XPK Start: Fri Apr 24 17:32:13 UTC 2026 2026-04-24 17:32:17.655274: 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 00:00:1777051937.668422 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777051937.672451 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777051937.684163 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777051937.684182 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777051937.684185 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777051937.684186 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 17:32:36.978982: 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 17:32:37.508497 138512696915776 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 17:32:46,549:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 17:32:46.549843 138512696915776 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 17:32:46,552:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-sv-bjgou-slice-job-0-0.mt-15-ocdbt-false-sv-bjgou:8482 I0424 17:32:46.552342 138512696915776 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-bjgou-slice-job-0-0.mt-15-ocdbt-false-sv-bjgou:8482 I0424 17:32:47.689333 138512696915776 max_utils.py:284] Jax distributed system initialized! I0424 17:32:54.506684 138512696915776 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 17:32:54.506790 138512696915776 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 17:32:54.506830 138512696915776 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 17:32:54.506867 138512696915776 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 17:32:55.125878 138512696915776 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 17:32:55.238491 138512696915776 checkpointing.py:688] Setting up checkpoint logger... I0424 17:32:55.238607 138512696915776 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0424 17:32:55.238652 138512696915776 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 17:32:55.238857 138512696915776 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7df974379340>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 17:32:58.193312 138512696915776 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 17:32:58.193651 138512696915776 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 0x7de672eed760>}, handler_registry=None I0424 17:32:58.193886 138512696915776 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7de672eed760>` 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 17:32:58.193934 138512696915776 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7de672eed4c0>` 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 17:32:58.193970 138512696915776 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 0x7de672eed760>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7de672eed760>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7de672eed4c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7de672eed4c0>}). I0424 17:32:58.194327 138512696915776 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 17:32:58.194403 138512696915776 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7de504230f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 17:32:58.998694 138512696915776 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints I0424 17:32:59.076293 138512696915776 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/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7de672eef080> I0424 17:32:59.076462 138512696915776 checkpointing.py:302] Checkpoint manager created! I0424 17:33:00.009080 138512696915776 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 17:33:00.009222 138512696915776 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 17:33:01.420661 138512696915776 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 17:33:01.420784 138512696915776 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:33:01.420838 138512696915776 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 17:33:01.420880 138512696915776 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 17:33:01.420917 138512696915776 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:33:01.420951 138512696915776 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 17:33:01.420987 138512696915776 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:33:01.421023 138512696915776 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 17:33:01.421057 138512696915776 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:33:01.421091 138512696915776 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 17:33:01.421141 138512696915776 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 17:33:01.421175 138512696915776 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 17:33:01.826442 138512696915776 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:33:01.826600 138512696915776 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:33:01.831926 138512696915776 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 17:33:01.832020 138512696915776 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:33:01.983724 138512696915776 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:33:01.983858 138512696915776 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:33:02.132735 138512696915776 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 17:33:02.132868 138512696915776 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 17:33:02.296427 138512696915776 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 17:33:02.296570 138512696915776 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:33:02.440431 138512696915776 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 17:33:02.440560 138512696915776 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:33:02.579555 138512696915776 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 17:33:02.579679 138512696915776 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:33:02.729235 138512696915776 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 17:33:02.729369 138512696915776 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 17:33:02.753151 138512696915776 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 17:33:02.753246 138512696915776 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 17:33:13.739140 138512696915776 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 17:33:13.741693 138512696915776 metric_logger.py:289] number parameters: 1.104 billion I0424 17:33:24.827790 138512696915776 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 17:33:24.978596 138512696915776 checkpointing.py:798] Waited 0.15079116821289062 seconds for step 0 to finish before starting checkpointing. I0424 17:33:24.980870 138512696915776 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:33:24.982482 138512696915776 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 17:33:24.985138 138512696915776 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/0. I0424 17:33:25.849004 138512696915776 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 17:33:25.850159 138512696915776 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:33:25.850219 138512696915776 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 17:33:26.096227 138512696915776 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.248328s I0424 17:33:26.096406 138512696915776 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.831 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5449395179748535 s) (per-host) I0424 17:33:26.096457 138512696915776 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.545000s (batch_requests_ready=0.290108s, total_serialization_initiated=0.254824s, others=0.000067s) I0424 17:33:26.096551 138512696915776 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.549169s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.549149) I0424 17:33:26.097566 138390649415424 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:33:26.097728 138512696915776 async_checkpointer.py:561] Finished blocking save. Time taken: 1.115172s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/0. I0424 17:33:26.112515 138512696915776 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:33:26.112817 138385859540736 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:33:26.112969 138512696915776 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052004.9808521, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052004.982523, 'checkpointer_blocking_duration_secs': 1.1153109073638916, 'get_old_steps_start_time': 1777052006.0978546, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1777052004.9789946, 'checkpoint_manager_blocking_duration_secs': 1.1339390277862549} I0424 17:33:26.113077 138512696915776 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 17:33:26.113164 138512696915776 max_utils.py:750] Memstats: After params initialized: I0424 17:33:26.113214 138512696915776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 17:33:26.113246 138512696915776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 17:33:26.113273 138512696915776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 17:33:26.113303 138512696915776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 17:33:26.375116 138512696915776 metric_logger.py:185] completed step: 0, seconds: 11.085, TFLOP/s/device: 1.226, Tokens/s/device: 184.753, total_weights: 65536, loss: 10.874 I0424 17:33:26.558195 138512696915776 metric_logger.py:185] completed step: 1, seconds: 1.546, TFLOP/s/device: 8.791, Tokens/s/device: 1325.131, total_weights: 65536, loss: 10.874 I0424 17:33:26.995997 138512696915776 metric_logger.py:185] completed step: 2, seconds: 0.038, TFLOP/s/device: 362.140, Tokens/s/device: 54585.677, total_weights: 65536, loss: 10.560 I0424 17:33:27.143478 138512696915776 metric_logger.py:185] completed step: 3, seconds: 0.428, TFLOP/s/device: 31.713, Tokens/s/device: 4780.066, total_weights: 65536, loss: 9.979 I0424 17:33:27.149367 138512696915776 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 17:33:27.438359 138512696915776 checkpointing.py:798] Waited 0.28896117210388184 seconds for step 5 to finish before starting checkpointing. I0424 17:33:27.443762 138512696915776 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:33:28.826857 2586 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 17:33:30.769773 138390657808128 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_6 I0424 17:33:36.274847 138390649415424 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 147.305 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.723319053649902 s) (per-host) I0424 17:33:36.274971 138390649415424 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.177297s. I0424 17:33:43.074841 138390649415424 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.977148s. I0424 17:33:43.075160 138385859540736 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:33:43.075286 138385859540736 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:33:43.075338 138385859540736 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:33:43.076712 138385859540736 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 17:33:43.076890 138512696915776 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 17:33:43.077021 138512696915776 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.633265 seconds. If this number is high, consider checkpointing less frequently. I0424 17:33:43.079200 138512696915776 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0424 17:33:43.083294 138512696915776 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/5. I0424 17:33:43.617981 138512696915776 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:33:43.618149 138512696915776 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 17:33:43.663185 138512696915776 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047407s I0424 17:33:43.663349 138512696915776 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.118 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3013927936553955 s) (per-host) I0424 17:33:43.663399 138512696915776 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.301452s (batch_requests_ready=0.248893s, total_serialization_initiated=0.052493s, others=0.000066s) I0424 17:33:43.663501 138512696915776 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.306327s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.306312) I0424 17:33:43.664394 138382075492096 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:33:43.664521 138512696915776 async_checkpointer.py:561] Finished blocking save. Time taken: 0.585251s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/5. I0424 17:33:44.020858 138512696915776 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:33:44.021242 138385859540736 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:33:44.021417 138512696915776 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052007.4437332, 'wait_for_prev_duration_secs': 15.633265018463135, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777052023.079239, 'checkpointer_blocking_duration_secs': 0.585453987121582, 'get_old_steps_start_time': 1777052023.6647112, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777052007.438702, 'checkpoint_manager_blocking_duration_secs': 16.582676887512207} I0424 17:33:44.021532 138512696915776 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 17:33:44.022338 138512696915776 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.357, Tokens/s/device: 12564.494, total_weights: 65536, loss: 9.457 I0424 17:33:44.029851 138512696915776 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.404, Tokens/s/device: 13928.183, total_weights: 65536, loss: 8.955 I0424 17:34:03.970604 138512696915776 metric_logger.py:185] completed step: 6, seconds: 16.879, TFLOP/s/device: 0.805, Tokens/s/device: 121.336, total_weights: 65536, loss: 8.465 I0424 17:34:04.118024 138512696915776 metric_logger.py:185] completed step: 7, seconds: 19.795, TFLOP/s/device: 0.686, Tokens/s/device: 103.459, total_weights: 65536, loss: 7.998 I0424 17:34:04.265297 138512696915776 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.399, Tokens/s/device: 13324.485, total_weights: 65536, loss: 7.567 I0424 17:34:04.413681 138512696915776 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 17:34:04.414303 138512696915776 checkpointing.py:798] Waited 0.0006334781646728516 seconds for step 9 to finish before starting checkpointing. I0424 17:34:04.416277 138512696915776 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:34:08.060600 138390657808128 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_6 I0424 17:34:13.991413 138382075492096 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 51.571 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 30.629413843154907 s) (per-host) I0424 17:34:13.991550 138382075492096 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.326948s. I0424 17:34:19.584715 138382075492096 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 35.920100s. I0424 17:34:19.585026 138385859540736 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:34:19.585165 138385859540736 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:34:19.585215 138385859540736 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:34:19.586592 138385859540736 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 17:34:19.586785 138512696915776 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 17:34:19.586910 138512696915776 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.170644 seconds. If this number is high, consider checkpointing less frequently. I0424 17:34:19.588606 138512696915776 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0424 17:34:19.592043 138512696915776 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/9. I0424 17:34:20.153008 138512696915776 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0424 17:34:20.153125 138512696915776 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 17:34:20.195740 138512696915776 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044746s I0424 17:34:20.195899 138512696915776 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.008 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3080451488494873 s) (per-host) I0424 17:34:20.195947 138512696915776 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.308101s (batch_requests_ready=0.258606s, total_serialization_initiated=0.049434s, others=0.000061s) I0424 17:34:20.196040 138512696915776 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312069s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.312058) I0424 17:34:20.196911 138383133222656 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 17:34:20.197067 138512696915776 async_checkpointer.py:561] Finished blocking save. Time taken: 0.608394s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/9. I0424 17:34:20.205746 138512696915776 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 17:34:20.206007 138385859540736 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 17:34:20.206152 138512696915776 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777052044.4162498, 'wait_for_prev_duration_secs': 15.170644283294678, 'time_between_consecutive_saves_sec': 21.339499950408936, 'checkpointer_blocking_start_time': 1777052059.5886443, 'checkpointer_blocking_duration_secs': 0.608543872833252, 'get_old_steps_start_time': 1777052060.197202, 'get_old_steps_duration_secs': 1.9073486328125e-05, 'checkpoint_manager_blocking_start_time': 1777052044.4145226, 'checkpoint_manager_blocking_duration_secs': 15.79159927368164} I0424 17:34:20.206319 138512696915776 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 17:34:20.206361 138512696915776 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 17:34:24.962905 138390657808128 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/nnx_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_6 I0424 17:34:30.821040 138383133222656 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 144.478 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.933144330978394 s) (per-host) I0424 17:34:30.821155 138383133222656 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.624134s. I0424 17:34:36.687550 138383133222656 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.490514s. I0424 17:34:36.687883 138385859540736 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 17:34:36.688002 138385859540736 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 17:34:36.688044 138385859540736 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 17:34:36.689732 138385859540736 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 17:34:36.689925 138512696915776 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 17:34:36.690092 138512696915776 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 17:34:36.690753 138512696915776 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.547, Tokens/s/device: 13949.623, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 24 17:34:49 UTC 2026 EXIT_CODE=0