XPK Start: Tue Apr 21 16:14:00 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config.
2026-04-21 16:14:25.405247: 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)
I0421 16:14:25.621375 137030557677376 max_utils.py:273] Attempting to initialize the jax distributed system...
I0421 16:14:34.662178 137030557677376 distributed.py:149] Starting JAX distributed service on [::]:8482
I0421 16:14:34.664583 137030557677376 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-rs-sq4b8-slice-job-0-0.mt-16-zarr3-false-rs-sq4b8:8482
I0421 16:14:35.913370 137030557677376 max_utils.py:284] Jax distributed system initialized!
I0421 16:14:42.052953 137030557677376 max_utils.py:800] System Information: Jax Version: 0.9.2
I0421 16:14:42.053058 137030557677376 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0421 16:14:42.053100 137030557677376 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
I0421 16:14:42.053135 137030557677376 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 16:14:42.751035 137030557677376 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 16:14:42.751325 137030557677376 checkpointing.py:677] Setting up checkpoint logger...
I0421 16:14:42.751379 137030557677376 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0421 16:14:42.751426 137030557677376 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0421 16:14:42.751779 137030557677376 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7ca04bb185c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 16:14:43.984400 137030557677376 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0421 16:14:43.984659 137030557677376 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c8c1836de80>}, handler_registry=None
I0421 16:14:43.984922 137030557677376 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c8c1836de80>` 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`.
I0421 16:14:43.984999 137030557677376 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8c1836b470>` 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`.
I0421 16:14:43.985054 137030557677376 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 0x7c8c1836de80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c8c1836de80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8c1836b470>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8c1836b470>}).
I0421 16:14:43.985407 137030557677376 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0421 16:14:43.985497 137030557677376 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c8b404ad580> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0421 16:14:45.521093 136909816829696 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776787871186732693, 'commit_timestamp_nsecs': 1776787916863480636, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA
I0421 16:14:45.782692 136909263206144 checkpoint_manager.py:1871] Missing metrics for step 9
I0421 16:14:45.930252 137030557677376 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints
I0421 16:14:47.070708 137030557677376 checkpoint_manager.py:929] [process=3][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_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c8c1836cb60>
I0421 16:14:47.070881 137030557677376 checkpointing.py:301] Checkpoint manager created!
I0421 16:14:47.994588 137030557677376 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 16:14:47.994714 137030557677376 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 16:14:48.377434 137030557677376 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 16:14:48.377525 137030557677376 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 16:14:48.394382 137030557677376 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 16:14:48.394443 137030557677376 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 16:14:48.418696 137030557677376 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 16:14:48.418767 137030557677376 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 16:14:48.435750 137030557677376 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 16:14:48.435813 137030557677376 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 16:14:48.452762 137030557677376 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 16:14:48.452823 137030557677376 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 16:14:48.478060 137030557677376 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 16:14:48.478132 137030557677376 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 16:14:48.499528 137030557677376 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 16:14:48.499600 137030557677376 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 16:14:48.711319 137030557677376 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 16:14:48.711427 137030557677376 checkpointing.py:581] restoring from this run's directory step 9
I0421 16:14:48.981399 137030557677376 event_tracking.py:70] [process=3] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/9.
I0421 16:14:49.392024 137030557677376 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/9.
I0421 16:14:49.392456 137030557677376 event_tracking.py:125] [process=3] [sync] Finished blocking load in 0.41 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/9.
I0421 16:14:50.821576 1448 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 16:14:57.399204 137030557677376 jax_array_handlers.py:843] [process=3] /jax/orbax/read/worker/io/requested throughput: 240.030 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.583651781082153 s) (per-host)
I0421 16:14:57.404840 137030557677376 base_pytree_checkpoint_handler.py:130] [process=3] /jax/checkpoint/read/gbytes_per_sec: 226.015 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.002148151397705 s) (per-host)
I0421 16:14:58.502980 137030557677376 event_tracking.py:138] [process=3] [sync] Finished load in 9.52 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/9
I0421 16:14:58.503138 137030557677376 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776788088.981292, 'checkpointer_duration_secs': 9.521810293197632, 'checkpoint_manager_start_time': 1776788088.7116988, 'checkpoint_manager_duration_secs': 9.79140591621399}
fsdp: 32
I0421 16:14:58.920900 137030557677376 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0421 16:14:58.921046 137030557677376 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0421 16:14:58.921102 137030557677376 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0421 16:14:58.921162 137030557677376 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0421 16:14:58.921216 137030557677376 maxtext_utils.py:1668] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0421 16:14:58.921253 137030557677376 maxtext_utils.py:1668] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0421 16:14:58.921304 137030557677376 maxtext_utils.py:1668] 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)
I0421 16:14:58.921354 137030557677376 maxtext_utils.py:1668] 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')
I0421 16:14:58.921393 137030557677376 maxtext_utils.py:1668] 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)
I0421 16:14:58.921429 137030557677376 maxtext_utils.py:1668] 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)
I0421 16:14:58.921475 137030557677376 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0421 16:14:58.921530 137030557677376 maxtext_utils.py:1668] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0421 16:14:59.412524 137030557677376 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 16:14:59.412617 137030557677376 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 16:14:59.428516 137030557677376 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 16:14:59.428578 137030557677376 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 16:15:10.456250 137030557677376 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.
I0421 16:15:10.457172 137030557677376 metric_logger.py:301] number parameters: 1.104 billion
I0421 16:15:21.595813 137030557677376 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0421 16:15:22.245610 137030557677376 checkpointing.py:776] Waited 0.6497752666473389 seconds for step 10 to finish before starting checkpointing.
I0421 16:15:22.248020 137030557677376 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 16:15:22.249814 137030557677376 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 10
I0421 16:15:22.251261 137030557677376 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/10.
I0421 16:15:23.017577 137030557677376 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 16:15:23.018536 137030557677376 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 16:15:23.018593 137030557677376 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
I0421 16:15:23.300380 137030557677376 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.282885s
I0421 16:15:23.300560 137030557677376 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.377 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28690242767333984 s) (per-host)
I0421 16:15:23.300617 137030557677376 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286971s (batch_requests_ready=0.002301s, total_serialization_initiated=0.284593s, others=0.000077s)
I0421 16:15:23.300711 137030557677376 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.291154s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.291138)
I0421 16:15:23.301398 137030557677376 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.05 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/10.
I0421 16:15:23.301735 136909229635328 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 16:35:23.301697
I0421 16:15:23.316188 137030557677376 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 16:15:23.316481 136908147500800 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0421 16:15:23.316637 137030557677376 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776788122.2480016, 'wait_for_prev_duration_secs': 5.888938903808594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776788122.2498522, 'checkpointer_blocking_duration_secs': 1.052032470703125, 'get_old_steps_start_time': 1776788123.3019087, 'get_old_steps_duration_secs': 4.291534423828125e-05, 'checkpoint_manager_blocking_start_time': 1776788122.2461145, 'checkpoint_manager_blocking_duration_secs': 1.0704853534698486}
I0421 16:15:23.316738 137030557677376 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0421 16:15:23.316787 137030557677376 max_utils.py:750]
Memstats: After params initialized:
I0421 16:15:23.316835 137030557677376 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0))
I0421 16:15:23.316874 137030557677376 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0))
I0421 16:15:23.316912 137030557677376 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0))
I0421 16:15:23.316951 137030557677376 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0))
I0421 16:15:23.633965 137030557677376 metric_logger.py:196] completed step: 10, seconds: 11.139, TFLOP/s/device: 1.220, Tokens/s/device: 183.866, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0421 16:15:24.209633 137030557677376 metric_logger.py:196] completed step: 11, seconds: 2.037, TFLOP/s/device: 6.672, Tokens/s/device: 1005.630, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0421 16:15:24.357026 137030557677376 metric_logger.py:196] completed step: 12, seconds: 0.430, TFLOP/s/device: 31.599, Tokens/s/device: 4762.990, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0421 16:15:24.504534 137030557677376 metric_logger.py:196] completed step: 13, seconds: 0.152, TFLOP/s/device: 89.420, Tokens/s/device: 13478.295, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0421 16:15:24.509241 137030557677376 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0421 16:15:24.798648 137030557677376 checkpointing.py:776] Waited 0.2893688678741455 seconds for step 15 to finish before starting checkpointing.
I0421 16:15:24.801761 137030557677376 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 16:15:28.495037 136908155893504 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/10/items/array_metadatas/process_3
I0421 16:16:05.903668 136909229635328 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.829 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.889976501464844 s) (per-host)
I0421 16:16:05.903799 136909229635328 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.601958s.
I0421 16:16:13.726771 136909229635328 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 50.424913s.
I0421 16:16:13.727092 136908147500800 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 16:16:13.727217 136908147500800 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0421 16:16:13.727263 136908147500800 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0421 16:16:13.728604 136908147500800 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0421 16:16:13.728785 137030557677376 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0421 16:16:13.728920 137030557677376 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.927159 seconds. If this number is high, consider checkpointing less frequently.
I0421 16:16:13.730888 137030557677376 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 15
I0421 16:16:13.733422 137030557677376 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/15.
I0421 16:16:14.025459 137030557677376 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 16:16:14.025553 137030557677376 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
I0421 16:16:14.050972 137030557677376 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026686s
I0421 16:16:14.051100 137030557677376 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 51.125 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030172348022460938 s) (per-host)
I0421 16:16:14.051146 137030557677376 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030228s (batch_requests_ready=0.001874s, total_serialization_initiated=0.028292s, others=0.000062s)
I0421 16:16:14.051222 137030557677376 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035232s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.035220)
I0421 16:16:14.051923 137030557677376 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.32 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/15.
I0421 16:16:14.052192 136908147500800 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 16:36:14.052170
I0421 16:16:14.064627 137030557677376 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 16:16:14.064952 136908155893504 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0421 16:16:14.065115 137030557677376 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776788124.801731, 'wait_for_prev_duration_secs': 48.92715930938721, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776788173.7309406, 'checkpointer_blocking_duration_secs': 0.3213491439819336, 'get_old_steps_start_time': 1776788174.0523157, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776788124.799225, 'checkpoint_manager_blocking_duration_secs': 49.26585531234741}
I0421 16:16:14.065222 137030557677376 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0421 16:16:14.066001 137030557677376 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.142, Tokens/s/device: 13888.606, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0421 16:16:14.072171 137030557677376 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.181, Tokens/s/device: 13894.543, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0421 16:16:19.250324 136908139108096 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/15/items/array_metadatas/process_3
I0421 16:16:34.861129 137030557677376 metric_logger.py:196] completed step: 16, seconds: 49.561, TFLOP/s/device: 0.274, Tokens/s/device: 41.322, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0421 16:16:35.008356 137030557677376 metric_logger.py:196] completed step: 17, seconds: 20.643, TFLOP/s/device: 0.658, Tokens/s/device: 99.211, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0421 16:16:35.155796 137030557677376 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.283, Tokens/s/device: 13457.659, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0421 16:16:35.302234 137030557677376 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0421 16:16:35.303012 137030557677376 checkpointing.py:776] Waited 0.0007963180541992188 seconds for step 19 to finish before starting checkpointing.
I0421 16:16:35.305041 137030557677376 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 16:16:56.583688 136908147500800 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.112 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.562710762023926 s) (per-host)
I0421 16:16:56.583807 136908147500800 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.531540s.
I0421 16:17:03.203457 136908147500800 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 49.151174s.
I0421 16:17:03.203754 136908155893504 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 16:17:03.203871 136908155893504 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0421 16:17:03.203918 136908155893504 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0421 16:17:03.205353 136908155893504 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0421 16:17:03.205626 137030557677376 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0421 16:17:03.205763 137030557677376 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.900725 seconds. If this number is high, consider checkpointing less frequently.
I0421 16:17:03.207366 137030557677376 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 19
I0421 16:17:03.209508 137030557677376 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/19.
I0421 16:17:04.351394 137030557677376 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 16:17:04.351489 137030557677376 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
I0421 16:17:04.377040 137030557677376 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026506s
I0421 16:17:04.377211 137030557677376 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 51.251 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03009819984436035 s) (per-host)
I0421 16:17:04.377279 137030557677376 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030159s (batch_requests_ready=0.001756s, total_serialization_initiated=0.028337s, others=0.000065s)
I0421 16:17:04.377380 137030557677376 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034324s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.034313)
I0421 16:17:04.378054 137030557677376 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/19.
I0421 16:17:04.378369 136903869318912 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 16:37:04.378332
I0421 16:17:04.381441 137030557677376 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 16:17:04.381686 136908155893504 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0421 16:17:04.381838 137030557677376 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776788195.3050122, 'wait_for_prev_duration_secs': 27.90072536468506, 'time_between_consecutive_saves_sec': 21.5763680934906, 'checkpointer_blocking_start_time': 1776788223.207405, 'checkpointer_blocking_duration_secs': 1.1711146831512451, 'get_old_steps_start_time': 1776788224.3785346, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1776788195.3032532, 'checkpoint_manager_blocking_duration_secs': 29.078554391860962}
I0421 16:17:04.381959 137030557677376 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0421 16:17:04.382004 137030557677376 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 16:17:09.975946 136908139108096 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_16_zarr3_false/checkpoints/19/items/array_metadatas/process_3
I0421 16:17:47.118851 136903869318912 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.931 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.7716965675354 s) (per-host)
I0421 16:17:47.118982 136903869318912 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.740498s.
I0421 16:17:52.990693 136903869318912 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.612187s.
I0421 16:17:52.990980 136908155893504 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 16:17:52.991094 136908155893504 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0421 16:17:52.991143 136908155893504 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0421 16:17:52.992716 136908155893504 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0421 16:17:52.992904 137030557677376 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0421 16:17:52.993069 137030557677376 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 16:17:52.994164 137030557677376 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.345, Tokens/s/device: 13919.284, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Tue Apr 21 16:18:05 UTC 2026
EXIT_CODE=0