XPK Start: Sat Apr 25 17:54:15 UTC 2026 2026-04-25 17:54:19.458615: 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 00:00:1777139659.471489 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777139659.475494 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777139659.487179 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777139659.487197 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777139659.487199 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777139659.487201 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:54:38.695666: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0425 17:54:39.221033 137335721744192 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 17:54:48,261:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 17:54:48.261960 137335721744192 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 17:54:48,264:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-gmuam-slice-job-0-0.mt-16-zarr3-false-sv-gmuam:8482 I0425 17:54:48.264393 137335721744192 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-gmuam-slice-job-0-0.mt-16-zarr3-false-sv-gmuam:8482 I0425 17:54:49.197297 137335721744192 max_utils.py:284] Jax distributed system initialized! I0425 17:54:56.299260 137335721744192 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 17:54:56.299363 137335721744192 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 17:54:56.299405 137335721744192 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 I0425 17:54:56.299440 137335721744192 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 17:54:56.984135 137335721744192 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 17:54:57.099158 137335721744192 checkpointing.py:688] Setting up checkpoint logger... I0425 17:54:57.099277 137335721744192 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0425 17:54:57.099322 137335721744192 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 17:54:57.099530 137335721744192 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 0x7ce782178ce0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 17:54:59.970857 137335721744192 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 17:54:59.971213 137335721744192 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 0x7cd2e02ac500>}, handler_registry=None I0425 17:54:59.971452 137335721744192 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cd2e02ac500>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0425 17:54:59.971506 137335721744192 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cd2f814a6c0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0425 17:54:59.971542 137335721744192 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 0x7cd2e02ac500>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cd2e02ac500>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cd2f814a6c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cd2f814a6c0>}). I0425 17:54:59.971889 137335721744192 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 17:54:59.971968 137335721744192 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7cd2e06a0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 17:55:00.739361 137335721744192 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints I0425 17:55:00.825573 137335721744192 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cd2f814b6e0> I0425 17:55:00.825742 137335721744192 checkpointing.py:302] Checkpoint manager created! I0425 17:55:01.773774 137335721744192 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 17:55:01.773895 137335721744192 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 17:55:03.195988 137335721744192 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 17:55:03.196127 137335721744192 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:55:03.196182 137335721744192 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:55:03.196228 137335721744192 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 17:55:03.196267 137335721744192 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:55:03.196300 137335721744192 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:55:03.196338 137335721744192 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:55:03.196376 137335721744192 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 17:55:03.196413 137335721744192 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:55:03.196450 137335721744192 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:55:03.196484 137335721744192 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 17:55:03.196523 137335721744192 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 17:55:03.592620 137335721744192 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:55:03.592773 137335721744192 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:55:03.598307 137335721744192 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:55:03.598407 137335721744192 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:55:03.760431 137335721744192 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:55:03.760579 137335721744192 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:55:03.922005 137335721744192 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:55:03.922158 137335721744192 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:55:04.093569 137335721744192 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:55:04.093710 137335721744192 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:55:04.253776 137335721744192 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:55:04.253912 137335721744192 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:55:04.410054 137335721744192 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:55:04.410204 137335721744192 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:55:04.573043 137335721744192 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 17:55:04.573205 137335721744192 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:55:04.597214 137335721744192 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 17:55:04.597318 137335721744192 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 17:55:15.584735 137335721744192 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0425 17:55:15.587218 137335721744192 metric_logger.py:289] number parameters: 1.104 billion I0425 17:55:26.812133 137335721744192 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 17:55:26.962958 137335721744192 checkpointing.py:798] Waited 0.15081167221069336 seconds for step 0 to finish before starting checkpointing. I0425 17:55:26.965198 137335721744192 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:55:26.966951 137335721744192 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 17:55:26.969639 137335721744192 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/0. I0425 17:55:27.531167 137335721744192 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 17:55:27.532137 137335721744192 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:55:27.532197 137335721744192 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0425 17:55:27.786965 137335721744192 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256931s I0425 17:55:27.787151 137335721744192 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.886 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5344736576080322 s) (per-host) I0425 17:55:27.787205 137335721744192 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.534538s (batch_requests_ready=0.270920s, total_serialization_initiated=0.263546s, others=0.000073s) I0425 17:55:27.787303 137335721744192 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.538526s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.538507) I0425 17:55:27.788324 137208963000064 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:55:27.788457 137335721744192 async_checkpointer.py:561] Finished blocking save. Time taken: 0.821432s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/0. I0425 17:55:27.813123 137335721744192 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:55:27.813418 137207855687424 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:55:27.813558 137335721744192 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777139726.9651787, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777139726.966991, 'checkpointer_blocking_duration_secs': 0.8216509819030762, 'get_old_steps_start_time': 1777139727.7886631, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777139726.9633842, 'checkpoint_manager_blocking_duration_secs': 0.8501369953155518} I0425 17:55:27.813667 137335721744192 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 17:55:27.813740 137335721744192 max_utils.py:750] Memstats: After params initialized: I0425 17:55:27.813793 137335721744192 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 17:55:27.813827 137335721744192 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 17:55:27.813855 137335721744192 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 17:55:27.813880 137335721744192 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 17:55:28.084107 137335721744192 metric_logger.py:185] completed step: 0, seconds: 11.224, TFLOP/s/device: 1.211, Tokens/s/device: 182.468, total_weights: 65536, loss: 10.874 I0425 17:55:28.257340 137335721744192 metric_logger.py:185] completed step: 1, seconds: 1.270, TFLOP/s/device: 10.697, Tokens/s/device: 1612.372, total_weights: 65536, loss: 10.874 I0425 17:55:28.695417 137335721744192 metric_logger.py:185] completed step: 2, seconds: 0.028, TFLOP/s/device: 490.457, Tokens/s/device: 73927.012, total_weights: 65536, loss: 10.560 I0425 17:55:28.842763 137335721744192 metric_logger.py:185] completed step: 3, seconds: 0.429, TFLOP/s/device: 31.686, Tokens/s/device: 4776.097, total_weights: 65536, loss: 9.979 I0425 17:55:28.848729 137335721744192 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 17:55:29.137840 137335721744192 checkpointing.py:798] Waited 0.289081335067749 seconds for step 5 to finish before starting checkpointing. I0425 17:55:29.140386 137335721744192 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:55:31.232047 2596 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 17:55:33.196714 137208971392768 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0425 17:56:09.828131 137208963000064 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: 42.57540154457092 s) (per-host) I0425 17:56:09.828255 137208963000064 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.039717s. I0425 17:56:18.159610 137208963000064 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.371056s. I0425 17:56:18.159851 137207855687424 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:56:18.159909 137207855687424 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:56:18.159959 137207855687424 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:56:18.161795 137207855687424 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 17:56:18.161969 137335721744192 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. W0425 17:56:18.162119 137335721744192 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.021720 seconds. If this number is high, consider checkpointing less frequently. I0425 17:56:18.164153 137335721744192 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0425 17:56:18.168121 137335721744192 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/5. I0425 17:56:18.727965 137335721744192 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:56:18.728125 137335721744192 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0425 17:56:18.773079 137335721744192 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047213s I0425 17:56:18.773268 137335721744192 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.090 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3030877113342285 s) (per-host) I0425 17:56:18.773319 137335721744192 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.303148s (batch_requests_ready=0.250672s, total_serialization_initiated=0.052403s, others=0.000073s) I0425 17:56:18.773410 137335721744192 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.307958s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.307941) I0425 17:56:18.774344 137208971392768 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:56:18.774502 137335721744192 async_checkpointer.py:561] Finished blocking save. Time taken: 0.610277s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/5. I0425 17:56:19.022137 137335721744192 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:56:19.022502 137207855687424 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:56:19.022672 137335721744192 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777139729.140356, 'wait_for_prev_duration_secs': 49.02172017097473, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777139778.1641915, 'checkpointer_blocking_duration_secs': 0.6104259490966797, 'get_old_steps_start_time': 1777139778.7746382, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1777139729.1382008, 'checkpoint_manager_blocking_duration_secs': 49.88443470001221} I0425 17:56:19.022782 137335721744192 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 17:56:19.023613 137335721744192 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.463, Tokens/s/device: 12580.394, total_weights: 65536, loss: 9.457 I0425 17:56:19.031244 137335721744192 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.394, Tokens/s/device: 13926.573, total_weights: 65536, loss: 8.955 I0425 17:56:38.890653 137335721744192 metric_logger.py:185] completed step: 6, seconds: 50.181, TFLOP/s/device: 0.271, Tokens/s/device: 40.812, total_weights: 65536, loss: 8.465 I0425 17:56:39.037984 137335721744192 metric_logger.py:185] completed step: 7, seconds: 19.714, TFLOP/s/device: 0.689, Tokens/s/device: 103.884, total_weights: 65536, loss: 7.998 I0425 17:56:39.185525 137335721744192 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.689, Tokens/s/device: 13368.146, total_weights: 65536, loss: 7.567 I0425 17:56:39.333942 137335721744192 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 17:56:39.334564 137335721744192 checkpointing.py:798] Waited 0.0006260871887207031 seconds for step 9 to finish before starting checkpointing. I0425 17:56:39.337291 137335721744192 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:56:42.313985 137206261860096 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0425 17:57:19.897417 137208971392768 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.715 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.4272027015686 s) (per-host) I0425 17:57:19.897529 137208971392768 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.123070s. I0425 17:57:26.856706 137208971392768 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.082238s. I0425 17:57:26.856995 137207855687424 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:57:26.857151 137207855687424 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:57:26.857202 137207855687424 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:57:26.859453 137207855687424 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 17:57:26.859622 137335721744192 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. W0425 17:57:26.859754 137335721744192 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.522472 seconds. If this number is high, consider checkpointing less frequently. I0425 17:57:26.861395 137335721744192 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 17:57:26.864737 137335721744192 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 17:57:27.446185 137335721744192 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:57:27.446285 137335721744192 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0425 17:57:27.495230 137335721744192 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051147s I0425 17:57:27.495397 137335721744192 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.023 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30712294578552246 s) (per-host) I0425 17:57:27.495447 137335721744192 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307183s (batch_requests_ready=0.251544s, total_serialization_initiated=0.055574s, others=0.000065s) I0425 17:57:27.495547 137335721744192 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311314s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.311302) I0425 17:57:27.496383 137208928306944 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:57:27.496514 137335721744192 async_checkpointer.py:561] Finished blocking save. Time taken: 0.635049s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 17:57:27.507666 137335721744192 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:57:27.507951 137207855687424 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:57:27.508116 137335721744192 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777139799.3372653, 'wait_for_prev_duration_secs': 47.52247214317322, 'time_between_consecutive_saves_sec': 21.17543125152588, 'checkpointer_blocking_start_time': 1777139846.8614347, 'checkpointer_blocking_duration_secs': 0.6352522373199463, 'get_old_steps_start_time': 1777139847.4967005, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1777139799.3347917, 'checkpoint_manager_blocking_duration_secs': 48.173282861709595} I0425 17:57:27.508228 137335721744192 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 17:57:27.508269 137335721744192 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:57:32.323558 137208963000064 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0425 17:58:10.321012 137208928306944 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.622 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.132697105407715 s) (per-host) I0425 17:58:10.321151 137208928306944 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.824558s. I0425 17:58:15.719227 137208928306944 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.222617s. I0425 17:58:15.719506 137207855687424 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:58:15.719621 137207855687424 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:58:15.719663 137207855687424 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:58:15.721126 137207855687424 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 17:58:15.721309 137335721744192 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. I0425 17:58:15.721478 137335721744192 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:58:15.722192 137335721744192 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.563, Tokens/s/device: 13952.094, 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: Sat Apr 25 17:58:25 UTC 2026 EXIT_CODE=0