XPK Start: Thu Apr 23 22:16:15 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 22:16:40.611459: 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 22:16:40.825693 134268608575296 max_utils.py:273] Attempting to initialize the jax distributed system...
I0423 22:16:49.867836 134268608575296 distributed.py:149] Starting JAX distributed service on [::]:8482
I0423 22:16:49.870209 134268608575296 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-ktnm9-slice-job-0-0.mt-15-ocdbt-true-rs-ktnm9:8482
I0423 22:16:50.794743 134268608575296 max_utils.py:284] Jax distributed system initialized!
I0423 22:16:56.784036 134268608575296 max_utils.py:800] System Information: Jax Version: 0.9.2
I0423 22:16:56.784162 134268608575296 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0423 22:16:56.784206 134268608575296 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 22:16:56.784242 134268608575296 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 22:16:57.490184 134268608575296 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 22:16:57.490467 134268608575296 checkpointing.py:677] Setting up checkpoint logger...
I0423 22:16:57.490522 134268608575296 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0423 22:16:57.490571 134268608575296 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0423 22:16:57.490914 134268608575296 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 0x7a1d3ab1ee40>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 22:16:58.796202 134268608575296 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0423 22:16:58.796438 134268608575296 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a09503df470>}, handler_registry=None
I0423 22:16:58.796681 134268608575296 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a09503df470>` 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 22:16:58.796731 134268608575296 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0950483170>` 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 22:16:58.796767 134268608575296 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 0x7a09503df470>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a09503df470>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0950483170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0950483170>}).
I0423 22:16:58.797086 134268608575296 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0423 22:16:58.797175 134268608575296 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a09502932e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0423 22:17:00.877821 134148001888000 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776982359431072390, 'commit_timestamp_nsecs': 1776982402136868967, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA
I0423 22:17:01.118563 134147985102592 checkpoint_manager.py:1871] Missing metrics for step 9
I0423 22:17:01.531680 134268608575296 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_15_ocdbt_true/checkpoints
I0423 22:17:01.694473 134268608575296 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a0950479f10>
I0423 22:17:01.694660 134268608575296 checkpointing.py:301] Checkpoint manager created!
I0423 22:17:02.632500 134268608575296 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 22:17:02.632628 134268608575296 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 22:17:03.011740 134268608575296 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 22:17:03.011832 134268608575296 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 22:17:03.028318 134268608575296 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 22:17:03.028382 134268608575296 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 22:17:03.051959 134268608575296 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 22:17:03.052030 134268608575296 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:17:03.068692 134268608575296 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 22:17:03.068755 134268608575296 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:17:03.085184 134268608575296 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 22:17:03.085244 134268608575296 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:17:03.110008 134268608575296 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 22:17:03.110080 134268608575296 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 22:17:03.131273 134268608575296 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 22:17:03.131350 134268608575296 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0423 22:17:03.343456 134268608575296 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 22:17:03.343564 134268608575296 checkpointing.py:581] restoring from this run's directory step 9
I0423 22:17:03.589066 134268608575296 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/9.
I0423 22:17:03.840233 134268608575296 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_15_ocdbt_true/checkpoints/9.
I0423 22:17:03.840674 134268608575296 event_tracking.py:125] [process=6] [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_15_ocdbt_true/checkpoints/9.
I0423 22:17:05.735138 1438 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0423 22:17:12.351476 134268608575296 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 238.626 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.6223742961883545 s) (per-host)
I0423 22:17:12.357126 134268608575296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 225.645 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.013659954071045 s) (per-host)
I0423 22:17:12.784753 134268608575296 event_tracking.py:138] [process=6] [sync] Finished load in 9.20 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/9
I0423 22:17:12.784912 134268608575296 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_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776982623.588968, 'checkpointer_duration_secs': 9.195912599563599, 'checkpoint_manager_start_time': 1776982623.3438141, 'checkpoint_manager_duration_secs': 9.44106936454773}
fsdp: 32
I0423 22:17:13.204986 134268608575296 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0423 22:17:13.205130 134268608575296 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 22:17:13.205188 134268608575296 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 22:17:13.205249 134268608575296 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 22:17:13.205301 134268608575296 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 22:17:13.205343 134268608575296 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 22:17:13.205395 134268608575296 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 22:17:13.205446 134268608575296 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 22:17:13.205486 134268608575296 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 22:17:13.205523 134268608575296 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 22:17:13.205569 134268608575296 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0423 22:17:13.205617 134268608575296 maxtext_utils.py:1668] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0423 22:17:13.699340 134268608575296 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 22:17:13.699438 134268608575296 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 22:17:13.714966 134268608575296 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 22:17:13.715028 134268608575296 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 22:17:24.764840 134268608575296 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 22:17:24.767068 134268608575296 metric_logger.py:301] number parameters: 1.104 billion
I0423 22:17:36.125737 134268608575296 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0423 22:17:36.476055 134268608575296 checkpointing.py:776] Waited 0.3502933979034424 seconds for step 10 to finish before starting checkpointing.
I0423 22:17:36.478532 134268608575296 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 22:17:36.480642 134268608575296 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0423 22:17:36.482502 134268608575296 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/10.
I0423 22:17:36.970472 134268608575296 signaling_client.py:364] Using JaxDistributedSignalingClient
I0423 22:17:36.971325 134268608575296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 22:17:36.971382 134268608575296 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 22:17:37.244821 134268608575296 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274425s
I0423 22:17:37.244992 134268608575296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.543 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27831101417541504 s) (per-host)
I0423 22:17:37.245044 134268608575296 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278374s (batch_requests_ready=0.002156s, total_serialization_initiated=0.276147s, others=0.000071s)
I0423 22:17:37.245151 134268608575296 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.282395s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.282378)
I0423 22:17:37.245853 134268608575296 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/10.
I0423 22:17:37.246204 134147397908224 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 22:37:37.246166
I0423 22:17:37.256484 134268608575296 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 22:17:37.256747 134143623026432 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0423 22:17:37.256873 134268608575296 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776982656.4785132, 'wait_for_prev_duration_secs': 6.079673767089844e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776982656.4806805, 'checkpointer_blocking_duration_secs': 0.7656869888305664, 'get_old_steps_start_time': 1776982657.2463927, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1776982656.4765549, 'checkpoint_manager_blocking_duration_secs': 0.7802829742431641}
I0423 22:17:37.256974 134268608575296 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0423 22:17:37.257022 134268608575296 max_utils.py:750]
Memstats: After params initialized:
I0423 22:17:37.257069 134268608575296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0423 22:17:37.257112 134268608575296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0423 22:17:37.257140 134268608575296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0423 22:17:37.257164 134268608575296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0423 22:17:37.571951 134268608575296 metric_logger.py:196] completed step: 10, seconds: 11.359, TFLOP/s/device: 1.196, Tokens/s/device: 180.305, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0423 22:17:37.760662 134268608575296 metric_logger.py:196] completed step: 11, seconds: 1.445, TFLOP/s/device: 9.405, Tokens/s/device: 1417.610, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0423 22:17:37.908042 134268608575296 metric_logger.py:196] completed step: 12, seconds: 0.041, TFLOP/s/device: 328.485, Tokens/s/device: 49512.850, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0423 22:17:38.055620 134268608575296 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.621, Tokens/s/device: 13357.945, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0423 22:17:38.060124 134268608575296 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0423 22:17:38.349482 134268608575296 checkpointing.py:776] Waited 0.28932690620422363 seconds for step 15 to finish before starting checkpointing.
I0423 22:17:38.352514 134268608575296 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 22:17:43.638608 134146290595584 array_metadata_store.py:203] [process=6][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_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6
I0423 22:18:19.670801 134147397908224 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.989 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.70409846305847 s) (per-host)
I0423 22:18:19.670911 134147397908224 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.424592s.
I0423 22:18:28.399849 134147397908224 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.153512s.
I0423 22:18:28.400158 134143623026432 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 22:18:28.400282 134143623026432 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0423 22:18:28.400331 134143623026432 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0423 22:18:28.404711 134143623026432 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0423 22:18:28.404875 134268608575296 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0423 22:18:28.405006 134268608575296 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.052493 seconds. If this number is high, consider checkpointing less frequently.
I0423 22:18:28.406976 134268608575296 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0423 22:18:28.409417 134268608575296 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/15.
I0423 22:18:28.706918 134268608575296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 22:18:28.707009 134268608575296 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 22:18:28.747981 134268608575296 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042155s
I0423 22:18:28.748158 134268608575296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.853 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.045566558837890625 s) (per-host)
I0423 22:18:28.748212 134268608575296 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045631s (batch_requests_ready=0.001762s, total_serialization_initiated=0.043797s, others=0.000072s)
I0423 22:18:28.748306 134268608575296 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050668s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.050655)
I0423 22:18:28.749021 134268608575296 event_tracking.py:125] [process=6] [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_15_ocdbt_true/checkpoints/15.
I0423 22:18:28.749349 134143623026432 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 22:38:28.749316
I0423 22:18:28.751389 134268608575296 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 22:18:28.751622 134140401800960 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0423 22:18:28.751745 134268608575296 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776982658.352484, 'wait_for_prev_duration_secs': 50.05249309539795, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776982708.4070148, 'checkpointer_blocking_duration_secs': 0.34249258041381836, 'get_old_steps_start_time': 1776982708.749535, 'get_old_steps_duration_secs': 3.600120544433594e-05, 'checkpoint_manager_blocking_start_time': 1776982658.3501348, 'checkpoint_manager_blocking_duration_secs': 50.40157723426819}
I0423 22:18:28.751859 134268608575296 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0423 22:18:28.752810 134268608575296 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.132, Tokens/s/device: 13887.194, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0423 22:18:28.759122 134268608575296 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.162, Tokens/s/device: 13891.621, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0423 22:18:34.929687 134146290595584 array_metadata_store.py:203] [process=6][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_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6
I0423 22:18:49.977040 134268608575296 metric_logger.py:196] completed step: 16, seconds: 50.697, TFLOP/s/device: 0.268, Tokens/s/device: 40.397, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0423 22:18:50.124130 134268608575296 metric_logger.py:196] completed step: 17, seconds: 21.072, TFLOP/s/device: 0.645, Tokens/s/device: 97.191, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0423 22:18:50.271553 134268608575296 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.293, Tokens/s/device: 13459.251, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0423 22:18:50.418135 134268608575296 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0423 22:18:50.418839 134268608575296 checkpointing.py:776] Waited 0.0007398128509521484 seconds for step 19 to finish before starting checkpointing.
I0423 22:18:50.420890 134268608575296 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 22:19:11.825575 134143623026432 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.630 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.12292957305908 s) (per-host)
I0423 22:19:11.825701 134143623026432 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.076239s.
I0423 22:19:19.946251 134143623026432 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.196773s.
I0423 22:19:19.946484 134140401800960 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 22:19:19.946555 134140401800960 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0423 22:19:19.946615 134140401800960 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0423 22:19:19.949077 134140401800960 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0423 22:19:19.949271 134268608575296 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0423 22:19:19.949399 134268608575296 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.528514 seconds. If this number is high, consider checkpointing less frequently.
I0423 22:19:19.950952 134268608575296 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0423 22:19:19.952958 134268608575296 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/19.
I0423 22:19:20.661184 134268608575296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 22:19:20.661288 134268608575296 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 22:19:20.703391 134268608575296 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043255s
I0423 22:19:20.703557 134268608575296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.008 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04673337936401367 s) (per-host)
I0423 22:19:20.703607 134268608575296 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046794s (batch_requests_ready=0.001680s, total_serialization_initiated=0.045048s, others=0.000066s)
I0423 22:19:20.703695 134268608575296 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051004s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.050993)
I0423 22:19:20.704403 134268608575296 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260423_211617/linen_xpk_test_pipeline_scan_nnx_20260423_211617_15_ocdbt_true/checkpoints/19.
I0423 22:19:20.704685 134140401800960 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 22:39:20.704652
I0423 22:19:20.706593 134268608575296 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 22:19:20.706796 134146290595584 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0423 22:19:20.706908 134268608575296 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776982730.4208589, 'wait_for_prev_duration_secs': 29.52851414680481, 'time_between_consecutive_saves_sec': 22.016105890274048, 'checkpointer_blocking_start_time': 1776982759.9509933, 'checkpointer_blocking_duration_secs': 0.7538332939147949, 'get_old_steps_start_time': 1776982760.7048454, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776982730.4190648, 'checkpoint_manager_blocking_duration_secs': 30.28781533241272}
I0423 22:19:20.707020 134268608575296 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0423 22:19:20.707062 134268608575296 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 22:19:26.236387 134141492328192 array_metadata_store.py:203] [process=6][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_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6
I0423 22:20:02.128144 134140401800960 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.089 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.471259355545044 s) (per-host)
I0423 22:20:02.128250 134140401800960 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.423455s.
I0423 22:20:10.096441 134140401800960 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.391629s.
I0423 22:20:10.096736 134146290595584 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 22:20:10.096864 134146290595584 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0423 22:20:10.096906 134146290595584 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0423 22:20:10.098702 134146290595584 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0423 22:20:10.098858 134268608575296 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0423 22:20:10.099005 134268608575296 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 22:20:10.100016 134268608575296 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.315, Tokens/s/device: 13914.650, 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 22:20:20 UTC 2026
EXIT_CODE=0