MaxView

← Back to run

Log Summary

XPK Start: Fri Apr 17 14:52:47 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 14:53:11.099945: 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 14:53:11.282804 132951197263680 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 14:53:20.324039 132951197263680 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 14:53:20.326377 132951197263680 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-rs-ga5bs-slice-job-0-0.mt-16-zarr3-true-rs-ga5bs:8482
I0417 14:53:21.403767 132951197263680 max_utils.py:284] Jax distributed system initialized!
I0417 14:53:26.632497 132951197263680 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 14:53:26.632601 132951197263680 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 14:53:26.632641 132951197263680 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 14:53:26.632676 132951197263680 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 14:53:27.298556 132951197263680 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 14:53:27.298826 132951197263680 checkpointing.py:677] Setting up checkpoint logger...
I0417 14:53:27.298897 132951197263680 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0417 14:53:27.298943 132951197263680 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 14:53:27.299269 132951197263680 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 0x78ea80f1b920>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 14:53:28.567200 132951197263680 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 14:53:28.567432 132951197263680 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78d6d45c0170>}, handler_registry=None
I0417 14:53:28.567672 132951197263680 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78d6d45c0170>` 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 14:53:28.567722 132951197263680 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78d6d45c21e0>` 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 14:53:28.567756 132951197263680 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 0x78d6d45c0170>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78d6d45c0170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78d6d45c21e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78d6d45c21e0>}).
I0417 14:53:28.568113 132951197263680 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 14:53:28.568186 132951197263680 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78d5fc50ca40> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 14:53:30.884682 132832852694784 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437348952059370, 'commit_timestamp_nsecs': 1776437390670671068, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA
I0417 14:53:31.126063 132832852694784 checkpoint_manager.py:1871] Missing metrics for step 0
I0417 14:53:31.571973 132951197263680 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints
I0417 14:53:31.574509 132951197263680 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78d6d4521d30>
I0417 14:53:31.574623 132951197263680 checkpointing.py:301] Checkpoint manager created!
I0417 14:53:32.482084 132951197263680 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 14:53:32.482200 132951197263680 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:53:32.863011 132951197263680 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:53:32.863103 132951197263680 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:53:32.879321 132951197263680 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:53:32.879377 132951197263680 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:53:32.902866 132951197263680 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 14:53:32.902930 132951197263680 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:53:32.919193 132951197263680 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 14:53:32.919252 132951197263680 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:53:32.935518 132951197263680 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 14:53:32.935580 132951197263680 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:53:32.961009 132951197263680 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 14:53:32.961077 132951197263680 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:53:32.981877 132951197263680 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 14:53:32.981935 132951197263680 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 14:53:33.198277 132951197263680 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 14:53:33.198383 132951197263680 checkpointing.py:581] restoring from this run's directory step 9
I0417 14:53:33.439360 132951197263680 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/9.
I0417 14:53:33.687031 132951197263680 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/9.
I0417 14:53:33.687466 132951197263680 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/9.
I0417 14:53:35.095393    1407 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 14:53:40.711827 132951197263680 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 281.077 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.622203350067139 s) (per-host)
I0417 14:53:40.725347 132951197263680 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 261.811 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.0447838306427 s) (per-host)
I0417 14:53:42.229025 132951197263680 event_tracking.py:138] [process=0] [sync] Finished load in 8.79 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/9
I0417 14:53:42.229188 132951197263680 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776437613.4392614, 'checkpointer_duration_secs': 8.789896488189697, 'checkpoint_manager_start_time': 1776437613.198627, 'checkpoint_manager_duration_secs': 9.0305335521698}
fsdp: 32
I0417 14:53:42.646048 132951197263680 maxtext_utils.py:1651]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0417 14:53:42.646183 132951197263680 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 14:53:42.646243 132951197263680 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 14:53:42.646302 132951197263680 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 14:53:42.646353 132951197263680 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 14:53:42.646391 132951197263680 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 14:53:42.646442 132951197263680 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 14:53:42.646495 132951197263680 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 14:53:42.646534 132951197263680 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 14:53:42.646570 132951197263680 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 14:53:42.646616 132951197263680 maxtext_utils.py:1651]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0417 14:53:42.646664 132951197263680 maxtext_utils.py:1651]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]

    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0417 14:53:43.143774 132951197263680 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:53:43.143881 132951197263680 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:53:43.159933 132951197263680 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:53:43.159995 132951197263680 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:53:54.004003 132951197263680 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 14:53:54.011316 132951197263680 metric_logger.py:301] number parameters: 1.104 billion
I0417 14:54:05.372670 132951197263680 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0417 14:54:05.526953 132951197263680 checkpointing.py:776] Waited 0.1542670726776123 seconds for step 10 to finish before starting checkpointing.
I0417 14:54:05.529248 132951197263680 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:54:05.531355 132951197263680 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10
I0417 14:54:05.532784 132951197263680 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10.
I0417 14:54:06.285577 132951197263680 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 14:54:06.298188 132951197263680 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:54:06.298255 132951197263680 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 14:54:06.535597 132832240322304 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10
I0417 14:54:06.574000 132951197263680 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276531s
I0417 14:54:06.574787 132951197263680 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.482 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28138089179992676 s) (per-host)
I0417 14:54:06.574885 132951197263680 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281497s (batch_requests_ready=0.002244s, total_serialization_initiated=0.278529s, others=0.000725s)
I0417 14:54:06.574966 132951197263680 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285462s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.285444)
I0417 14:54:06.575731 132951197263680 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10.
I0417 14:54:06.576027 132832248715008 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:14:06.576004
I0417 14:54:06.584044 132951197263680 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:54:06.584351 132827970533120 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:54:06.584507 132951197263680 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776437645.5292304, 'wait_for_prev_duration_secs': 6.103515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776437645.5313926, 'checkpointer_blocking_duration_secs': 1.0447437763214111, 'get_old_steps_start_time': 1776437646.576161, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776437645.5273216, 'checkpoint_manager_blocking_duration_secs': 1.0571479797363281}
I0417 14:54:06.584619 132951197263680 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0417 14:54:06.584670 132951197263680 max_utils.py:750] 
Memstats: After params initialized:
I0417 14:54:06.584721 132951197263680 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 14:54:06.584756 132951197263680 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 14:54:06.584782 132951197263680 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 14:54:06.584808 132951197263680 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 14:54:06.916742 132951197263680 metric_logger.py:196] completed step: 10, seconds: 11.061, TFLOP/s/device: 1.228, Tokens/s/device: 185.152, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0417 14:54:07.061995 132951197263680 metric_logger.py:196] completed step: 11, seconds: 1.542, TFLOP/s/device: 8.810, Tokens/s/device: 1328.012, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0417 14:54:07.209345 132951197263680 metric_logger.py:196] completed step: 12, seconds: 0.011, TFLOP/s/device: 1254.930, Tokens/s/device: 189156.738, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0417 14:54:07.357134 132951197263680 metric_logger.py:196] completed step: 13, seconds: 0.145, TFLOP/s/device: 93.550, Tokens/s/device: 14100.896, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0417 14:54:07.365820 132951197263680 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0417 14:54:07.651175 132951197263680 checkpointing.py:776] Waited 0.28519701957702637 seconds for step 15 to finish before starting checkpointing.
I0417 14:54:07.654181 132951197263680 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:54:08.492765 132827962140416 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437648093608111, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776437648093608111, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10/_CHECKPOINT_METADATA
I0417 14:54:08.774021 132829027489536 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10/items
I0417 14:54:13.048516 132829010704128 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10/items/array_metadatas/process_0
I0417 14:54:52.840535 132829002311424 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.082933s (commit=40.875228s, array_metadata_write=1.207705s)
I0417 14:54:52.841896 132832248715008 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.934 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.54843831062317 s) (per-host)
I0417 14:54:52.842063 132832248715008 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.265951s.
I0417 14:54:54.398094 132827962140416 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776437648093608111, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10/_CHECKPOINT_METADATA
I0417 14:54:54.604034 132832248715008 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:54:54.740281 132832248715008 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000428s.
I0417 14:54:55.693362 132832248715008 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.937128s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10/items
I0417 14:54:55.695735 132832248715008 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10/items
I0417 14:54:56.216748 132832248715008 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10
I0417 14:54:57.606421 132832248715008 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10`.
I0417 14:54:57.607455 132832248715008 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.08 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/10
I0417 14:54:57.609120 132832248715008 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.033007s.
I0417 14:54:57.609313 132827970533120 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:54:57.609406 132827970533120 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:54:57.609505 132827970533120 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:54:57.614041 132827970533120 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0417 14:54:57.614342 132951197263680 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0417 14:54:57.614535 132951197263680 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.960351 seconds. If this number is high, consider checkpointing less frequently.
I0417 14:54:57.616432 132951197263680 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15
I0417 14:54:57.618859 132951197263680 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15.
I0417 14:54:58.754636 132951197263680 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:54:58.754771 132951197263680 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 14:54:58.782667 132951197263680 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029042s
I0417 14:54:58.783568 132951197263680 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.674 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03792572021484375 s) (per-host)
I0417 14:54:58.783670 132951197263680 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038049s (batch_requests_ready=0.004038s, total_serialization_initiated=0.033180s, others=0.000830s)
I0417 14:54:58.783790 132951197263680 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042912s (all_items=0.000024s, per_item={'items': '0.00002384'}, temp_paths=0.042888)
I0417 14:54:58.784763 132951197263680 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15.
I0417 14:54:58.785167 132827970533120 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:14:58.785129
I0417 14:54:58.787687 132951197263680 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:54:58.788052 132829019096832 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:54:58.788206 132951197263680 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776437647.654138, 'wait_for_prev_duration_secs': 49.96035099029541, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776437697.6164992, 'checkpointer_blocking_duration_secs': 1.1688563823699951, 'get_old_steps_start_time': 1776437698.785387, 'get_old_steps_duration_secs': 4.673004150390625e-05, 'checkpoint_manager_blocking_start_time': 1776437647.6518638, 'checkpoint_manager_blocking_duration_secs': 51.13628911972046}
I0417 14:54:58.788432 132951197263680 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0417 14:54:58.988539 132832248715008 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15
I0417 14:55:00.748175 132827945355008 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15/items
I0417 14:55:03.941518 132831154001664 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15/items/array_metadatas/process_0
I0417 14:55:19.222388 132951197263680 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.236, Tokens/s/device: 13902.749, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0417 14:55:19.231586 132951197263680 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.116, Tokens/s/device: 13884.652, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0417 14:55:19.377098 132951197263680 metric_logger.py:196] completed step: 16, seconds: 71.864, TFLOP/s/device: 0.189, Tokens/s/device: 28.498, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0417 14:55:19.524631 132951197263680 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1864.315, Tokens/s/device: 281009.879, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0417 14:55:19.671624 132951197263680 metric_logger.py:196] completed step: 18, seconds: 0.146, TFLOP/s/device: 92.748, Tokens/s/device: 13979.999, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0417 14:55:19.818460 132951197263680 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0417 14:55:19.819195 132951197263680 checkpointing.py:776] Waited 0.0007565021514892578 seconds for step 19 to finish before starting checkpointing.
I0417 14:55:19.821297 132951197263680 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:55:43.198987 132829027489536 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.016042s (commit=39.388538s, array_metadata_write=1.627504s)
I0417 14:55:43.200049 132827970533120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.533 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.454389333724976 s) (per-host)
I0417 14:55:43.200156 132827970533120 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.414842s.
I0417 14:55:44.896637 132827970533120 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:55:45.041040 132827970533120 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000262s.
I0417 14:55:45.520469 132827970533120 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.445733s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15/items
I0417 14:55:45.522274 132827970533120 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15/items
I0417 14:55:46.039904 132827970533120 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15
I0417 14:55:47.460676 132827970533120 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15`.
I0417 14:55:47.461501 132827970533120 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.84 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/15
I0417 14:55:47.462788 132827970533120 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.677471s.
I0417 14:55:47.463066 132829019096832 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:55:47.463190 132829019096832 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:55:47.463255 132829019096832 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:55:47.465167 132829019096832 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0417 14:55:47.465375 132951197263680 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0417 14:55:47.465511 132951197263680 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.644218 seconds. If this number is high, consider checkpointing less frequently.
I0417 14:55:47.466879 132951197263680 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19
I0417 14:55:47.469010 132951197263680 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19.
I0417 14:55:48.174909 132951197263680 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:55:48.175006 132951197263680 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 14:55:48.200394 132951197263680 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026437s
I0417 14:55:48.201142 132951197263680 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.970 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030869722366333008 s) (per-host)
I0417 14:55:48.201212 132951197263680 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030958s (batch_requests_ready=0.001850s, total_serialization_initiated=0.028409s, others=0.000700s)
I0417 14:55:48.201286 132951197263680 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034895s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.034879)
I0417 14:55:48.201965 132951197263680 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.73 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19.
I0417 14:55:48.202252 132831158187776 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 15:15:48.202215
I0417 14:55:48.204308 132951197263680 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:55:48.204543 132829019096832 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:55:48.204715 132951197263680 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776437719.8212636, 'wait_for_prev_duration_secs': 27.64421844482422, 'time_between_consecutive_saves_sec': 22.207145929336548, 'checkpointer_blocking_start_time': 1776437747.4669251, 'checkpointer_blocking_duration_secs': 0.7354633808135986, 'get_old_steps_start_time': 1776437748.2024078, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776437719.8194308, 'checkpoint_manager_blocking_duration_secs': 28.38524842262268}
I0417 14:55:48.204847 132951197263680 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0417 14:55:48.204898 132951197263680 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:55:48.410948 132827970533120 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19
I0417 14:55:50.190300 132831678273280 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19/items
I0417 14:55:53.043472 132832240322304 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19/items/array_metadatas/process_0
I0417 14:56:31.700805 132831166580480 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.388259s (commit=38.752292s, array_metadata_write=1.635967s)
I0417 14:56:31.701938 132831158187776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 36.286 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.53161692619324 s) (per-host)
I0417 14:56:31.702059 132831158187776 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.499694s.
I0417 14:56:33.116174 132831158187776 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000276s.
I0417 14:56:33.363880 132831158187776 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:56:33.995395 132831158187776 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.466285s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19/items
I0417 14:56:33.997099 132831158187776 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19/items
I0417 14:56:34.512638 132831158187776 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19
I0417 14:56:36.764473 132831158187776 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19`.
I0417 14:56:36.765322 132831158187776 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.30 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_16_zarr3_true/checkpoints/19
I0417 14:56:36.766990 132831158187776 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.564623s.
I0417 14:56:36.767123 132829019096832 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:56:36.767183 132829019096832 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:56:36.767230 132829019096832 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:56:36.769157 132829019096832 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0417 14:56:36.769323 132951197263680 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0417 14:56:36.769468 132951197263680 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:56:36.770562 132951197263680 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.163, Tokens/s/device: 13891.809, 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: Fri Apr 17 14:56:48 UTC 2026
EXIT_CODE=0