XPK Start: Sun Apr 19 19:24:38 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-19 19:25:02.422142: 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)
I0419 19:25:02.952359 137589409388352 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 19:25:11.993729 137589409388352 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 19:25:11.995975 137589409388352 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-ht02x-slice-job-0-0.mt-15-ocdbt-false-sv-ht02x:8482
I0419 19:25:14.368622 137589409388352 max_utils.py:284] Jax distributed system initialized!
I0419 19:25:20.664736 137589409388352 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 19:25:20.664838 137589409388352 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 19:25:20.664878 137589409388352 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
I0419 19:25:20.664913 137589409388352 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 19:25:21.360181 137589409388352 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 19:25:21.360500 137589409388352 checkpointing.py:677] Setting up checkpoint logger...
I0419 19:25:21.360571 137589409388352 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True
I0419 19:25:21.360633 137589409388352 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 19:25:21.361007 137589409388352 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7d225033c410>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 19:25:24.694178 137589409388352 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 19:25:24.694427 137589409388352 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 0x7d0e5446b560>}, handler_registry=None
I0419 19:25:24.694661 137589409388352 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d0e5446b560>` 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`.
I0419 19:25:24.694709 137589409388352 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d0e54083230>` 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`.
I0419 19:25:24.694746 137589409388352 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 0x7d0e5446b560>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d0e5446b560>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d0e54083230>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d0e54083230>}).
I0419 19:25:24.695068 137589409388352 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 19:25:24.695146 137589409388352 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d0e1c5cf880> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 19:25:25.865797 137589409388352 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints
I0419 19:25:26.304532 137589409388352 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=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d0e540813a0>
I0419 19:25:26.304699 137589409388352 checkpointing.py:301] Checkpoint manager created!
I0419 19:25:27.226991 137589409388352 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 19:25:27.227107 137589409388352 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 19:25:27.610910 137589409388352 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 19:25:27.611003 137589409388352 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 19:25:27.627060 137589409388352 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 19:25:27.627116 137589409388352 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 19:25:27.651251 137589409388352 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 19:25:27.651318 137589409388352 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:25:27.667478 137589409388352 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 19:25:27.667538 137589409388352 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:25:27.683681 137589409388352 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 19:25:27.683744 137589409388352 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:25:27.708168 137589409388352 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 19:25:27.708234 137589409388352 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:25:27.729138 137589409388352 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 19:25:27.729200 137589409388352 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 19:25:27.945749 137589409388352 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 19:25:27.945857 137589409388352 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 19:25:29.374253 137589409388352 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0419 19:25:29.374376 137589409388352 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, None)
I0419 19:25:29.374447 137589409388352 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, None)
I0419 19:25:29.374508 137589409388352 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0419 19:25:29.374560 137589409388352 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0419 19:25:29.374598 137589409388352 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0419 19:25:29.374649 137589409388352 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, None, None)
I0419 19:25:29.374700 137589409388352 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: (None, None, None, 'fsdp')
I0419 19:25:29.374739 137589409388352 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, None, None)
I0419 19:25:29.374775 137589409388352 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, None, None)
I0419 19:25:29.374822 137589409388352 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0419 19:25:29.374870 137589409388352 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0419 19:25:29.860491 137589409388352 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 19:25:29.860586 137589409388352 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 19:25:29.875758 137589409388352 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 19:25:29.875821 137589409388352 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 19:25:40.900435 137589409388352 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.
I0419 19:25:40.901218 137589409388352 metric_logger.py:301] number parameters: 1.104 billion
I0419 19:25:52.139256 137589409388352 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0419 19:25:52.291747 137589409388352 checkpointing.py:776] Waited 0.1524674892425537 seconds for step 0 to finish before starting checkpointing.
I0419 19:25:52.294748 137589409388352 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 19:25:52.296132 137589409388352 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0419 19:25:52.297610 137589409388352 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/0.
I0419 19:25:53.030111 137589409388352 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 19:25:53.031117 137589409388352 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 19:25:53.031179 137589409388352 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
I0419 19:25:53.303503 137589409388352 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273464s
I0419 19:25:53.303682 137589409388352 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.531 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2789154052734375 s) (per-host)
I0419 19:25:53.303745 137589409388352 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278986s (batch_requests_ready=0.002225s, total_serialization_initiated=0.276682s, others=0.000079s)
I0419 19:25:53.303857 137589409388352 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283134s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.283114)
I0419 19:25:53.304739 137589409388352 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/0.
I0419 19:25:53.305078 137461300721408 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:45:53.305037
I0419 19:25:53.319470 137589409388352 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 19:25:53.319725 137459102279424 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 19:25:53.319881 137589409388352 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776626752.294729, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776626752.2961762, 'checkpointer_blocking_duration_secs': 1.0090632438659668, 'get_old_steps_start_time': 1776626753.305266, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776626752.292228, 'checkpoint_manager_blocking_duration_secs': 1.0276153087615967}
I0419 19:25:53.319983 137589409388352 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0419 19:25:53.320031 137589409388352 max_utils.py:750]
Memstats: After params initialized:
I0419 19:25:53.320080 137589409388352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0419 19:25:53.320113 137589409388352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0419 19:25:53.320139 137589409388352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0419 19:25:53.320163 137589409388352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0419 19:25:53.632340 137589409388352 metric_logger.py:196] completed step: 0, seconds: 11.238, TFLOP/s/device: 1.209, Tokens/s/device: 182.240, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0419 19:25:53.809034 137589409388352 metric_logger.py:196] completed step: 1, seconds: 1.491, TFLOP/s/device: 9.110, Tokens/s/device: 1373.128, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0419 19:25:54.667127 137589409388352 metric_logger.py:196] completed step: 2, seconds: 0.029, TFLOP/s/device: 461.488, Tokens/s/device: 69560.492, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0419 19:25:54.814640 137589409388352 metric_logger.py:196] completed step: 3, seconds: 0.859, TFLOP/s/device: 15.823, Tokens/s/device: 2385.034, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0419 19:25:54.819644 137589409388352 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0419 19:25:55.108909 137589409388352 checkpointing.py:776] Waited 0.2892279624938965 seconds for step 5 to finish before starting checkpointing.
I0419 19:25:55.111960 137589409388352 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 19:25:56.735630 2836 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 19:25:59.099456 137459110672128 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_6
I0419 19:26:03.876338 137461300721408 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 145.564 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.851547479629517 s) (per-host)
I0419 19:26:03.876459 137461300721408 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.571263s.
I0419 19:26:10.399049 137461300721408 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 17.093835s.
I0419 19:26:10.399332 137459102279424 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 19:26:10.399480 137459102279424 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 19:26:10.399546 137459102279424 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 19:26:10.401232 137459102279424 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 19:26:10.401439 137589409388352 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.
W0419 19:26:10.401576 137589409388352 checkpoint_manager.py:1452] Waiting for previous save to complete took 15.289617 seconds. If this number is high, consider checkpointing less frequently.
I0419 19:26:10.403440 137589409388352 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5
I0419 19:26:10.405895 137589409388352 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/5.
I0419 19:26:10.702887 137589409388352 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 19:26:10.702981 137589409388352 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
I0419 19:26:10.741543 137589409388352 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039842s
I0419 19:26:10.741701 137589409388352 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 35.697 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0432133674621582 s) (per-host)
I0419 19:26:10.741752 137589409388352 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.043273s (batch_requests_ready=0.001749s, total_serialization_initiated=0.041456s, others=0.000068s)
I0419 19:26:10.741838 137589409388352 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048115s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.048099)
I0419 19:26:10.742508 137589409388352 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/5.
I0419 19:26:10.742841 137459102279424 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:46:10.742803
I0419 19:26:10.744787 137589409388352 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 19:26:10.745026 137460226979584 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 19:26:10.745166 137589409388352 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776626755.1119301, 'wait_for_prev_duration_secs': 15.289616584777832, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776626770.4034772, 'checkpointer_blocking_duration_secs': 0.3395092487335205, 'get_old_steps_start_time': 1776626770.74301, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776626755.109588, 'checkpoint_manager_blocking_duration_secs': 15.63553524017334}
I0419 19:26:10.745275 137589409388352 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0419 19:26:10.746098 137589409388352 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.855, Tokens/s/device: 13393.237, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0419 19:26:10.752164 137589409388352 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 91.957, Tokens/s/device: 13860.783, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0419 19:26:16.309008 137459110672128 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_6
I0419 19:26:21.588270 137459102279424 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 145.053 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.889754056930542 s) (per-host)
I0419 19:26:21.588386 137459102279424 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.845434s.
I0419 19:26:30.300516 137459102279424 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 19.557546s.
I0419 19:26:30.300794 137460226979584 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 19:26:30.300913 137460226979584 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 19:26:30.300962 137460226979584 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0419 19:26:30.302528 137460226979584 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0419 19:26:33.384071 137589409388352 metric_logger.py:196] completed step: 6, seconds: 15.931, TFLOP/s/device: 0.853, Tokens/s/device: 128.554, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0419 19:26:33.531580 137589409388352 metric_logger.py:196] completed step: 7, seconds: 22.486, TFLOP/s/device: 0.604, Tokens/s/device: 91.080, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0419 19:26:33.678664 137589409388352 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.298, Tokens/s/device: 13459.959, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0419 19:26:33.825689 137589409388352 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0419 19:26:33.826324 137589409388352 checkpointing.py:776] Waited 0.0006573200225830078 seconds for step 9 to finish before starting checkpointing.
I0419 19:26:33.828619 137589409388352 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 19:26:33.830183 137589409388352 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0419 19:26:33.831754 137589409388352 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/9.
I0419 19:26:34.548137 137589409388352 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 19:26:34.548237 137589409388352 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
I0419 19:26:34.589772 137589409388352 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042649s
I0419 19:26:34.589931 137589409388352 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.501 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04604625701904297 s) (per-host)
I0419 19:26:34.589980 137589409388352 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046105s (batch_requests_ready=0.001641s, total_serialization_initiated=0.044400s, others=0.000064s)
I0419 19:26:34.590091 137589409388352 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050367s (all_items=0.000010s, per_item={'items': '0.00001049'}, temp_paths=0.050356)
I0419 19:26:34.590867 137589409388352 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/9.
I0419 19:26:34.591163 137460226979584 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:46:34.591131
I0419 19:26:34.593271 137589409388352 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 19:26:34.593562 137438064236288 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 19:26:34.593698 137589409388352 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776626793.8285832, 'wait_for_prev_duration_secs': 8.654594421386719e-05, 'time_between_consecutive_saves_sec': 3.526015520095825, 'checkpointer_blocking_start_time': 1776626793.8302207, 'checkpointer_blocking_duration_secs': 0.7610878944396973, 'get_old_steps_start_time': 1776626794.5913262, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776626793.826572, 'checkpoint_manager_blocking_duration_secs': 0.7670941352844238}
I0419 19:26:34.593884 137589409388352 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0419 19:26:34.593935 137589409388352 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 19:26:41.110556 137438525638400 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_6
I0419 19:26:46.094188 137460226979584 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 136.759 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.550263166427612 s) (per-host)
I0419 19:26:46.094305 137460226979584 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.503031s.
I0419 19:26:52.396960 137460226979584 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 17.805669s.
I0419 19:26:52.397291 137438064236288 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 19:26:52.397430 137438064236288 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 19:26:52.397475 137438064236288 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0419 19:26:52.398758 137438064236288 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0419 19:26:52.398955 137589409388352 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.
I0419 19:26:52.399108 137589409388352 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 19:26:52.400241 137589409388352 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.053, Tokens/s/device: 13875.245, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 19:27:03 UTC 2026
EXIT_CODE=0