XPK Start: Sat Apr 18 13:41:00 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 13:41:24.294294: 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) I0418 13:41:24.474529 134978752309056 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 13:41:33.516049 134978752309056 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 13:41:33.518676 134978752309056 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-t8if9-slice-job-0-0.mt-11-optimizer-offload-t8if9:8482 I0418 13:41:35.429291 134978752309056 max_utils.py:284] Jax distributed system initialized! I0418 13:41:41.813580 134978752309056 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 13:41:41.813734 134978752309056 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 13:41:41.813775 134978752309056 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 I0418 13:41:41.813809 134978752309056 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 13:41:42.504529 134978752309056 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 13:41:42.504826 134978752309056 checkpointing.py:677] Setting up checkpoint logger... I0418 13:41:42.504883 134978752309056 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 13:41:42.504929 134978752309056 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 13:41:42.505283 134978752309056 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 0x7ac2a450e6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 13:41:46.082215 134978752309056 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 13:41:46.082446 134978752309056 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 0x7aae0c65ac90>}, handler_registry=None I0418 13:41:46.082725 134978752309056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aae0c65ac90>` 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`. I0418 13:41:46.082776 134978752309056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aae0c44c920>` 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`. I0418 13:41:46.082813 134978752309056 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 0x7aae0c65ac90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aae0c65ac90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aae0c44c920>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aae0c44c920>}). I0418 13:41:46.083134 134978752309056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 13:41:46.083203 134978752309056 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7aad2c7dbd80> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 13:41:47.739155 134978752309056 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints I0418 13:41:48.592320 134978752309056 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=10), 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_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7aae0c44e750> I0418 13:41:48.592488 134978752309056 checkpointing.py:301] Checkpoint manager created! I0418 13:41:49.531388 134978752309056 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 13:41:49.531506 134978752309056 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 13:41:49.906674 134978752309056 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 13:41:49.906768 134978752309056 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 13:41:49.923088 134978752309056 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 13:41:49.923158 134978752309056 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 13:41:49.946763 134978752309056 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 13:41:49.946826 134978752309056 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:41:49.963141 134978752309056 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 13:41:49.963207 134978752309056 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:41:49.979437 134978752309056 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 13:41:49.979497 134978752309056 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:41:50.004172 134978752309056 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 13:41:50.004245 134978752309056 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 13:41:50.026385 134978752309056 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 13:41:50.026454 134978752309056 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 13:41:50.259994 134978752309056 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 13:41:50.260100 134978752309056 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 13:41:51.710923 134978752309056 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 13:41:51.711057 134978752309056 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 13:41:51.711110 134978752309056 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 13:41:51.711168 134978752309056 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0418 13:41:51.711241 134978752309056 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 13:41:51.711285 134978752309056 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 13:41:51.711336 134978752309056 maxtext_utils.py:1654] 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) I0418 13:41:51.711389 134978752309056 maxtext_utils.py:1654] 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') I0418 13:41:51.711429 134978752309056 maxtext_utils.py:1654] 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) I0418 13:41:51.711466 134978752309056 maxtext_utils.py:1654] 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) I0418 13:41:51.711514 134978752309056 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 13:41:51.711563 134978752309056 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 13:41:52.210274 134978752309056 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 13:41:52.210366 134978752309056 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 13:41:52.225957 134978752309056 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 13:41:52.226020 134978752309056 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 13:42:03.168220 134978752309056 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. I0418 13:42:03.169068 134978752309056 metric_logger.py:301] number parameters: 1.104 billion I0418 13:42:14.386177 134978752309056 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0418 13:42:14.540803 134978752309056 checkpointing.py:776] Waited 0.1546032428741455 seconds for step 0 to finish before starting checkpointing. I0418 13:42:14.543145 134978752309056 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 13:42:14.544939 134978752309056 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0 I0418 13:42:14.546518 134978752309056 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints/0. I0418 13:42:15.742102 134978752309056 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 13:42:15.742992 134978752309056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 13:42:15.743049 134978752309056 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 I0418 13:42:16.023759 134978752309056 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281730s I0418 13:42:16.023934 134978752309056 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.370 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28725457191467285 s) (per-host) I0418 13:42:16.023984 134978752309056 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.287315s (batch_requests_ready=0.002309s, total_serialization_initiated=0.284938s, others=0.000068s) I0418 13:42:16.024083 134978752309056 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.291605s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.291588) I0418 13:42:16.024912 134978752309056 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.48 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints/0. I0418 13:42:16.025252 134849481676544 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 14:02:16.025216 I0418 13:42:16.027348 134978752309056 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 13:42:16.027655 134849438242560 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 13:42:16.027812 134978752309056 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776519734.5431244, 'wait_for_prev_duration_secs': 7.009506225585938e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776519734.5449789, 'checkpointer_blocking_duration_secs': 1.4804351329803467, 'get_old_steps_start_time': 1776519736.0254433, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776519734.5412924, 'checkpoint_manager_blocking_duration_secs': 1.4864821434020996} I0418 13:42:16.027919 134978752309056 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0418 13:42:16.027971 134978752309056 max_utils.py:750] Memstats: After params initialized: I0418 13:42:16.028018 134978752309056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0418 13:42:16.028049 134978752309056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0418 13:42:16.028076 134978752309056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0418 13:42:16.028102 134978752309056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0418 13:42:16.355730 134978752309056 metric_logger.py:196] completed step: 0, seconds: 11.217, TFLOP/s/device: 1.211, Tokens/s/device: 182.580, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 13:42:16.526836 134978752309056 metric_logger.py:196] completed step: 1, seconds: 1.968, TFLOP/s/device: 6.904, Tokens/s/device: 1040.658, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 13:42:16.965960 134978752309056 metric_logger.py:196] completed step: 2, seconds: 0.024, TFLOP/s/device: 566.343, Tokens/s/device: 85365.345, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0418 13:42:17.113210 134978752309056 metric_logger.py:196] completed step: 3, seconds: 0.439, TFLOP/s/device: 30.935, Tokens/s/device: 4662.907, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0418 13:42:17.408202 134978752309056 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.701, Tokens/s/device: 13369.979, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0418 13:42:17.414505 134978752309056 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.291, Tokens/s/device: 13911.153, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0418 13:42:20.225805 2827 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 13:42:23.632864 134849446635264 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints/0/items/array_metadatas/process_3 I0418 13:42:39.452235 134978752309056 metric_logger.py:196] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.952, Tokens/s/device: 6926.384, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0418 13:42:39.599730 134978752309056 metric_logger.py:196] completed step: 7, seconds: 21.891, TFLOP/s/device: 0.621, Tokens/s/device: 93.555, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0418 13:42:39.747220 134978752309056 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.135, Tokens/s/device: 13435.411, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0418 13:42:39.894137 134978752309056 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0418 13:42:39.894900 134978752309056 checkpointing.py:776] Waited 0.0007805824279785156 seconds for step 9 to finish before starting checkpointing. I0418 13:42:39.896910 134978752309056 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 13:42:53.901007 134849481676544 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 41.389 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.16428899765015 s) (per-host) I0418 13:42:53.901142 134849481676544 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.875768s. I0418 13:43:02.995787 134849481676544 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 46.970397s. I0418 13:43:02.996074 134849438242560 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 13:43:02.996191 134849438242560 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 13:43:02.996249 134849438242560 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 13:43:02.998123 134849438242560 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 13:43:02.998293 134978752309056 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. W0418 13:43:02.998424 134978752309056 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.101516 seconds. If this number is high, consider checkpointing less frequently. I0418 13:43:03.000415 134978752309056 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9 I0418 13:43:03.002530 134978752309056 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints/9. I0418 13:43:03.714483 134978752309056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 13:43:03.714577 134978752309056 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 I0418 13:43:03.750309 134978752309056 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036949s I0418 13:43:03.750476 134978752309056 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 38.049 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0405421257019043 s) (per-host) I0418 13:43:03.750526 134978752309056 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040603s (batch_requests_ready=0.001925s, total_serialization_initiated=0.038611s, others=0.000068s) I0418 13:43:03.750630 134978752309056 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.044938s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.044921) I0418 13:43:03.751362 134978752309056 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints/9. I0418 13:43:03.751676 134849438242560 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 14:03:03.751645 I0418 13:43:03.753715 134978752309056 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 13:43:03.753992 134843002242816 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 13:43:03.754160 134978752309056 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776519759.8968785, 'wait_for_prev_duration_secs': 23.101515531539917, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776519783.0004547, 'checkpointer_blocking_duration_secs': 0.7513701915740967, 'get_old_steps_start_time': 1776519783.7518466, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776519759.895165, 'checkpoint_manager_blocking_duration_secs': 23.85896134376526} I0418 13:43:03.754266 134978752309056 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0418 13:43:03.754312 134978752309056 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 13:43:11.539193 134849446635264 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_11_optimizer_offload_false/checkpoints/9/items/array_metadatas/process_3 I0418 13:43:45.210742 134849438242560 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 38.062 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.500771045684814 s) (per-host) I0418 13:43:45.210867 134849438242560 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.459075s. I0418 13:43:54.835190 134849438242560 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 51.083382s. I0418 13:43:54.835460 134843002242816 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 13:43:54.835579 134843002242816 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 13:43:54.835642 134843002242816 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 13:43:54.837110 134843002242816 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 13:43:54.837325 134978752309056 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. I0418 13:43:54.837471 134978752309056 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 13:43:54.838593 134978752309056 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.132, Tokens/s/device: 13887.100, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 13:44:05 UTC 2026 EXIT_CODE=0