 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0417 13:58:18.624930 139268957189888 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0417 13:58:18.629672 139268957189888 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0417 13:58:18.629869 139397101459264 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0417 13:58:18.630006 139397101459264 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.515280 seconds. If this number is high, consider checkpointing less frequently.
I0417 13:58:18.632082 139397101459264 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0417 13:58:18.634268 139397101459264 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints/9.
I0417 13:58:19.346880 139397101459264 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 13:58:19.346976 139397101459264 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0417 13:58:19.382773 139397101459264 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037084s
I0417 13:58:19.382959 139397101459264 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 37.943 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04065513610839844 s) (per-host)
I0417 13:58:19.383010 139397101459264 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040715s (batch_requests_ready=0.001909s, total_serialization_initiated=0.038737s, others=0.000069s)
I0417 13:58:19.383100 139397101459264 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.044878s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.044861)
I0417 13:58:19.383726 139397101459264 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints/9.
I0417 13:58:19.384066 139268997601024 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:18:19.384029
I0417 13:58:19.386025 139397101459264 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 13:58:19.386310 139268957189888 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0417 13:58:19.386442 139397101459264 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776434277.114696, 'wait_for_prev_duration_secs': 21.51528000831604, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776434298.6321208, 'checkpointer_blocking_duration_secs': 0.7520937919616699, 'get_old_steps_start_time': 1776434299.3842394, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776434277.1129718, 'checkpoint_manager_blocking_duration_secs': 22.27343773841858}
I0417 13:58:19.386593 139397101459264 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0417 13:58:19.386644 139397101459264 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 13:58:25.212197 139248736921344 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_10_shardy_false/checkpoints/9/items/array_metadatas/process_6
I0417 13:59:01.180330 139268997601024 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.755 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.83798837661743 s) (per-host)
I0417 13:59:01.180456 139268997601024 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.796275s.
I0417 13:59:09.554722 139268997601024 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.170528s.
I0417 13:59:09.555044 139268957189888 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 13:59:09.555172 139268957189888 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0417 13:59:09.555221 139268957189888 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0417 13:59:09.557508 139268957189888 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0417 13:59:09.557714 139397101459264 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0417 13:59:09.557881 139397101459264 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 13:59:09.558958 139397101459264 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.237, Tokens/s/device: 13902.937, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 17 13:59:19 UTC 2026
EXIT_CODE=0