XPK Start: Sat Apr 25 08:24:38 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 08:25:03.806734: 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 08:25:04.012476 134365769791296 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 08:25:13.052645 134365769791296 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 08:25:13.055004 134365769791296 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-sv-481ic-slice-job-0-0.mt-16-zarr3-true-sv-481ic:8482 I0425 08:25:14.043148 134365769791296 max_utils.py:284] Jax distributed system initialized! I0425 08:25:20.177426 134365769791296 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 08:25:20.177535 134365769791296 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 08:25:20.177578 134365769791296 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 08:25:20.177615 134365769791296 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 08:25:20.868419 134365769791296 maxtext_utils.py:1604] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1) I0425 08:25:20.868709 134365769791296 checkpointing.py:677] Setting up checkpoint logger... I0425 08:25:20.868764 134365769791296 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 08:25:20.868808 134365769791296 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 08:25:20.869171 134365769791296 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 0x7a33e1992660>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 08:25:23.786613 134365769791296 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0425 08:25:23.786852 134365769791296 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1f7c27c170>}, handler_registry=None I0425 08:25:23.787102 134365769791296 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1f7c27c170>` 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 08:25:23.787152 134365769791296 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1f7c27fd70>` 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 08:25:23.787189 134365769791296 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 0x7a1f7c27c170>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1f7c27c170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1f7c27fd70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1f7c27fd70>}). I0425 08:25:23.787505 134365769791296 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 08:25:23.787575 134365769791296 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a1e8c579a80> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 08:25:24.893040 134365769791296 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints I0425 08:25:25.271875 134365769791296 checkpoint_manager.py:929] [process=3][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a34281d3140> I0425 08:25:25.272085 134365769791296 checkpointing.py:301] Checkpoint manager created! I0425 08:25:26.203698 134365769791296 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 08:25:26.203816 134365769791296 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 08:25:26.584432 134365769791296 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 08:25:26.584524 134365769791296 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 08:25:26.601078 134365769791296 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 08:25:26.601138 134365769791296 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 08:25:26.625315 134365769791296 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 08:25:26.625386 134365769791296 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:25:26.642076 134365769791296 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 08:25:26.642136 134365769791296 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:25:26.658688 134365769791296 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 08:25:26.658746 134365769791296 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:25:26.683625 134365769791296 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0425 08:25:26.683697 134365769791296 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:25:26.704998 134365769791296 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 08:25:26.705068 134365769791296 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 08:25:26.916920 134365769791296 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0425 08:25:26.917038 134365769791296 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 08:25:28.349372 134365769791296 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0425 08:25:28.349503 134365769791296 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0425 08:25:28.349557 134365769791296 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0425 08:25:28.349615 134365769791296 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0425 08:25:28.349672 134365769791296 maxtext_utils.py:1707] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0425 08:25:28.349713 134365769791296 maxtext_utils.py:1707] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0425 08:25:28.349774 134365769791296 maxtext_utils.py:1707] 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 08:25:28.349827 134365769791296 maxtext_utils.py:1707] 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 08:25:28.349867 134365769791296 maxtext_utils.py:1707] 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 08:25:28.349904 134365769791296 maxtext_utils.py:1707] 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 08:25:28.349967 134365769791296 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0425 08:25:28.350020 134365769791296 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0425 08:25:28.844969 134365769791296 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 08:25:28.845066 134365769791296 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 08:25:28.860685 134365769791296 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 08:25:28.860754 134365769791296 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 08:25:39.863549 134365769791296 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 08:25:39.864334 134365769791296 metric_logger.py:301] number parameters: 1.104 billion I0425 08:25:51.792248 134365769791296 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0425 08:25:51.948171 134365769791296 checkpointing.py:776] Waited 0.15590119361877441 seconds for step 0 to finish before starting checkpointing. I0425 08:25:51.950522 134365769791296 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 08:25:51.952374 134365769791296 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0 I0425 08:25:51.953681 134365769791296 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/0. I0425 08:25:52.677213 134365769791296 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 08:25:52.678144 134365769791296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 08:25:52.678202 134365769791296 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 08:25:52.967106 134365769791296 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.289970s I0425 08:25:52.967283 134365769791296 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.220 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29550695419311523 s) (per-host) I0425 08:25:52.967338 134365769791296 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.295572s (batch_requests_ready=0.002261s, total_serialization_initiated=0.293238s, others=0.000073s) I0425 08:25:52.967428 134365769791296 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.299549s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.299532) I0425 08:25:52.968270 134365769791296 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/0. I0425 08:25:52.968572 134237189568256 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 08:45:52.968537 I0425 08:25:52.970577 134365769791296 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 08:25:52.970888 134236603037440 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 08:25:52.971060 134365769791296 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777105551.9505026, 'wait_for_prev_duration_secs': 6.103515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777105551.952413, 'checkpointer_blocking_duration_secs': 1.0163085460662842, 'get_old_steps_start_time': 1777105552.9687488, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1777105551.9487145, 'checkpoint_manager_blocking_duration_secs': 1.022305965423584} I0425 08:25:52.971174 134365769791296 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0425 08:25:52.971224 134365769791296 max_utils.py:750] Memstats: After params initialized: I0425 08:25:52.971274 134365769791296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0425 08:25:52.971308 134365769791296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0425 08:25:52.971336 134365769791296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0425 08:25:52.971362 134365769791296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0425 08:25:53.284318 134365769791296 metric_logger.py:196] completed step: 0, seconds: 11.928, TFLOP/s/device: 1.139, Tokens/s/device: 171.700, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0425 08:25:53.461005 134365769791296 metric_logger.py:196] completed step: 1, seconds: 1.491, TFLOP/s/device: 9.116, Tokens/s/device: 1374.008, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0425 08:25:53.876269 134365769791296 metric_logger.py:196] completed step: 2, seconds: 0.029, TFLOP/s/device: 463.867, Tokens/s/device: 69919.088, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0425 08:25:54.023580 134365769791296 metric_logger.py:196] completed step: 3, seconds: 0.416, TFLOP/s/device: 32.699, Tokens/s/device: 4928.823, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0425 08:25:54.028037 134365769791296 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0425 08:25:54.317687 134365769791296 checkpointing.py:776] Waited 0.28960609436035156 seconds for step 5 to finish before starting checkpointing. I0425 08:25:54.320622 134365769791296 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 08:25:55.860823 2774 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 08:25:57.899538 134236611430144 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/0/items/array_metadatas/process_3 I0425 08:26:28.941267 134237189568256 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 43.552 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.26943516731262 s) (per-host) I0425 08:26:28.941396 134237189568256 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.972710s. I0425 08:26:36.173874 134237189568256 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 43.205172s. I0425 08:26:36.174175 134236603037440 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 08:26:36.174294 134236603037440 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 08:26:36.174346 134236603037440 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 08:26:36.176784 134236603037440 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 08:26:36.176975 134365769791296 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0425 08:26:36.177108 134365769791296 checkpoint_manager.py:1452] Waiting for previous save to complete took 41.856486 seconds. If this number is high, consider checkpointing less frequently. I0425 08:26:36.179121 134365769791296 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 5 I0425 08:26:36.181574 134365769791296 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/5. I0425 08:26:36.753439 134365769791296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 08:26:36.753537 134365769791296 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 08:26:36.798955 134365769791296 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046676s I0425 08:26:36.799120 134365769791296 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 30.767 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05013704299926758 s) (per-host) I0425 08:26:36.799170 134365769791296 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.050197s (batch_requests_ready=0.001774s, total_serialization_initiated=0.048356s, others=0.000067s) I0425 08:26:36.799263 134365769791296 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.055302s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.055287) I0425 08:26:36.799979 134365769791296 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/5. I0425 08:26:36.800270 134236603037440 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 08:46:36.800238 I0425 08:26:36.802141 134365769791296 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 08:26:36.802453 134214867461888 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 08:26:36.802595 134365769791296 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777105554.3205905, 'wait_for_prev_duration_secs': 41.856486320495605, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777105596.1791613, 'checkpointer_blocking_duration_secs': 0.621248722076416, 'get_old_steps_start_time': 1777105596.800439, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1777105554.3183074, 'checkpoint_manager_blocking_duration_secs': 42.484251976013184} I0425 08:26:36.802759 134365769791296 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0425 08:26:36.803539 134365769791296 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.689, Tokens/s/device: 13368.233, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0425 08:26:36.809605 134365769791296 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.343, Tokens/s/device: 13918.906, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0425 08:26:41.689067 134236611430144 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/5/items/array_metadatas/process_3 I0425 08:26:59.030376 134365769791296 metric_logger.py:196] completed step: 6, seconds: 42.780, TFLOP/s/device: 0.318, Tokens/s/device: 47.873, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0425 08:26:59.177941 134365769791296 metric_logger.py:196] completed step: 7, seconds: 22.074, TFLOP/s/device: 0.616, Tokens/s/device: 92.777, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0425 08:26:59.324947 134365769791296 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 89.058, Tokens/s/device: 13423.787, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0425 08:26:59.471539 134365769791296 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0425 08:26:59.472239 134365769791296 checkpointing.py:776] Waited 0.0007205009460449219 seconds for step 9 to finish before starting checkpointing. I0425 08:26:59.474258 134365769791296 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 08:27:18.580053 134236603037440 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.761 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.83104705810547 s) (per-host) I0425 08:27:18.580161 134236603037440 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.779779s. I0425 08:27:25.567879 134236603037440 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.767481s. I0425 08:27:25.568192 134214867461888 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 08:27:25.568339 134214867461888 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 08:27:25.568401 134214867461888 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 08:27:25.569897 134214867461888 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 08:27:25.570113 134365769791296 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0425 08:27:25.570261 134365769791296 checkpoint_manager.py:1452] Waiting for previous save to complete took 26.096011 seconds. If this number is high, consider checkpointing less frequently. I0425 08:27:25.571714 134365769791296 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9 I0425 08:27:25.573843 134365769791296 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/9. I0425 08:27:25.865714 134365769791296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 08:27:25.865814 134365769791296 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 08:27:25.913319 134365769791296 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048452s I0425 08:27:25.913486 134365769791296 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 29.694 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.051949501037597656 s) (per-host) I0425 08:27:25.913533 134365769791296 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.052008s (batch_requests_ready=0.001666s, total_serialization_initiated=0.050279s, others=0.000063s) I0425 08:27:25.913625 134365769791296 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.056013s (all_items=0.000010s, per_item={'items': '0.00000954'}, temp_paths=0.056003) I0425 08:27:25.914261 134365769791296 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/9. I0425 08:27:25.914664 134233940109056 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 08:47:25.914626 I0425 08:27:25.916488 134365769791296 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 08:27:25.916771 134214867461888 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 08:27:25.916904 134365769791296 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777105619.474227, 'wait_for_prev_duration_secs': 26.09601140022278, 'time_between_consecutive_saves_sec': 23.29740285873413, 'checkpointer_blocking_start_time': 1777105645.571754, 'checkpointer_blocking_duration_secs': 0.34305739402770996, 'get_old_steps_start_time': 1777105645.9148283, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777105619.472474, 'checkpoint_manager_blocking_duration_secs': 26.44440007209778} I0425 08:27:25.917085 134365769791296 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0425 08:27:25.917130 134365769791296 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 08:27:31.008166 134236611430144 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_true/checkpoints/9/items/array_metadatas/process_3 I0425 08:28:08.125092 134233940109056 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.375 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.263516426086426 s) (per-host) I0425 08:28:08.125213 134233940109056 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.210431s. I0425 08:28:13.987033 134233940109056 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.072235s. I0425 08:28:13.987317 134214867461888 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 08:28:13.987449 134214867461888 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 08:28:13.987492 134214867461888 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 08:28:13.992406 134214867461888 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 08:28:13.992581 134365769791296 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0425 08:28:13.992747 134365769791296 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 08:28:13.993705 134365769791296 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.226, Tokens/s/device: 13901.239, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 08:28:24 UTC 2026 EXIT_CODE=0