XPK Start: Sat Apr 25 08:39:32 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 08:39:57.228131: 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 08:39:57.432916 136500740904768 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 08:40:06.474650 136500740904768 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 08:40:06.477065 136500740904768 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-rs-4o1xi-slice-job-0-0.mt-16-zarr3-false-rs-4o1xi:8482 I0425 08:40:07.532723 136500740904768 max_utils.py:284] Jax distributed system initialized! I0425 08:40:13.853931 136500740904768 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 08:40:13.854033 136500740904768 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 08:40:13.854073 136500740904768 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 08:40:13.854125 136500740904768 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 08:40:14.553048 136500740904768 maxtext_utils.py:1604] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1) I0425 08:40:14.553375 136500740904768 checkpointing.py:677] Setting up checkpoint logger... I0425 08:40:14.553445 136500740904768 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0425 08:40:14.553511 136500740904768 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 08:40:14.553885 136500740904768 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7c24f731c260>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 08:40:15.849464 136500740904768 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0425 08:40:15.849714 136500740904768 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 0x7c10d81316a0>}, handler_registry=None I0425 08:40:15.849973 136500740904768 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c10d81316a0>` 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 08:40:15.850034 136500740904768 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c10d8096f00>` 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 08:40:15.850087 136500740904768 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 0x7c10d81316a0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c10d81316a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c10d8096f00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c10d8096f00>}). I0425 08:40:15.850483 136500740904768 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 08:40:15.850718 136500740904768 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c0fcc3d1a80> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 08:40:17.296713 136380143351552 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777106198148456128, 'commit_timestamp_nsecs': 1777106245772085107, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA I0425 08:40:17.544564 136379572942592 checkpoint_manager.py:1871] Missing metrics for step 9 I0425 08:40:17.566035 136500740904768 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints I0425 08:40:18.718584 136500740904768 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_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c10b8707cb0> I0425 08:40:18.718759 136500740904768 checkpointing.py:301] Checkpoint manager created! I0425 08:40:19.649883 136500740904768 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 08:40:19.650003 136500740904768 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 08:40:20.027441 136500740904768 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 08:40:20.027538 136500740904768 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 08:40:20.044080 136500740904768 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 08:40:20.044149 136500740904768 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 08:40:20.068371 136500740904768 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 08:40:20.068439 136500740904768 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:40:20.085029 136500740904768 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 08:40:20.085113 136500740904768 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:40:20.101742 136500740904768 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 08:40:20.101803 136500740904768 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:40:20.126647 136500740904768 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0425 08:40:20.126717 136500740904768 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 08:40:20.148398 136500740904768 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 08:40:20.148461 136500740904768 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 08:40:20.358808 136500740904768 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0425 08:40:20.358916 136500740904768 checkpointing.py:581] restoring from this run's directory step 9 I0425 08:40:20.605419 136500740904768 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/9. I0425 08:40:20.856663 136500740904768 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/9. I0425 08:40:20.857085 136500740904768 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/9. I0425 08:40:22.297649 1447 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 08:40:28.787174 136500740904768 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 243.297 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.495235919952393 s) (per-host) I0425 08:40:28.793027 136500740904768 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 227.874 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.945024251937866 s) (per-host) I0425 08:40:30.241474 136500740904768 event_tracking.py:138] [process=6] [sync] Finished load in 9.64 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/9 I0425 08:40:30.241650 136500740904768 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1777106420.60532, 'checkpointer_duration_secs': 9.636296510696411, 'checkpoint_manager_start_time': 1777106420.3592286, 'checkpoint_manager_duration_secs': 9.882389307022095} fsdp: 32 I0425 08:40:30.666051 136500740904768 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0425 08:40:30.666197 136500740904768 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 08:40:30.666252 136500740904768 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 08:40:30.666315 136500740904768 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 08:40:30.666369 136500740904768 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 08:40:30.666409 136500740904768 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 08:40:30.666460 136500740904768 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 08:40:30.666512 136500740904768 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 08:40:30.666552 136500740904768 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 08:40:30.666588 136500740904768 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 08:40:30.666635 136500740904768 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0425 08:40:30.666680 136500740904768 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0425 08:40:31.157686 136500740904768 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 08:40:31.157781 136500740904768 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 08:40:31.173658 136500740904768 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 08:40:31.173720 136500740904768 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 08:40:42.373304 136500740904768 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 08:40:42.374255 136500740904768 metric_logger.py:301] number parameters: 1.104 billion I0425 08:40:54.223748 136500740904768 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0425 08:40:54.377447 136500740904768 checkpointing.py:776] Waited 0.15367937088012695 seconds for step 10 to finish before starting checkpointing. I0425 08:40:54.379779 136500740904768 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 08:40:54.381886 136500740904768 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10 I0425 08:40:54.383657 136500740904768 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/10. I0425 08:40:54.706662 136500740904768 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 08:40:54.707728 136500740904768 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 08:40:54.707783 136500740904768 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 08:40:54.993144 136500740904768 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.286556s I0425 08:40:54.993315 136500740904768 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.310 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2904849052429199 s) (per-host) I0425 08:40:54.993366 136500740904768 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.290547s (batch_requests_ready=0.002178s, total_serialization_initiated=0.288300s, others=0.000069s) I0425 08:40:54.993456 136500740904768 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.294700s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.294683) I0425 08:40:54.994139 136500740904768 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/10. I0425 08:40:54.994512 136379539371776 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 09:00:54.994469 I0425 08:40:54.996506 136500740904768 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 08:40:54.996823 136378444658432 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 08:40:54.996950 136500740904768 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777106454.37976, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777106454.381972, 'checkpointer_blocking_duration_secs': 0.6126971244812012, 'get_old_steps_start_time': 1777106454.9946916, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1777106454.3779738, 'checkpoint_manager_blocking_duration_secs': 0.6189408302307129} I0425 08:40:54.997128 136500740904768 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0425 08:40:54.997180 136500740904768 max_utils.py:750] Memstats: After params initialized: I0425 08:40:54.997234 136500740904768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 08:40:54.997268 136500740904768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 08:40:54.997297 136500740904768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 08:40:54.997323 136500740904768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 08:40:55.311832 136500740904768 metric_logger.py:196] completed step: 10, seconds: 11.849, TFLOP/s/device: 1.147, Tokens/s/device: 172.836, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0425 08:40:55.488165 136500740904768 metric_logger.py:196] completed step: 11, seconds: 1.087, TFLOP/s/device: 12.505, Tokens/s/device: 1884.893, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0425 08:40:55.635279 136500740904768 metric_logger.py:196] completed step: 12, seconds: 0.029, TFLOP/s/device: 468.570, Tokens/s/device: 70627.996, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0425 08:40:55.782692 136500740904768 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.568, Tokens/s/device: 13349.934, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0425 08:40:55.787162 136500740904768 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0425 08:40:56.076815 136500740904768 checkpointing.py:776] Waited 0.28961706161499023 seconds for step 15 to finish before starting checkpointing. I0425 08:40:56.079856 136500740904768 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 08:40:59.850042 136378453051136 array_metadata_store.py:203] [process=6][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_false/checkpoints/10/items/array_metadatas/process_6 I0425 08:41:35.631880 136379539371776 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.594 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.92902708053589 s) (per-host) I0425 08:41:35.631993 136379539371776 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.637369s. I0425 08:41:44.737168 136379539371776 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.742526s. I0425 08:41:44.737450 136378444658432 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 08:41:44.737563 136378444658432 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 08:41:44.737608 136378444658432 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 08:41:44.741840 136378444658432 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 08:41:44.742035 136500740904768 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 08:41:44.742179 136500740904768 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.662325 seconds. If this number is high, consider checkpointing less frequently. I0425 08:41:44.744259 136500740904768 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15 I0425 08:41:44.746792 136500740904768 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/15. I0425 08:41:45.050480 136500740904768 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 08:41:45.050574 136500740904768 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 08:41:45.076666 136500740904768 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027358s I0425 08:41:45.076841 136500740904768 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 49.962 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03087472915649414 s) (per-host) I0425 08:41:45.076892 136500740904768 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030935s (batch_requests_ready=0.001785s, total_serialization_initiated=0.029083s, others=0.000067s) I0425 08:41:45.076989 136500740904768 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036565s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.036553) I0425 08:41:45.077614 136500740904768 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/15. I0425 08:41:45.077906 136378453051136 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 09:01:45.077870 I0425 08:41:45.089546 136500740904768 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 08:41:45.089816 136378444658432 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 08:41:45.089972 136500740904768 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777106456.079826, 'wait_for_prev_duration_secs': 48.66232466697693, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777106504.7442992, 'checkpointer_blocking_duration_secs': 0.33375000953674316, 'get_old_steps_start_time': 1777106505.0780683, 'get_old_steps_duration_secs': 4.1961669921875e-05, 'checkpoint_manager_blocking_start_time': 1777106456.0774124, 'checkpoint_manager_blocking_duration_secs': 49.01252722740173} I0425 08:41:45.090079 136500740904768 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0425 08:41:45.090919 136500740904768 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.326, Tokens/s/device: 13916.352, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0425 08:41:45.096901 136500740904768 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.313, Tokens/s/device: 13914.367, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0425 08:41:50.823610 136374204233472 array_metadata_store.py:203] [process=6][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_false/checkpoints/15/items/array_metadatas/process_6 I0425 08:42:06.099863 136500740904768 metric_logger.py:196] completed step: 16, seconds: 49.308, TFLOP/s/device: 0.276, Tokens/s/device: 41.535, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0425 08:42:06.247298 136500740904768 metric_logger.py:196] completed step: 17, seconds: 20.857, TFLOP/s/device: 0.651, Tokens/s/device: 98.193, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0425 08:42:06.394693 136500740904768 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.260, Tokens/s/device: 13454.299, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0425 08:42:06.541248 136500740904768 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0425 08:42:06.541951 136500740904768 checkpointing.py:776] Waited 0.00072479248046875 seconds for step 19 to finish before starting checkpointing. I0425 08:42:06.544060 136500740904768 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 08:42:26.460512 136378453051136 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.141 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.41452121734619 s) (per-host) I0425 08:42:26.460626 136378453051136 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.382608s. I0425 08:42:33.582317 136378453051136 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.504283s. I0425 08:42:33.582617 136378444658432 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 08:42:33.582759 136378444658432 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 08:42:33.582818 136378444658432 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 08:42:33.584479 136378444658432 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 08:42:33.584664 136500740904768 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0425 08:42:33.584787 136500740904768 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.040733 seconds. If this number is high, consider checkpointing less frequently. I0425 08:42:33.586352 136500740904768 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19 I0425 08:42:33.588632 136500740904768 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/19. I0425 08:42:34.320141 136500740904768 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 08:42:34.320232 136500740904768 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 08:42:34.344928 136500740904768 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.025719s I0425 08:42:34.345073 136500740904768 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 52.780 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.02922654151916504 s) (per-host) I0425 08:42:34.345141 136500740904768 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.029307s (batch_requests_ready=0.001670s, total_serialization_initiated=0.027550s, others=0.000087s) I0425 08:42:34.345215 136500740904768 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.033472s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.033461) I0425 08:42:34.345925 136500740904768 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints/19. I0425 08:42:34.346275 136378444658432 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 09:02:34.346237 I0425 08:42:34.350330 136500740904768 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 08:42:34.350565 136374204233472 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 08:42:34.350692 136500740904768 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260425_071456/linen_xpk_main_20260425_071456_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777106526.5440307, 'wait_for_prev_duration_secs': 27.040733337402344, 'time_between_consecutive_saves_sec': 21.802149772644043, 'checkpointer_blocking_start_time': 1777106553.5863914, 'checkpointer_blocking_duration_secs': 0.7600321769714355, 'get_old_steps_start_time': 1777106554.3464417, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1777106526.5422041, 'checkpoint_manager_blocking_duration_secs': 27.808454275131226} I0425 08:42:34.350809 136500740904768 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0425 08:42:34.350850 136500740904768 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 08:42:39.444448 136373667362560 array_metadata_store.py:203] [process=6][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_false/checkpoints/19/items/array_metadatas/process_6 I0425 08:43:14.846362 136378444658432 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.973 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.530478715896606 s) (per-host) I0425 08:43:14.846483 136378444658432 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.500093s. I0425 08:43:23.000396 136378444658432 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.653989s. I0425 08:43:23.000690 136374204233472 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 08:43:23.000820 136374204233472 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 08:43:23.000861 136374204233472 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 08:43:23.002639 136374204233472 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 08:43:23.002805 136500740904768 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 08:43:23.002954 136500740904768 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 08:43:23.003871 136500740904768 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.201, Tokens/s/device: 13897.560, 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 08:43:32 UTC 2026 EXIT_CODE=0