XPK Start: Sun Apr 19 05:04:18 UTC 2026 2026-04-19 05:04:22.908728: 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 00:00:1776575062.921270 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776575062.924954 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776575062.936144 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776575062.936161 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776575062.936163 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776575062.936165 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 05:04:42.066772: 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) I0419 05:04:42.594877 139097093625664 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 05:04:51,636:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 05:04:51.636202 139097093625664 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 05:04:51,638:jax._src.distributed:157: Connecting to JAX distributed service on mt-04-int8-c2q2x-slice-job-0-0.mt-04-int8-c2q2x:8482 I0419 05:04:51.638539 139097093625664 distributed.py:157] Connecting to JAX distributed service on mt-04-int8-c2q2x-slice-job-0-0.mt-04-int8-c2q2x:8482 I0419 05:04:53.183069 139097093625664 max_utils.py:284] Jax distributed system initialized! I0419 05:05:00.074766 139097093625664 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 05:05:00.074871 139097093625664 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 05:05:00.074913 139097093625664 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 I0419 05:05:00.074952 139097093625664 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 05:05:00.697822 139097093625664 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 05:05:01.262380 139097093625664 checkpointing.py:688] Setting up checkpoint logger... I0419 05:05:01.262498 139097093625664 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 05:05:01.262543 139097093625664 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 05:05:01.262763 139097093625664 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e8187f792e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 05:05:04.983312 139097093625664 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 05:05:04.983666 139097093625664 checkpoint_manager.py:709] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e804b8d1910>}, handler_registry=None I0419 05:05:04.983945 139097093625664 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e804b8d1910>` 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`. I0419 05:05:04.983999 139097093625664 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e6f204a14f0>` 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`. I0419 05:05:04.984037 139097093625664 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 0x7e804b8d1910>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e804b8d1910>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e6f204a14f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e6f204a14f0>}). I0419 05:05:04.984405 139097093625664 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 05:05:04.984485 139097093625664 async_checkpointer.py:177] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e6f21408cc0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 05:05:06.962773 139097093625664 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints I0419 05:05:06.964928 139097093625664 checkpoint_manager.py:929] [process=3][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e6f21ed5a90> I0419 05:05:06.965043 139097093625664 checkpointing.py:302] Checkpoint manager created! I0419 05:05:08.440702 139097093625664 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 05:05:08.440829 139097093625664 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 05:05:10.807078 139097093625664 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 05:05:10.807200 139097093625664 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 05:05:10.807269 139097093625664 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 05:05:10.807318 139097093625664 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 05:05:10.807356 139097093625664 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 05:05:10.807391 139097093625664 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 05:05:10.807428 139097093625664 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:05:10.807464 139097093625664 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 05:05:10.807498 139097093625664 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:05:10.807531 139097093625664 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:05:10.807563 139097093625664 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 05:05:10.807595 139097093625664 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 05:05:11.220892 139097093625664 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 05:05:11.221045 139097093625664 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:05:11.226506 139097093625664 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 05:05:11.226599 139097093625664 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:05:11.377521 139097093625664 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 05:05:11.377654 139097093625664 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:05:11.528420 139097093625664 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 05:05:11.528547 139097093625664 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:05:11.732143 139097093625664 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 05:05:11.732291 139097093625664 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:05:11.886173 139097093625664 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 05:05:11.886710 139097093625664 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:05:12.041821 139097093625664 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 05:05:12.041958 139097093625664 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:05:12.203791 139097093625664 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 05:05:12.203927 139097093625664 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:05:12.697833 139097093625664 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 05:05:12.697957 139097093625664 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 05:05:26.357940 139097093625664 max_utils.py:791] Total memory size: 1.8 GB, Output size: 0.4 GB, Temp size: 1.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0419 05:05:26.361512 139097093625664 metric_logger.py:289] number parameters: 1.104 billion I0419 05:05:40.774645 139097093625664 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 05:05:41.009565 139097093625664 checkpointing.py:798] Waited 0.23490405082702637 seconds for step 0 to finish before starting checkpointing. I0419 05:05:41.012091 139097093625664 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 05:05:41.014000 139097093625664 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 0 I0419 05:05:41.017042 139097093625664 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints/0. I0419 05:05:42.045147 139097093625664 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 05:05:42.046217 139097093625664 jax_array_handlers.py:358] Scheduling D2H of 111 prioritized jax.Array. I0419 05:05:42.046292 139097093625664 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 I0419 05:05:42.296972 139097093625664 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254287s I0419 05:05:42.297144 139097093625664 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.840 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5432195663452148 s) (per-host) I0419 05:05:42.297196 139097093625664 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.543283s (batch_requests_ready=0.278979s, total_serialization_initiated=0.264233s, others=0.000072s) I0419 05:05:42.297337 139097093625664 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.547449s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.547430) I0419 05:05:42.298293 138973934515968 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0419 05:05:42.298464 139097093625664 async_checkpointer.py:561] Finished blocking save. Time taken: 1.284395s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints/0. I0419 05:05:42.714468 139097093625664 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 05:05:42.714854 138962794444544 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0419 05:05:42.715036 139097093625664 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776575141.0120716, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776575141.0140405, 'checkpointer_blocking_duration_secs': 1.2845308780670166, 'get_old_steps_start_time': 1776575142.2985928, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776575141.01005, 'checkpoint_manager_blocking_duration_secs': 1.7049455642700195} I0419 05:05:42.715151 139097093625664 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 05:05:42.715248 139097093625664 max_utils.py:750] Memstats: After params initialized: I0419 05:05:42.715303 139097093625664 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0419 05:05:42.715335 139097093625664 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0419 05:05:42.715359 139097093625664 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0419 05:05:42.715382 139097093625664 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0419 05:05:42.992298 139097093625664 metric_logger.py:185] completed step: 0, seconds: 14.411, TFLOP/s/device: 0.943, Tokens/s/device: 142.110, total_weights: 65536, loss: 10.872 I0419 05:05:43.257203 139097093625664 metric_logger.py:185] completed step: 1, seconds: 2.215, TFLOP/s/device: 6.134, Tokens/s/device: 924.574, total_weights: 65536, loss: 10.872 I0419 05:05:43.675381 139097093625664 metric_logger.py:185] completed step: 2, seconds: 0.040, TFLOP/s/device: 340.248, Tokens/s/device: 51285.904, total_weights: 65536, loss: 10.258 I0419 05:05:43.904653 139097093625664 metric_logger.py:185] completed step: 3, seconds: 0.400, TFLOP/s/device: 33.982, Tokens/s/device: 5122.100, total_weights: 65536, loss: 9.727 I0419 05:05:44.363907 139097093625664 metric_logger.py:185] completed step: 4, seconds: 0.253, TFLOP/s/device: 53.671, Tokens/s/device: 8089.873, total_weights: 65536, loss: 9.270 I0419 05:05:44.372689 139097093625664 metric_logger.py:185] completed step: 5, seconds: 0.229, TFLOP/s/device: 59.351, Tokens/s/device: 8945.966, total_weights: 65536, loss: 8.884 I0419 05:06:07.680650 139097093625664 metric_logger.py:185] completed step: 6, seconds: 0.460, TFLOP/s/device: 29.561, Tokens/s/device: 4455.768, total_weights: 65536, loss: 8.584 I0419 05:06:07.909822 139097093625664 metric_logger.py:185] completed step: 7, seconds: 23.082, TFLOP/s/device: 0.589, Tokens/s/device: 88.728, total_weights: 65536, loss: 8.376 I0419 05:06:08.139209 139097093625664 metric_logger.py:185] completed step: 8, seconds: 0.235, TFLOP/s/device: 57.868, Tokens/s/device: 8722.540, total_weights: 65536, loss: 8.247 I0419 05:06:08.370366 139097093625664 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 05:06:08.371281 139097093625664 checkpointing.py:798] Waited 0.0009286403656005859 seconds for step 9 to finish before starting checkpointing. I0419 05:06:08.373266 139097093625664 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 05:06:09.116825 2817 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 05:06:11.369723 138974421030656 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 111 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints/0/items/array_metadatas/process_3 I0419 05:06:37.514874 138973934515968 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 28.328 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 55.7609076499939 s) (per-host) I0419 05:06:37.514996 138973934515968 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.216593s. I0419 05:06:48.308470 138973934515968 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 66.010051s. I0419 05:06:48.308796 138962794444544 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 05:06:48.308933 138962794444544 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0419 05:06:48.308983 138962794444544 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 05:06:48.310627 138962794444544 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 05:06:48.310828 139097093625664 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 05:06:48.310964 139097093625664 checkpoint_manager.py:1458] Waiting for previous save to complete took 39.937718 seconds. If this number is high, consider checkpointing less frequently. I0419 05:06:48.312956 139097093625664 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 9 I0419 05:06:48.316541 139097093625664 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints/9. I0419 05:06:49.717124 139097093625664 jax_array_handlers.py:358] Scheduling D2H of 111 prioritized jax.Array. I0419 05:06:49.717229 139097093625664 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 I0419 05:06:49.760526 139097093625664 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046870s I0419 05:06:49.760691 139097093625664 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.114 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7295324802398682 s) (per-host) I0419 05:06:49.760743 139097093625664 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.729603s (batch_requests_ready=0.674530s, total_serialization_initiated=0.054994s, others=0.000078s) I0419 05:06:49.760859 139097093625664 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.733691s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.733675) I0419 05:06:49.761738 138967332669184 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0419 05:06:49.761865 139097093625664 async_checkpointer.py:561] Finished blocking save. Time taken: 1.448838s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints/9. I0419 05:06:49.781013 139097093625664 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 05:06:49.781302 138962794444544 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0419 05:06:49.781441 139097093625664 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776575168.373221, 'wait_for_prev_duration_secs': 39.93771839141846, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776575208.3129954, 'checkpointer_blocking_duration_secs': 1.4490489959716797, 'get_old_steps_start_time': 1776575209.7620633, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776575168.3715775, 'checkpoint_manager_blocking_duration_secs': 41.40982913970947} I0419 05:06:49.781567 139097093625664 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 05:06:49.781613 139097093625664 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 05:06:55.345360 138974421030656 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 111 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_04_int8/checkpoints/9/items/array_metadatas/process_3 I0419 05:07:31.935446 138967332669184 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 36.817 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.9042592048645 s) (per-host) I0419 05:07:31.935550 138967332669184 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.173606s. I0419 05:07:40.367798 138967332669184 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 50.605837s. I0419 05:07:40.368114 138962794444544 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 05:07:40.368263 138962794444544 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0419 05:07:40.368314 138962794444544 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 05:07:40.370828 138962794444544 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 05:07:40.371005 139097093625664 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 05:07:40.371190 139097093625664 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 05:07:40.371969 139097093625664 metric_logger.py:185] completed step: 9, seconds: 0.229, TFLOP/s/device: 59.324, Tokens/s/device: 8941.943, total_weights: 65536, loss: 8.171 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 05:07:49 UTC 2026 EXIT_CODE=0