MaxView

XPK Start: Tue Apr 21 07:29:03 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 07:29:27.989309: 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 07:29:28.202195 134310155450176 max_utils.py:273] Attempting to initialize the jax distributed system...
I0421 07:29:37.242072 134310155450176 distributed.py:149] Starting JAX distributed service on [::]:8482
I0421 07:29:37.244389 134310155450176 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-n2i0y-slice-job-0-0.mt-15-ocdbt-false-rs-n2i0y:8482
I0421 07:29:38.121561 134310155450176 max_utils.py:284] Jax distributed system initialized!
I0421 07:29:43.069645 134310155450176 max_utils.py:800] System Information: Jax Version: 0.9.2
I0421 07:29:43.069767 134310155450176 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0421 07:29:43.069810 134310155450176 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 07:29:43.069846 134310155450176 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 07:29:43.760076 134310155450176 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 07:29:43.760357 134310155450176 checkpointing.py:677] Setting up checkpoint logger...
I0421 07:29:43.760411 134310155450176 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True
I0421 07:29:43.760456 134310155450176 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0421 07:29:43.760841 134310155450176 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7a273693a0f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 07:29:45.058498 134310155450176 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0421 07:29:45.058750 134310155450176 checkpoint_manager.py:708] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a132058dac0>}, handler_registry=None
I0421 07:29:45.058994 134310155450176 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a132058dac0>` 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 07:29:45.059045 134310155450176 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1320590440>` 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 07:29:45.059082 134310155450176 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 0x7a132058dac0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a132058dac0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1320590440>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1320590440>}).
I0421 07:29:45.059406 134310155450176 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0421 07:29:45.059478 134310155450176 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a1308119800> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0421 07:29:46.957878 134191186425600 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776756432970482776, 'commit_timestamp_nsecs': 1776756446111157077, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/5/_CHECKPOINT_METADATA
I0421 07:29:47.196894 134191186425600 checkpoint_manager.py:1871] Missing metrics for step 5
I0421 07:29:47.816850 134310155450176 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints
I0421 07:29:48.051959 134310155450176 checkpoint_manager.py:929] [process=7][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_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a13207ba870>
I0421 07:29:48.052130 134310155450176 checkpointing.py:301] Checkpoint manager created!
I0421 07:29:48.985115 134310155450176 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 07:29:48.985233 134310155450176 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 07:29:49.367708 134310155450176 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 07:29:49.367821 134310155450176 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 07:29:49.384194 134310155450176 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 07:29:49.384257 134310155450176 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 07:29:49.407958 134310155450176 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 07:29:49.408029 134310155450176 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:29:49.424504 134310155450176 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 07:29:49.424565 134310155450176 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:29:49.441011 134310155450176 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 07:29:49.441076 134310155450176 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:29:49.465778 134310155450176 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 07:29:49.465852 134310155450176 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:29:49.487937 134310155450176 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 07:29:49.488013 134310155450176 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 07:29:49.705166 134310155450176 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 07:29:49.705280 134310155450176 checkpointing.py:581] restoring from this run's directory step 9
I0421 07:29:50.367421 134310155450176 event_tracking.py:70] [process=7] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/9.
I0421 07:29:50.612515 134310155450176 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/9.
I0421 07:29:50.612972 134310155450176 event_tracking.py:125] [process=7] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/9.
I0421 07:29:53.184871    1448 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 07:30:02.658063 134310155450176 jax_array_handlers.py:843] [process=7] /jax/orbax/read/worker/io/requested throughput: 149.823 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.547575235366821 s) (per-host)
I0421 07:30:02.665699 134310155450176 base_pytree_checkpoint_handler.py:130] [process=7] /jax/checkpoint/read/gbytes_per_sec: 144.596 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.944917440414429 s) (per-host)
I0421 07:30:03.663611 134310155450176 event_tracking.py:138] [process=7] [sync] Finished load in 13.30 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/9
I0421 07:30:03.663807 134310155450176 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776756590.3673215, 'checkpointer_duration_secs': 13.296451807022095, 'checkpoint_manager_start_time': 1776756589.705524, 'checkpoint_manager_duration_secs': 13.958251476287842}
fsdp: 32
I0421 07:30:04.080366 134310155450176 maxtext_utils.py:1668]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0421 07:30:04.080507 134310155450176 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 07:30:04.080588 134310155450176 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 07:30:04.080682 134310155450176 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 07:30:04.080782 134310155450176 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 07:30:04.080857 134310155450176 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 07:30:04.080954 134310155450176 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 07:30:04.081051 134310155450176 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 07:30:04.081144 134310155450176 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 07:30:04.081231 134310155450176 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 07:30:04.081333 134310155450176 maxtext_utils.py:1668]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0421 07:30:04.081428 134310155450176 maxtext_utils.py:1668]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0421 07:30:04.565804 134310155450176 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 07:30:04.565898 134310155450176 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 07:30:04.581309 134310155450176 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 07:30:04.581369 134310155450176 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 07:30:15.697499 134310155450176 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 07:30:15.698324 134310155450176 metric_logger.py:301] number parameters: 1.104 billion
I0421 07:30:27.588554 134310155450176 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0421 07:30:27.742141 134310155450176 checkpointing.py:776] Waited 0.15357041358947754 seconds for step 10 to finish before starting checkpointing.
I0421 07:30:27.744424 134310155450176 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 07:30:27.746479 134310155450176 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 10
I0421 07:30:27.747895 134310155450176 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/10.
I0421 07:30:28.081459 134310155450176 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 07:30:28.082477 134310155450176 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 07:30:28.082534 134310155450176 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 07:30:28.363248 134310155450176 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281869s
I0421 07:30:28.363422 134310155450176 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.398 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2857940196990967 s) (per-host)
I0421 07:30:28.363480 134310155450176 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285864s (batch_requests_ready=0.002192s, total_serialization_initiated=0.283592s, others=0.000080s)
I0421 07:30:28.363577 134310155450176 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290174s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.290157)
I0421 07:30:28.364308 134310155450176 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/10.
I0421 07:30:28.364657 134188485285632 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 07:50:28.364619
I0421 07:30:28.368398 134310155450176 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 07:30:28.368687 134185806964480 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0421 07:30:28.368850 134310155450176 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776756627.7444048, 'wait_for_prev_duration_secs': 6.175041198730469e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776756627.7465186, 'checkpointer_blocking_duration_secs': 0.6182475090026855, 'get_old_steps_start_time': 1776756628.3647897, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776756627.742523, 'checkpoint_manager_blocking_duration_secs': 0.6262912750244141}
I0421 07:30:28.368956 134310155450176 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0421 07:30:28.369007 134310155450176 max_utils.py:750] 
Memstats: After params initialized:
I0421 07:30:28.369062 134310155450176 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0))
I0421 07:30:28.369110 134310155450176 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0))
I0421 07:30:28.369144 134310155450176 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0))
I0421 07:30:28.369171 134310155450176 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0))
I0421 07:30:28.679714 134310155450176 metric_logger.py:196] completed step: 10, seconds: 11.890, TFLOP/s/device: 1.143, Tokens/s/device: 172.244, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0421 07:30:28.870719 134310155450176 metric_logger.py:196] completed step: 11, seconds: 1.090, TFLOP/s/device: 12.469, Tokens/s/device: 1879.435, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0421 07:30:29.018398 134310155450176 metric_logger.py:196] completed step: 12, seconds: 0.044, TFLOP/s/device: 311.132, Tokens/s/device: 46897.183, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0421 07:30:29.166136 134310155450176 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.568, Tokens/s/device: 13349.934, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0421 07:30:29.170709 134310155450176 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0421 07:30:29.460282 134310155450176 checkpointing.py:776] Waited 0.2895381450653076 seconds for step 15 to finish before starting checkpointing.
I0421 07:30:29.463217 134310155450176 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 07:30:33.426035 134186892965632 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_7
I0421 07:30:38.813204 134188485285632 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 147.137 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.735541820526123 s) (per-host)
I0421 07:30:38.813328 134188485285632 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.448038s.
I0421 07:30:45.935753 134188485285632 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 17.570425s.
I0421 07:30:45.936033 134185806964480 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 07:30:45.936145 134185806964480 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0421 07:30:45.936191 134185806964480 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0421 07:30:45.938713 134185806964480 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0421 07:30:45.938901 134310155450176 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0421 07:30:45.939034 134310155450176 checkpoint_manager.py:1452] Waiting for previous save to complete took 16.475817 seconds. If this number is high, consider checkpointing less frequently.
I0421 07:30:45.941078 134310155450176 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 15
I0421 07:30:45.943442 134310155450176 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/15.
I0421 07:30:46.229664 134310155450176 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 07:30:46.229773 134310155450176 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 07:30:46.255232 134310155450176 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026584s
I0421 07:30:46.255370 134310155450176 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 50.271 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0306851863861084 s) (per-host)
I0421 07:30:46.255420 134310155450176 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030743s (batch_requests_ready=0.002427s, total_serialization_initiated=0.028252s, others=0.000064s)
I0421 07:30:46.255496 134310155450176 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035665s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.035652)
I0421 07:30:46.256141 134310155450176 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.31 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/15.
I0421 07:30:46.256448 134182076405504 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 07:50:46.256409
I0421 07:30:46.267245 134310155450176 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 07:30:46.267506 134185806964480 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0421 07:30:46.267640 134310155450176 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776756629.4631867, 'wait_for_prev_duration_secs': 16.47581720352173, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776756645.9411588, 'checkpointer_blocking_duration_secs': 0.31543684005737305, 'get_old_steps_start_time': 1776756646.2566204, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776756629.4608812, 'checkpoint_manager_blocking_duration_secs': 16.806724071502686}
I0421 07:30:46.267768 134310155450176 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0421 07:30:46.268693 134310155450176 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.101, Tokens/s/device: 13882.487, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0421 07:30:46.274902 134310155450176 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.951, Tokens/s/device: 13859.845, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0421 07:30:51.625874 134186892965632 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_7
I0421 07:30:56.466341 134182076405504 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 154.233 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.241618394851685 s) (per-host)
I0421 07:30:56.466468 134182076405504 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.209908s.
I0421 07:31:02.926103 134182076405504 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 16.669518s.
I0421 07:31:02.926367 134185806964480 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 07:31:02.926479 134185806964480 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0421 07:31:02.926525 134185806964480 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0421 07:31:02.928164 134185806964480 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0421 07:31:06.877436 134310155450176 metric_logger.py:196] completed step: 16, seconds: 17.103, TFLOP/s/device: 0.794, Tokens/s/device: 119.748, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0421 07:31:07.024491 134310155450176 metric_logger.py:196] completed step: 17, seconds: 20.456, TFLOP/s/device: 0.664, Tokens/s/device: 100.116, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0421 07:31:07.171996 134310155450176 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.127, Tokens/s/device: 13434.265, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0421 07:31:07.318679 134310155450176 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0421 07:31:07.319357 134310155450176 checkpointing.py:776] Waited 0.0006992816925048828 seconds for step 19 to finish before starting checkpointing.
I0421 07:31:07.321439 134310155450176 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 07:31:07.323264 134310155450176 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 19
I0421 07:31:07.324597 134310155450176 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/19.
I0421 07:31:07.644575 134310155450176 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 07:31:07.644681 134310155450176 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 07:31:07.670273 134310155450176 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026532s
I0421 07:31:07.670393 134310155450176 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 51.402 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030010223388671875 s) (per-host)
I0421 07:31:07.670441 134310155450176 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030067s (batch_requests_ready=0.001662s, total_serialization_initiated=0.028343s, others=0.000062s)
I0421 07:31:07.670515 134310155450176 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034266s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.034255)
I0421 07:31:07.671264 134310155450176 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/19.
I0421 07:31:07.671529 134185806964480 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 07:51:07.671507
I0421 07:31:07.675062 134310155450176 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 07:31:07.675345 134181522749184 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0421 07:31:07.675502 134310155450176 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776756667.3214087, 'wait_for_prev_duration_secs': 7.295608520507812e-05, 'time_between_consecutive_saves_sec': 4.393206834793091, 'checkpointer_blocking_start_time': 1776756667.323301, 'checkpointer_blocking_duration_secs': 0.3483254909515381, 'get_old_steps_start_time': 1776756667.6716435, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776756667.3195941, 'checkpoint_manager_blocking_duration_secs': 0.35587525367736816}
I0421 07:31:07.675610 134310155450176 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0421 07:31:07.675655 134310155450176 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 07:31:13.243512 134186892965632 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_7
I0421 07:31:17.692692 134185806964480 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 157.138 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.052269220352173 s) (per-host)
I0421 07:31:17.692833 134185806964480 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.021229s.
I0421 07:31:23.051977 134185806964480 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 15.380357s.
I0421 07:31:23.052268 134181522749184 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 07:31:23.052388 134181522749184 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0421 07:31:23.052428 134181522749184 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0421 07:31:23.055269 134181522749184 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0421 07:31:23.055435 134310155450176 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0421 07:31:23.055579 134310155450176 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 07:31:23.056573 134310155450176 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.387, Tokens/s/device: 13925.531, 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 07:31:33 UTC 2026
EXIT_CODE=0