MaxView

XPK Start: Thu Apr 23 22:43:32 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. 
2026-04-23 22:43:56.939791: 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)
I0423 22:43:57.151864 136191410386752 max_utils.py:273] Attempting to initialize the jax distributed system...
I0423 22:44:06.191043 136191410386752 distributed.py:149] Starting JAX distributed service on [::]:8482
I0423 22:44:06.193313 136191410386752 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-rs-uqcqv-slice-job-0-0.mt-16-zarr3-false-rs-uqcqv:8482
I0423 22:44:07.270931 136191410386752 max_utils.py:284] Jax distributed system initialized!
I0423 22:44:12.440663 136191410386752 max_utils.py:800] System Information: Jax Version: 0.9.2
I0423 22:44:12.440768 136191410386752 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0423 22:44:12.440809 136191410386752 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
I0423 22:44:12.440845 136191410386752 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 22:44:13.137195 136191410386752 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 22:44:13.137471 136191410386752 checkpointing.py:677] Setting up checkpoint logger...
I0423 22:44:13.137527 136191410386752 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0423 22:44:13.137572 136191410386752 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0423 22:44:13.137938 136191410386752 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 0x7bdcd0e4c9b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 22:44:14.796765 136191410386752 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0423 22:44:14.797023 136191410386752 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 0x7bc97018f800>}, handler_registry=None
I0423 22:44:14.797261 136191410386752 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc97018f800>` 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`.
I0423 22:44:14.797308 136191410386752 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc9701934a0>` 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`.
I0423 22:44:14.797344 136191410386752 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 0x7bc97018f800>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc97018f800>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc9701934a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc9701934a0>}).
I0423 22:44:14.797663 136191410386752 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0423 22:44:14.797738 136191410386752 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bc9403972e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0423 22:44:16.374958 136070939277056 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776983990497477395, 'commit_timestamp_nsecs': 1776984037386100227, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0423 22:44:16.618654 136070939277056 checkpoint_manager.py:1871] Missing metrics for step 0
I0423 22:44:17.044026 136191410386752 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints
I0423 22:44:17.356175 136191410386752 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_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bc970191940>
I0423 22:44:17.356350 136191410386752 checkpointing.py:301] Checkpoint manager created!
I0423 22:44:18.286718 136191410386752 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 22:44:18.286839 136191410386752 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 22:44:18.674175 136191410386752 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 22:44:18.674267 136191410386752 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 22:44:18.690523 136191410386752 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 22:44:18.690586 136191410386752 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 22:44:18.714140 136191410386752 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 22:44:18.714208 136191410386752 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:44:18.730660 136191410386752 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 22:44:18.730723 136191410386752 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:44:18.746995 136191410386752 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 22:44:18.747056 136191410386752 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:44:18.771740 136191410386752 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 22:44:18.771818 136191410386752 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:44:18.792858 136191410386752 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 22:44:18.792945 136191410386752 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0423 22:44:19.009221 136191410386752 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 22:44:19.009334 136191410386752 checkpointing.py:581] restoring from this run's directory step 9
I0423 22:44:19.244834 136191410386752 event_tracking.py:70] [process=3] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/9.
I0423 22:44:19.495968 136191410386752 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/9.
I0423 22:44:19.496393 136191410386752 event_tracking.py:125] [process=3] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/9.
I0423 22:44:21.364895    1438 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0423 22:44:27.694171 136191410386752 jax_array_handlers.py:843] [process=3] /jax/orbax/read/worker/io/requested throughput: 249.448 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.3350746631622314 s) (per-host)
I0423 22:44:27.698997 136191410386752 base_pytree_checkpoint_handler.py:130] [process=3] /jax/checkpoint/read/gbytes_per_sec: 233.687 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.772275447845459 s) (per-host)
I0423 22:44:28.567547 136191410386752 event_tracking.py:138] [process=3] [sync] Finished load in 9.32 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/9
I0423 22:44:28.567701 136191410386752 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776984259.244737, 'checkpointer_duration_secs': 9.322932958602905, 'checkpoint_manager_start_time': 1776984259.0095835, 'checkpoint_manager_duration_secs': 9.558088541030884}
fsdp: 32
I0423 22:44:28.981631 136191410386752 maxtext_utils.py:1668]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0423 22:44:28.981758 136191410386752 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)
I0423 22:44:28.981819 136191410386752 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)
I0423 22:44:28.981880 136191410386752 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')
I0423 22:44:28.981952 136191410386752 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)
I0423 22:44:28.982002 136191410386752 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)
I0423 22:44:28.982059 136191410386752 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)
I0423 22:44:28.982112 136191410386752 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')
I0423 22:44:28.982153 136191410386752 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)
I0423 22:44:28.982190 136191410386752 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)
I0423 22:44:28.982236 136191410386752 maxtext_utils.py:1668]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)

I0423 22:44:28.982280 136191410386752 maxtext_utils.py:1668]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0423 22:44:29.463882 136191410386752 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 22:44:29.463982 136191410386752 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 22:44:29.479480 136191410386752 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 22:44:29.479547 136191410386752 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 22:44:40.573347 136191410386752 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.
I0423 22:44:40.574172 136191410386752 metric_logger.py:301] number parameters: 1.104 billion
I0423 22:44:51.763485 136191410386752 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0423 22:44:52.399401 136191410386752 checkpointing.py:776] Waited 0.6358959674835205 seconds for step 10 to finish before starting checkpointing.
I0423 22:44:52.401721 136191410386752 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 22:44:52.403625 136191410386752 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 10
I0423 22:44:52.405131 136191410386752 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/10.
I0423 22:44:52.721027 136191410386752 signaling_client.py:364] Using JaxDistributedSignalingClient
I0423 22:44:52.722064 136191410386752 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 22:44:52.722120 136191410386752 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
I0423 22:44:52.998156 136191410386752 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277215s
I0423 22:44:52.998324 136191410386752 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.489 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2810509204864502 s) (per-host)
I0423 22:44:52.998378 136191410386752 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281116s (batch_requests_ready=0.002156s, total_serialization_initiated=0.278886s, others=0.000073s)
I0423 22:44:52.998466 136191410386752 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285917s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.285900)
I0423 22:44:52.999235 136191410386752 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.60 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/10.
I0423 22:44:52.999626 136070888920832 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 23:04:52.999588
I0423 22:44:53.008508 136191410386752 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 22:44:53.008765 136066526844672 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0423 22:44:53.008888 136191410386752 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776984292.4017031, 'wait_for_prev_duration_secs': 6.580352783203125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776984292.4036627, 'checkpointer_blocking_duration_secs': 0.5960657596588135, 'get_old_steps_start_time': 1776984292.9997497, 'get_old_steps_duration_secs': 3.2901763916015625e-05, 'checkpoint_manager_blocking_start_time': 1776984292.3998065, 'checkpoint_manager_blocking_duration_secs': 0.6090455055236816}
I0423 22:44:53.009007 136191410386752 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0423 22:44:53.009059 136191410386752 max_utils.py:750] 
Memstats: After params initialized:
I0423 22:44:53.009114 136191410386752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0))
I0423 22:44:53.009147 136191410386752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0))
I0423 22:44:53.009175 136191410386752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0))
I0423 22:44:53.009200 136191410386752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0))
I0423 22:44:53.317794 136191410386752 metric_logger.py:196] completed step: 10, seconds: 11.189, TFLOP/s/device: 1.214, Tokens/s/device: 183.034, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0423 22:44:53.900745 136191410386752 metric_logger.py:196] completed step: 11, seconds: 1.553, TFLOP/s/device: 8.750, Tokens/s/device: 1318.943, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0423 22:44:54.048138 136191410386752 metric_logger.py:196] completed step: 12, seconds: 0.437, TFLOP/s/device: 31.095, Tokens/s/device: 4686.917, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0423 22:44:54.195617 136191410386752 metric_logger.py:196] completed step: 13, seconds: 0.152, TFLOP/s/device: 89.382, Tokens/s/device: 13472.621, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0423 22:44:54.200271 136191410386752 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0423 22:44:54.489885 136191410386752 checkpointing.py:776] Waited 0.2895798683166504 seconds for step 15 to finish before starting checkpointing.
I0423 22:44:54.493093 136191410386752 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 22:44:57.863690 136066535237376 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/10/items/array_metadatas/process_3
I0423 22:45:34.649977 136070888920832 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.670 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.932650566101074 s) (per-host)
I0423 22:45:34.650097 136070888920832 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.649833s.
I0423 22:45:42.949165 136070888920832 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 49.948884s.
I0423 22:45:42.949386 136066526844672 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 22:45:42.949519 136066526844672 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0423 22:45:42.949575 136066526844672 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0423 22:45:42.951719 136066526844672 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0423 22:45:42.951883 136191410386752 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.
W0423 22:45:42.952026 136191410386752 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.458934 seconds. If this number is high, consider checkpointing less frequently.
I0423 22:45:42.954148 136191410386752 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 15
I0423 22:45:42.956663 136191410386752 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/15.
I0423 22:45:43.264405 136191410386752 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 22:45:43.264502 136191410386752 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
I0423 22:45:43.290388 136191410386752 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027135s
I0423 22:45:43.290559 136191410386752 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 50.501 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03054523468017578 s) (per-host)
I0423 22:45:43.290613 136191410386752 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030608s (batch_requests_ready=0.001762s, total_serialization_initiated=0.028776s, others=0.000070s)
I0423 22:45:43.290710 136191410386752 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035735s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.035722)
I0423 22:45:43.291330 136191410386752 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/15.
I0423 22:45:43.291659 136066552022784 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 23:05:43.291612
I0423 22:45:43.302143 136191410386752 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 22:45:43.302440 136066526844672 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0423 22:45:43.302599 136191410386752 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776984294.4930649, 'wait_for_prev_duration_secs': 48.45893406867981, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776984342.954188, 'checkpointer_blocking_duration_secs': 0.3376181125640869, 'get_old_steps_start_time': 1776984343.2918277, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776984294.490542, 'checkpoint_manager_blocking_duration_secs': 48.81202554702759}
I0423 22:45:43.302706 136191410386752 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0423 22:45:43.303508 136191410386752 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.159, Tokens/s/device: 13891.150, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0423 22:45:43.309578 136191410386752 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.101, Tokens/s/device: 13882.487, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0423 22:45:49.327070 136063355975424 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/15/items/array_metadatas/process_3
I0423 22:46:04.585469 136191410386752 metric_logger.py:196] completed step: 16, seconds: 49.108, TFLOP/s/device: 0.277, Tokens/s/device: 41.704, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0423 22:46:04.732977 136191410386752 metric_logger.py:196] completed step: 17, seconds: 21.130, TFLOP/s/device: 0.643, Tokens/s/device: 96.926, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0423 22:46:04.880252 136191410386752 metric_logger.py:196] completed step: 18, seconds: 0.153, TFLOP/s/device: 89.060, Tokens/s/device: 13424.139, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0423 22:46:05.026995 136191410386752 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0423 22:46:05.027696 136191410386752 checkpointing.py:776] Waited 0.0007607936859130859 seconds for step 19 to finish before starting checkpointing.
I0423 22:46:05.029826 136191410386752 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 22:46:25.509998 136066552022784 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.387 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.24992656707764 s) (per-host)
I0423 22:46:25.510124 136066552022784 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.218351s.
I0423 22:46:33.728232 136066552022784 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 50.436444s.
I0423 22:46:33.728527 136066526844672 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 22:46:33.728644 136066526844672 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0423 22:46:33.728692 136066526844672 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0423 22:46:33.730251 136066526844672 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0423 22:46:33.730432 136191410386752 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.
W0423 22:46:33.730559 136191410386752 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.700740 seconds. If this number is high, consider checkpointing less frequently.
I0423 22:46:33.732097 136191410386752 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 19
I0423 22:46:33.734115 136191410386752 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/19.
I0423 22:46:34.011518 136191410386752 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 22:46:34.011609 136191410386752 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
I0423 22:46:34.037536 136191410386752 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026868s
I0423 22:46:34.037722 136191410386752 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 50.818 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030354976654052734 s) (per-host)
I0423 22:46:34.037776 136191410386752 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030421s (batch_requests_ready=0.001654s, total_serialization_initiated=0.028696s, others=0.000071s)
I0423 22:46:34.037872 136191410386752 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034378s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.034367)
I0423 22:46:34.038562 136191410386752 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.31 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/19.
I0423 22:46:34.038892 136066526844672 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 23:06:34.038856
I0423 22:46:34.042099 136191410386752 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 22:46:34.042337 136063355975424 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0423 22:46:34.042464 136191410386752 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776984365.029796, 'wait_for_prev_duration_secs': 28.700740098953247, 'time_between_consecutive_saves_sec': 22.078036308288574, 'checkpointer_blocking_start_time': 1776984393.7321377, 'checkpointer_blocking_duration_secs': 0.3069169521331787, 'get_old_steps_start_time': 1776984394.0390718, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776984365.0279593, 'checkpoint_manager_blocking_duration_secs': 29.014469623565674}
I0423 22:46:34.042577 136191410386752 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0423 22:46:34.042619 136191410386752 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 22:46:39.350521 136064941410048 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_16_zarr3_false/checkpoints/19/items/array_metadatas/process_3
I0423 22:47:17.856172 136066526844672 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.024 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.84876585006714 s) (per-host)
I0423 22:47:17.856293 136066526844672 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.817274s.
I0423 22:47:23.725487 136066526844672 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 49.686452s.
I0423 22:47:23.725721 136063355975424 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 22:47:23.725782 136063355975424 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0423 22:47:23.725819 136063355975424 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0423 22:47:23.727357 136063355975424 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0423 22:47:23.727552 136191410386752 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.
I0423 22:47:23.727716 136191410386752 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 22:47:23.728723 136191410386752 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.199, Tokens/s/device: 13897.277, 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: Thu Apr 23 22:47:32 UTC 2026
EXIT_CODE=0