XPK Start: Sat Apr 25 21:18:30 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. 2026-04-25 21:18:55.561226: 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) I0425 21:18:55.772563 137835727042368 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 21:19:04.813709 137835727042368 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 21:19:04.816072 137835727042368 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-x931h-slice-job-0-0.mt-15-ocdbt-false-rs-x931h:8482 I0425 21:19:05.904501 137835727042368 max_utils.py:284] Jax distributed system initialized! I0425 21:19:11.006742 137835727042368 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 21:19:11.006848 137835727042368 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 21:19:11.006888 137835727042368 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 I0425 21:19:11.006923 137835727042368 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 21:19:11.694690 137835727042368 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 21:19:11.694969 137835727042368 checkpointing.py:677] Setting up checkpoint logger... I0425 21:19:11.695026 137835727042368 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True I0425 21:19:11.695070 137835727042368 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 21:19:11.695436 137835727042368 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 0x7d5bcd118c20>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 21:19:13.012843 137835727042368 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0425 21:19:13.013088 137835727042368 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 0x7d47383f8800>}, handler_registry=None I0425 21:19:13.013349 137835727042368 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d47383f8800>` 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`. I0425 21:19:13.013399 137835727042368 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d47383d7d40>` 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`. I0425 21:19:13.013436 137835727042368 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 0x7d47383f8800>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d47383f8800>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d47383d7d40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d47383d7d40>}). I0425 21:19:13.013764 137835727042368 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 21:19:13.013833 137835727042368 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d47381e3420> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 21:19:14.951864 137716574447360 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777151778896221725, 'commit_timestamp_nsecs': 1777151793675360625, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0425 21:19:15.197184 137716574447360 checkpoint_manager.py:1871] Missing metrics for step 0 I0425 21:19:16.027329 137835727042368 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints I0425 21:19:16.296731 137835727042368 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_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d47383d9730> I0425 21:19:16.296910 137835727042368 checkpointing.py:301] Checkpoint manager created! I0425 21:19:17.227286 137835727042368 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 21:19:17.227403 137835727042368 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 21:19:17.606770 137835727042368 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 21:19:17.606865 137835727042368 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 21:19:17.623309 137835727042368 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 21:19:17.623369 137835727042368 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 21:19:17.646864 137835727042368 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 21:19:17.646932 137835727042368 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 21:19:17.663373 137835727042368 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 21:19:17.663432 137835727042368 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 21:19:17.679753 137835727042368 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 21:19:17.679813 137835727042368 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 21:19:17.704717 137835727042368 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 21:19:17.704789 137835727042368 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 21:19:17.726675 137835727042368 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 21:19:17.726747 137835727042368 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 21:19:17.935700 137835727042368 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0425 21:19:17.935826 137835727042368 checkpointing.py:581] restoring from this run's directory step 9 I0425 21:19:18.182329 137835727042368 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/9. I0425 21:19:18.431849 137835727042368 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/9. I0425 21:19:18.432296 137835727042368 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_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/9. I0425 21:19:21.101682 1449 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 21:19:29.709351 137835727042368 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 162.444 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 9.728121995925903 s) (per-host) I0425 21:19:29.716772 137835727042368 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 156.008 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.14431643486023 s) (per-host) I0425 21:19:30.977285 137835727042368 event_tracking.py:138] [process=6] [sync] Finished load in 12.80 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/9 I0425 21:19:30.977442 137835727042368 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1777151958.1822255, 'checkpointer_duration_secs': 12.795186042785645, 'checkpoint_manager_start_time': 1777151957.9361062, 'checkpoint_manager_duration_secs': 13.0413076877594} fsdp: 32 I0425 21:19:31.395158 137835727042368 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0425 21:19:31.395291 137835727042368 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) I0425 21:19:31.395347 137835727042368 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) I0425 21:19:31.395407 137835727042368 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') I0425 21:19:31.395462 137835727042368 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) I0425 21:19:31.395501 137835727042368 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) I0425 21:19:31.395558 137835727042368 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) I0425 21:19:31.395613 137835727042368 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') I0425 21:19:31.395653 137835727042368 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) I0425 21:19:31.395691 137835727042368 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) I0425 21:19:31.395739 137835727042368 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0425 21:19:31.395789 137835727042368 maxtext_utils.py:1668] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0425 21:19:31.885031 137835727042368 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 21:19:31.885139 137835727042368 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 21:19:31.901138 137835727042368 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 21:19:31.901195 137835727042368 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 21:19:42.987939 137835727042368 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. I0425 21:19:42.988812 137835727042368 metric_logger.py:301] number parameters: 1.104 billion I0425 21:19:54.705059 137835727042368 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0425 21:19:54.858164 137835727042368 checkpointing.py:776] Waited 0.15308141708374023 seconds for step 10 to finish before starting checkpointing. I0425 21:19:54.860506 137835727042368 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 21:19:54.862527 137835727042368 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10 I0425 21:19:54.864163 137835727042368 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/10. I0425 21:19:55.613058 137835727042368 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 21:19:55.614014 137835727042368 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 21:19:55.614070 137835727042368 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 I0425 21:19:55.888633 137835727042368 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275650s I0425 21:19:55.888808 137835727042368 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.519 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2794816493988037 s) (per-host) I0425 21:19:55.888862 137835727042368 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.279547s (batch_requests_ready=0.002133s, total_serialization_initiated=0.277340s, others=0.000074s) I0425 21:19:55.888955 137835727042368 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283825s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.283807) I0425 21:19:55.889709 137835727042368 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/10. I0425 21:19:55.890042 137713789413120 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 21:39:55.890003 I0425 21:19:55.900505 137835727042368 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 21:19:55.900750 137711662503680 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 21:19:55.900874 137835727042368 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777151994.8604808, 'wait_for_prev_duration_secs': 6.842613220214844e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777151994.8625662, 'checkpointer_blocking_duration_secs': 1.0276405811309814, 'get_old_steps_start_time': 1777151995.8902302, 'get_old_steps_duration_secs': 3.361701965332031e-05, 'checkpoint_manager_blocking_start_time': 1777151994.8587327, 'checkpoint_manager_blocking_duration_secs': 1.0421066284179688} I0425 21:19:55.900979 137835727042368 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0425 21:19:55.901027 137835727042368 max_utils.py:750] Memstats: After params initialized: I0425 21:19:55.901075 137835727042368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 21:19:55.901133 137835727042368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 21:19:55.901165 137835727042368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 21:19:55.901190 137835727042368 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 21:19:56.213653 137835727042368 metric_logger.py:196] completed step: 10, seconds: 11.716, TFLOP/s/device: 1.160, Tokens/s/device: 174.802, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0425 21:19:56.812812 137835727042368 metric_logger.py:196] completed step: 11, seconds: 1.507, TFLOP/s/device: 9.016, Tokens/s/device: 1359.044, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0425 21:19:56.960121 137835727042368 metric_logger.py:196] completed step: 12, seconds: 0.453, TFLOP/s/device: 29.987, Tokens/s/device: 4520.013, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0425 21:19:57.107378 137835727042368 metric_logger.py:196] completed step: 13, seconds: 0.152, TFLOP/s/device: 89.170, Tokens/s/device: 13440.701, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0425 21:19:57.112033 137835727042368 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0425 21:19:57.401563 137835727042368 checkpointing.py:776] Waited 0.28949403762817383 seconds for step 15 to finish before starting checkpointing. I0425 21:19:57.404582 137835727042368 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 21:20:02.467397 137712212371200 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_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6 I0425 21:20:07.150226 137713789413120 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 136.870 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.540863275527954 s) (per-host) I0425 21:20:07.150355 137713789413120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.260189s. I0425 21:20:15.012455 137713789413120 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 19.122272s. I0425 21:20:15.012762 137711662503680 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 21:20:15.012898 137711662503680 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 21:20:15.012960 137711662503680 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 21:20:15.015353 137711662503680 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 21:20:15.015544 137835727042368 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. W0425 21:20:15.015676 137835727042368 checkpoint_manager.py:1452] Waiting for previous save to complete took 17.611094 seconds. If this number is high, consider checkpointing less frequently. I0425 21:20:15.017920 137835727042368 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15 I0425 21:20:15.020401 137835727042368 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/15. I0425 21:20:15.600946 137835727042368 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 21:20:15.601061 137835727042368 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 I0425 21:20:15.625815 137835727042368 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026853s I0425 21:20:15.625945 137835727042368 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 50.931 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030287504196166992 s) (per-host) I0425 21:20:15.626010 137835727042368 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030358s (batch_requests_ready=0.001779s, total_serialization_initiated=0.028502s, others=0.000077s) I0425 21:20:15.626131 137835727042368 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035522s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=0.035509) I0425 21:20:15.626844 137835727042368 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/15. I0425 21:20:15.627158 137711662503680 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 21:40:15.627124 I0425 21:20:15.640784 137835727042368 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 21:20:15.641042 137708462667520 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 21:20:15.641168 137835727042368 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777151997.4045517, 'wait_for_prev_duration_secs': 17.6110942363739, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777152015.0179622, 'checkpointer_blocking_duration_secs': 0.6093554496765137, 'get_old_steps_start_time': 1777152015.6273434, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1777151997.40218, 'checkpoint_manager_blocking_duration_secs': 18.238954544067383} I0425 21:20:15.641335 137835727042368 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0425 21:20:15.642138 137835727042368 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.414, Tokens/s/device: 13929.699, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0425 21:20:15.648408 137835727042368 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.160, Tokens/s/device: 13891.338, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0425 21:20:21.484614 137712212371200 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_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6 I0425 21:20:26.467022 137711662503680 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 145.299 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.871325969696045 s) (per-host) I0425 21:20:26.467165 137711662503680 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.839889s. I0425 21:20:32.181052 137711662503680 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.553760s. I0425 21:20:32.181307 137708462667520 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 21:20:32.181366 137708462667520 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 21:20:32.181412 137708462667520 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 21:20:32.184012 137708462667520 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 21:20:37.007050 137835727042368 metric_logger.py:196] completed step: 16, seconds: 18.535, TFLOP/s/device: 0.733, Tokens/s/device: 110.495, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0425 21:20:37.154471 137835727042368 metric_logger.py:196] completed step: 17, seconds: 21.213, TFLOP/s/device: 0.641, Tokens/s/device: 96.546, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0425 21:20:37.302276 137835727042368 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.197, Tokens/s/device: 13444.760, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0425 21:20:37.448762 137835727042368 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0425 21:20:37.449487 137835727042368 checkpointing.py:776] Waited 0.0007455348968505859 seconds for step 19 to finish before starting checkpointing. I0425 21:20:37.451655 137835727042368 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 21:20:37.453215 137835727042368 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19 I0425 21:20:37.454620 137835727042368 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/19. I0425 21:20:37.741785 137835727042368 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 21:20:37.741878 137835727042368 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 I0425 21:20:37.766125 137835727042368 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.025436s I0425 21:20:37.766262 137835727042368 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 53.432 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.02886962890625 s) (per-host) I0425 21:20:37.766310 137835727042368 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.028927s (batch_requests_ready=0.001756s, total_serialization_initiated=0.027110s, others=0.000061s) I0425 21:20:37.766389 137835727042368 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.033038s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.033026) I0425 21:20:37.767116 137835727042368 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.31 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/19. I0425 21:20:37.767465 137708462667520 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 21:40:37.767424 I0425 21:20:37.773140 137835727042368 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 21:20:37.773395 137707909011200 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 21:20:37.773526 137835727042368 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777152037.4516237, 'wait_for_prev_duration_secs': 7.271766662597656e-05, 'time_between_consecutive_saves_sec': 5.2675700187683105, 'checkpointer_blocking_start_time': 1777152037.4532537, 'checkpointer_blocking_duration_secs': 0.31437039375305176, 'get_old_steps_start_time': 1777152037.7676423, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1777152037.449733, 'checkpoint_manager_blocking_duration_secs': 0.3237588405609131} I0425 21:20:37.773636 137835727042368 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0425 21:20:37.773688 137835727042368 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 21:20:42.889471 137712212371200 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_20260425_201227/linen_xpk_test_pipeline_scan_nnx_20260425_201227_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6 I0425 21:20:48.043273 137708462667520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 153.272 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.305842399597168 s) (per-host) I0425 21:20:48.043398 137708462667520 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.275806s. I0425 21:20:53.340322 137708462667520 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 15.572700s. I0425 21:20:53.340608 137707909011200 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 21:20:53.340726 137707909011200 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 21:20:53.340768 137707909011200 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 21:20:53.343632 137707909011200 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 21:20:53.343807 137835727042368 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. I0425 21:20:53.343956 137835727042368 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 21:20:53.345043 137835727042368 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.878, Tokens/s/device: 13848.880, 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: Sat Apr 25 21:21:05 UTC 2026 EXIT_CODE=0