XPK Start: Fri Apr 24 16:21:14 UTC 2026 2026-04-24 16:21:18.598353: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1777047678.611738 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777047678.615483 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777047678.627196 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047678.627214 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047678.627217 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047678.627219 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 16:21:37.888313: 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 16:21:38.419917 140248041047872 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 16:21:47,461:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 16:21:47.461369 140248041047872 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 16:21:47,463:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-pphz4-slice-job-0-0.mt-16-zarr3-false-sv-pphz4:8482 I0424 16:21:47.463794 140248041047872 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-pphz4-slice-job-0-0.mt-16-zarr3-false-sv-pphz4:8482 I0424 16:21:48.229756 140248041047872 max_utils.py:284] Jax distributed system initialized! I0424 16:21:55.184713 140248041047872 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 16:21:55.184822 140248041047872 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 16:21:55.184862 140248041047872 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0424 16:21:55.184896 140248041047872 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 16:21:55.813476 140248041047872 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 16:21:55.814079 140248041047872 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 16:21:55.814283 140248041047872 checkpointing.py:688] Setting up checkpoint logger... I0424 16:21:55.814335 140248041047872 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0424 16:21:55.814379 140248041047872 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 16:21:55.814757 140248041047872 base_pytree_checkpoint_handler.py:415] 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 0x7f8d93575310>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 16:21:58.711058 140248041047872 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 16:21:58.711327 140248041047872 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f790034a390>}, handler_registry=None I0424 16:21:58.711570 140248041047872 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f790034a390>` 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 16:21:58.711620 140248041047872 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f79003524e0>` 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 16:21:58.711656 140248041047872 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 0x7f790034a390>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f790034a390>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f79003524e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f79003524e0>}). I0424 16:21:58.712174 140248041047872 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 16:21:58.712252 140248041047872 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f78c04fc180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 16:21:59.771900 140248041047872 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints I0424 16:22:00.202876 140248041047872 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_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f8c46388b60> I0424 16:22:00.203052 140248041047872 checkpointing.py:302] Checkpoint manager created! I0424 16:22:01.097207 140248041047872 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 16:22:01.097315 140248041047872 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:22:01.466703 140248041047872 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 16:22:01.466803 140248041047872 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:22:01.482601 140248041047872 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 16:22:01.482663 140248041047872 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:22:01.505485 140248041047872 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:22:01.505553 140248041047872 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:22:01.521436 140248041047872 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:22:01.521495 140248041047872 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:22:01.537240 140248041047872 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:22:01.537305 140248041047872 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:22:01.562090 140248041047872 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 16:22:01.562173 140248041047872 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:22:01.581897 140248041047872 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 16:22:01.581963 140248041047872 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 16:22:01.779326 140248041047872 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 16:22:01.779439 140248041047872 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 16:22:03.223767 140248041047872 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0424 16:22:03.223898 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0424 16:22:03.223953 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0424 16:22:03.224011 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0424 16:22:03.224064 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0424 16:22:03.224117 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0424 16:22:03.224176 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0424 16:22:03.224232 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0424 16:22:03.224273 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0424 16:22:03.224311 140248041047872 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0424 16:22:03.224359 140248041047872 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0424 16:22:03.224406 140248041047872 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0424 16:22:03.727060 140248041047872 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 16:22:03.727173 140248041047872 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 16:22:03.742350 140248041047872 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 16:22:03.742413 140248041047872 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 16:22:14.550692 140248041047872 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 16:22:14.551620 140248041047872 metric_logger.py:289] number parameters: 1.104 billion I0424 16:22:25.798805 140248041047872 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 16:22:25.955667 140248041047872 checkpointing.py:798] Waited 0.1568460464477539 seconds for step 0 to finish before starting checkpointing. I0424 16:22:25.958041 140248041047872 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 16:22:25.959765 140248041047872 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 16:22:25.962626 140248041047872 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/0. I0424 16:22:26.940889 140248041047872 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 16:22:26.941941 140248041047872 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:22:26.942000 140248041047872 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 16:22:27.193998 140248041047872 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.253188s I0424 16:22:27.194193 140248041047872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.010 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5123989582061768 s) (per-host) I0424 16:22:27.194249 140248041047872 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.512466s (batch_requests_ready=0.255602s, total_serialization_initiated=0.256788s, others=0.000075s) I0424 16:22:27.194339 140248041047872 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.516663s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.516644) I0424 16:22:27.195316 140124381771520 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:22:27.195496 140248041047872 async_checkpointer.py:561] Finished blocking save. Time taken: 1.235657s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/0. I0424 16:22:27.232129 140248041047872 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:22:27.232449 140121152153344 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:22:27.232609 140248041047872 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047745.9580224, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777047745.9598045, 'checkpointer_blocking_duration_secs': 1.235799789428711, 'get_old_steps_start_time': 1777047747.195625, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1777047745.9562526, 'checkpoint_manager_blocking_duration_secs': 1.2763171195983887} I0424 16:22:27.232712 140248041047872 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 16:22:27.232763 140248041047872 max_utils.py:750] Memstats: After params initialized: I0424 16:22:27.232815 140248041047872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 16:22:27.232849 140248041047872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 16:22:27.232877 140248041047872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 16:22:27.232902 140248041047872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 16:22:27.499400 140248041047872 metric_logger.py:185] completed step: 0, seconds: 11.247, TFLOP/s/device: 1.208, Tokens/s/device: 182.092, total_weights: 65536, loss: 10.874 I0424 16:22:27.668221 140248041047872 metric_logger.py:185] completed step: 1, seconds: 1.699, TFLOP/s/device: 7.996, Tokens/s/device: 1205.216, total_weights: 65536, loss: 10.874 I0424 16:22:28.085227 140248041047872 metric_logger.py:185] completed step: 2, seconds: 0.022, TFLOP/s/device: 621.979, Tokens/s/device: 93751.431, total_weights: 65536, loss: 10.560 I0424 16:22:28.232743 140248041047872 metric_logger.py:185] completed step: 3, seconds: 0.417, TFLOP/s/device: 32.550, Tokens/s/device: 4906.224, total_weights: 65536, loss: 9.980 I0424 16:22:28.237167 140248041047872 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 16:22:28.527239 140248041047872 checkpointing.py:798] Waited 0.2900416851043701 seconds for step 5 to finish before starting checkpointing. I0424 16:22:28.530191 140248041047872 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:22:30.360749 2694 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 16:22:32.656400 140124868286208 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0424 16:23:09.280191 140124381771520 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.081 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.598355770111084 s) (per-host) I0424 16:23:09.280316 140124381771520 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.084884s. I0424 16:23:17.464782 140124381771520 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.269335s. I0424 16:23:17.465083 140121152153344 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:23:17.465242 140121152153344 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:23:17.465297 140121152153344 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:23:17.466717 140121152153344 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 16:23:17.466888 140248041047872 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0424 16:23:17.467027 140248041047872 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.936840 seconds. If this number is high, consider checkpointing less frequently. I0424 16:23:17.469200 140248041047872 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0424 16:23:17.473299 140248041047872 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/5. I0424 16:23:18.068153 140248041047872 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:23:18.068260 140248041047872 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 16:23:18.112794 140248041047872 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046026s I0424 16:23:18.112959 140248041047872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.290 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29160070419311523 s) (per-host) I0424 16:23:18.113011 140248041047872 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.291661s (batch_requests_ready=0.243716s, total_serialization_initiated=0.047878s, others=0.000067s) I0424 16:23:18.113108 140248041047872 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.296623s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.296607) I0424 16:23:18.114033 140124868286208 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:23:18.114172 140248041047872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.644899s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/5. I0424 16:23:18.142581 140248041047872 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:23:18.142868 140121152153344 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:23:18.143030 140248041047872 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047748.5301623, 'wait_for_prev_duration_secs': 48.93683958053589, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777047797.4692392, 'checkpointer_blocking_duration_secs': 0.6451089382171631, 'get_old_steps_start_time': 1777047798.1143684, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1777047748.5277991, 'checkpoint_manager_blocking_duration_secs': 49.615195512771606} I0424 16:23:18.143151 140248041047872 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 16:23:18.143699 140248041047872 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.211, Tokens/s/device: 13446.879, total_weights: 65536, loss: 9.460 I0424 16:23:18.149231 140248041047872 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.186, Tokens/s/device: 13895.202, total_weights: 65536, loss: 8.959 I0424 16:23:37.536558 140248041047872 metric_logger.py:185] completed step: 6, seconds: 49.911, TFLOP/s/device: 0.272, Tokens/s/device: 41.033, total_weights: 65536, loss: 8.469 I0424 16:23:37.684160 140248041047872 metric_logger.py:185] completed step: 7, seconds: 19.241, TFLOP/s/device: 0.706, Tokens/s/device: 106.437, total_weights: 65536, loss: 8.003 I0424 16:23:37.831708 140248041047872 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.596, Tokens/s/device: 13504.870, total_weights: 65536, loss: 7.572 I0424 16:23:37.978779 140248041047872 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 16:23:37.979368 140248041047872 checkpointing.py:798] Waited 0.0006077289581298828 seconds for step 9 to finish before starting checkpointing. I0424 16:23:37.981423 140248041047872 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:23:41.958684 140117914142464 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0424 16:24:18.719038 140124868286208 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.939 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.89763593673706 s) (per-host) I0424 16:24:18.719174 140124868286208 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.604923s. I0424 16:24:26.998598 140124868286208 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.884332s. I0424 16:24:26.998895 140121152153344 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:24:26.999029 140121152153344 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:24:26.999082 140121152153344 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:24:27.000523 140121152153344 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 16:24:27.000694 140248041047872 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0424 16:24:27.000829 140248041047872 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.019417 seconds. If this number is high, consider checkpointing less frequently. I0424 16:24:27.002322 140248041047872 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0424 16:24:27.006493 140248041047872 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/9. I0424 16:24:27.567469 140248041047872 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:24:27.567570 140248041047872 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 16:24:27.608827 140248041047872 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042270s I0424 16:24:27.608994 140248041047872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.053 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30530500411987305 s) (per-host) I0424 16:24:27.609046 140248041047872 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.305367s (batch_requests_ready=0.261020s, total_serialization_initiated=0.044280s, others=0.000067s) I0424 16:24:27.609167 140248041047872 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.309409s (all_items=0.000010s, per_item={'items': '0.00001001'}, temp_paths=0.309399) I0424 16:24:27.610005 140121135367936 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:24:27.610186 140248041047872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.607795s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/9. I0424 16:24:27.618827 140248041047872 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:24:27.619090 140121152153344 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:24:27.619261 140248041047872 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047817.9813955, 'wait_for_prev_duration_secs': 49.01941657066345, 'time_between_consecutive_saves_sec': 20.51463794708252, 'checkpointer_blocking_start_time': 1777047867.0023644, 'checkpointer_blocking_duration_secs': 0.607928991317749, 'get_old_steps_start_time': 1777047867.6103077, 'get_old_steps_duration_secs': 1.9073486328125e-05, 'checkpoint_manager_blocking_start_time': 1777047817.9796054, 'checkpoint_manager_blocking_duration_secs': 49.639625549316406} I0424 16:24:27.619369 140248041047872 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 16:24:27.619412 140248041047872 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:24:32.415401 140124381771520 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0424 16:25:09.879473 140121135367936 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.101 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.57574415206909 s) (per-host) I0424 16:25:09.879603 140121135367936 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.269469s. I0424 16:25:16.384343 140121135367936 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.774194s. I0424 16:25:16.384628 140121152153344 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:25:16.384752 140121152153344 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:25:16.384794 140121152153344 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:25:16.386384 140121152153344 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 16:25:16.386542 140248041047872 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0424 16:25:16.386688 140248041047872 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 16:25:16.387315 140248041047872 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.972, Tokens/s/device: 13863.035, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 24 16:25:30 UTC 2026 EXIT_CODE=0