XPK Start: Fri Apr 24 07:59:29 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. 2026-04-24 07:59:54.659722: 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) I0424 07:59:54.863277 132215042836288 max_utils.py:273] Attempting to initialize the jax distributed system... I0424 08:00:03.903813 132215042836288 distributed.py:149] Starting JAX distributed service on [::]:8482 I0424 08:00:03.906272 132215042836288 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-wn2t5-slice-job-0-0.mt-15-ocdbt-true-sv-wn2t5:8482 I0424 08:00:05.009634 132215042836288 max_utils.py:284] Jax distributed system initialized! I0424 08:00:11.157711 132215042836288 max_utils.py:800] System Information: Jax Version: 0.9.2 I0424 08:00:11.157817 132215042836288 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0424 08:00:11.157857 132215042836288 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 I0424 08:00:11.157893 132215042836288 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 08:00:11.850472 132215042836288 maxtext_utils.py:1604] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1) I0424 08:00:11.850765 132215042836288 checkpointing.py:677] Setting up checkpoint logger... I0424 08:00:11.850819 132215042836288 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 08:00:11.850863 132215042836288 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0424 08:00:11.851221 132215042836288 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 0x783eff1342c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 08:00:14.789743 132215042836288 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0424 08:00:14.789977 132215042836288 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 0x782b98179160>}, handler_registry=None I0424 08:00:14.790240 132215042836288 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x782b98179160>` 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`. I0424 08:00:14.790289 132215042836288 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x782b807244a0>` 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`. I0424 08:00:14.790326 132215042836288 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 0x782b98179160>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x782b98179160>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x782b807244a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x782b807244a0>}). I0424 08:00:14.790658 132215042836288 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0424 08:00:14.790728 132215042836288 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x782b80441a80> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0424 08:00:15.880880 132215042836288 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints I0424 08:00:16.209074 132215042836288 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_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x782b80721f40> I0424 08:00:16.209291 132215042836288 checkpointing.py:301] Checkpoint manager created! I0424 08:00:17.161121 132215042836288 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 08:00:17.161246 132215042836288 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 08:00:17.542229 132215042836288 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 08:00:17.542320 132215042836288 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 08:00:17.558773 132215042836288 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 08:00:17.558835 132215042836288 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 08:00:17.582590 132215042836288 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 08:00:17.582657 132215042836288 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 08:00:17.599307 132215042836288 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 08:00:17.599370 132215042836288 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 08:00:17.615913 132215042836288 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 08:00:17.615974 132215042836288 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 08:00:17.640824 132215042836288 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0424 08:00:17.640898 132215042836288 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 08:00:17.662301 132215042836288 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 08:00:17.662365 132215042836288 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 08:00:17.873600 132215042836288 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0424 08:00:17.873713 132215042836288 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 08:00:19.302199 132215042836288 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0424 08:00:19.302331 132215042836288 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) I0424 08:00:19.302385 132215042836288 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) I0424 08:00:19.302444 132215042836288 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') I0424 08:00:19.302501 132215042836288 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) I0424 08:00:19.302541 132215042836288 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) I0424 08:00:19.302600 132215042836288 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) I0424 08:00:19.302652 132215042836288 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') I0424 08:00:19.302693 132215042836288 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) I0424 08:00:19.302730 132215042836288 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) I0424 08:00:19.302779 132215042836288 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0424 08:00:19.302828 132215042836288 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0424 08:00:19.802068 132215042836288 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 08:00:19.802185 132215042836288 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 08:00:19.817905 132215042836288 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 08:00:19.817967 132215042836288 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 08:00:30.907982 132215042836288 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. I0424 08:00:30.908816 132215042836288 metric_logger.py:301] number parameters: 1.104 billion I0424 08:00:42.737912 132215042836288 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0424 08:00:42.892649 132215042836288 checkpointing.py:776] Waited 0.15472006797790527 seconds for step 0 to finish before starting checkpointing. I0424 08:00:42.895029 132215042836288 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 08:00:42.896599 132215042836288 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0424 08:00:42.897992 132215042836288 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/0. I0424 08:00:43.248967 132215042836288 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 08:00:43.249800 132215042836288 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 08:00:43.249857 132215042836288 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 I0424 08:00:43.525616 132215042836288 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276721s I0424 08:00:43.525791 132215042836288 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.462 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2824099063873291 s) (per-host) I0424 08:00:43.525841 132215042836288 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282470s (batch_requests_ready=0.002374s, total_serialization_initiated=0.280027s, others=0.000069s) I0424 08:00:43.525934 132215042836288 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286855s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.286837) I0424 08:00:43.526690 132215042836288 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.63 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/0. I0424 08:00:43.527061 132089901135616 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 08:20:43.527023 I0424 08:00:43.530411 132215042836288 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 08:00:43.530631 132089319667456 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 08:00:43.530745 132215042836288 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777017642.8950098, 'wait_for_prev_duration_secs': 6.29425048828125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777017642.8966374, 'checkpointer_blocking_duration_secs': 0.630537748336792, 'get_old_steps_start_time': 1777017643.5272, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1777017642.8931932, 'checkpoint_manager_blocking_duration_secs': 0.6375143527984619} I0424 08:00:43.530849 132215042836288 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0424 08:00:43.530899 132215042836288 max_utils.py:750] Memstats: After params initialized: I0424 08:00:43.530947 132215042836288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 08:00:43.530982 132215042836288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 08:00:43.531010 132215042836288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 08:00:43.531036 132215042836288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 08:00:43.848896 132215042836288 metric_logger.py:196] completed step: 0, seconds: 11.829, TFLOP/s/device: 1.149, Tokens/s/device: 173.134, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0424 08:00:44.106809 132215042836288 metric_logger.py:196] completed step: 1, seconds: 1.109, TFLOP/s/device: 12.247, Tokens/s/device: 1846.061, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0424 08:00:44.520590 132215042836288 metric_logger.py:196] completed step: 2, seconds: 0.110, TFLOP/s/device: 123.112, Tokens/s/device: 18556.776, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0424 08:00:44.667885 132215042836288 metric_logger.py:196] completed step: 3, seconds: 0.414, TFLOP/s/device: 32.833, Tokens/s/device: 4948.988, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0424 08:00:44.672506 132215042836288 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0424 08:00:44.961817 132215042836288 checkpointing.py:776] Waited 0.2892723083496094 seconds for step 5 to finish before starting checkpointing. I0424 08:00:44.964711 132215042836288 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 08:00:46.883304 2747 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 08:00:48.720487 132089328060160 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0424 08:01:20.733436 132089901135616 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 42.134 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.49003291130066 s) (per-host) I0424 08:01:20.733545 132089901135616 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.205847s. I0424 08:01:28.576564 132089901135616 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 45.048848s. I0424 08:01:28.576853 132089319667456 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 08:01:28.576978 132089319667456 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 08:01:28.577029 132089319667456 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 08:01:28.579155 132089319667456 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 08:01:28.579320 132215042836288 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0424 08:01:28.579449 132215042836288 checkpoint_manager.py:1452] Waiting for previous save to complete took 43.614737 seconds. If this number is high, consider checkpointing less frequently. I0424 08:01:28.581383 132215042836288 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0424 08:01:28.584023 132215042836288 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/5. I0424 08:01:28.893621 132215042836288 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 08:01:28.893715 132215042836288 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 I0424 08:01:28.934436 132215042836288 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041900s I0424 08:01:28.934611 132215042836288 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.889 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04551863670349121 s) (per-host) I0424 08:01:28.934658 132215042836288 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045576s (batch_requests_ready=0.001925s, total_serialization_initiated=0.043586s, others=0.000065s) I0424 08:01:28.934747 132215042836288 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050400s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.050385) I0424 08:01:28.935459 132215042836288 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/5. I0424 08:01:28.935817 132089319667456 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 08:21:28.935779 I0424 08:01:28.945374 132215042836288 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 08:01:28.945636 132066947622656 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 08:01:28.945796 132215042836288 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777017644.9646816, 'wait_for_prev_duration_secs': 43.61473727226257, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777017688.5814238, 'checkpointer_blocking_duration_secs': 0.3545420169830322, 'get_old_steps_start_time': 1777017688.9359915, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1777017644.9625132, 'checkpoint_manager_blocking_duration_secs': 43.98324775695801} I0424 08:01:28.945909 132215042836288 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0424 08:01:28.946711 132215042836288 metric_logger.py:196] completed step: 4, seconds: 0.154, TFLOP/s/device: 88.412, Tokens/s/device: 13326.393, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0424 08:01:28.952931 132215042836288 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.289, Tokens/s/device: 13910.775, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0424 08:01:33.995157 132089328060160 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0424 08:01:51.282595 132215042836288 metric_logger.py:196] completed step: 6, seconds: 44.279, TFLOP/s/device: 0.307, Tokens/s/device: 46.252, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0424 08:01:51.429975 132215042836288 metric_logger.py:196] completed step: 7, seconds: 22.184, TFLOP/s/device: 0.612, Tokens/s/device: 92.319, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0424 08:01:51.577286 132215042836288 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.438, Tokens/s/device: 13481.046, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0424 08:01:51.723929 132215042836288 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0424 08:01:51.724623 132215042836288 checkpointing.py:776] Waited 0.0007135868072509766 seconds for step 9 to finish before starting checkpointing. I0424 08:01:51.726821 132215042836288 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 08:02:11.396238 132089319667456 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.161 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.5071063041687 s) (per-host) I0424 08:02:11.396366 132089319667456 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.460435s. I0424 08:02:18.952470 132089319667456 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.016524s. I0424 08:02:18.952711 132066947622656 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 08:02:18.952776 132066947622656 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 08:02:18.952824 132066947622656 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 08:02:18.954408 132066947622656 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 08:02:18.954605 132215042836288 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0424 08:02:18.954734 132215042836288 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.227918 seconds. If this number is high, consider checkpointing less frequently. I0424 08:02:18.956243 132215042836288 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0424 08:02:18.958273 132215042836288 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/9. I0424 08:02:19.663468 132215042836288 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 08:02:19.663562 132215042836288 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 I0424 08:02:19.703174 132215042836288 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040472s I0424 08:02:19.703343 132215042836288 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 34.963 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04412031173706055 s) (per-host) I0424 08:02:19.703392 132215042836288 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.044179s (batch_requests_ready=0.001808s, total_serialization_initiated=0.042307s, others=0.000064s) I0424 08:02:19.703483 132215042836288 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048553s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.048543) I0424 08:02:19.704213 132215042836288 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/9. I0424 08:02:19.704563 132066947622656 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 08:22:19.704519 I0424 08:02:19.707733 132215042836288 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 08:02:19.707971 132087222892288 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 08:02:19.708139 132215042836288 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777017711.72679, 'wait_for_prev_duration_secs': 27.227917909622192, 'time_between_consecutive_saves_sec': 23.147594213485718, 'checkpointer_blocking_start_time': 1777017738.9562826, 'checkpointer_blocking_duration_secs': 0.7484302520751953, 'get_old_steps_start_time': 1777017739.7047298, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1777017711.7248495, 'checkpoint_manager_blocking_duration_secs': 27.98325777053833} I0424 08:02:19.708251 132215042836288 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0424 08:02:19.708293 132215042836288 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 08:02:25.119073 132089328060160 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227/linen_xpk_main_20260424_070227_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0424 08:03:01.735125 132066947622656 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.542 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.07584810256958 s) (per-host) I0424 08:03:01.735246 132066947622656 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.030568s. I0424 08:03:09.886920 132066947622656 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.182228s. I0424 08:03:09.887210 132087222892288 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 08:03:09.887331 132087222892288 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 08:03:09.887372 132087222892288 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 08:03:09.889107 132087222892288 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 08:03:09.889317 132215042836288 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0424 08:03:09.889472 132215042836288 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 08:03:09.890396 132215042836288 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.196, Tokens/s/device: 13896.711, 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: Fri Apr 24 08:03:19 UTC 2026 EXIT_CODE=0