XPK Start: Sat Apr 25 17:02:58 UTC 2026 2026-04-25 17:03:02.216313: 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 00:00:1777136582.229226 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777136582.233187 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777136582.244715 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777136582.244734 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777136582.244736 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777136582.244738 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:03:21.476790: 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 17:03:22.004485 132949805041472 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 17:03:31,046:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 17:03:31.046013 132949805041472 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 17:03:31,048:jax._src.distributed:157: Connecting to JAX distributed service on mt-09-pdb-lt-1-o5a8c-slice-job-0-0.mt-09-pdb-lt-1-o5a8c:8482 I0425 17:03:31.048333 132949805041472 distributed.py:157] Connecting to JAX distributed service on mt-09-pdb-lt-1-o5a8c-slice-job-0-0.mt-09-pdb-lt-1-o5a8c:8482 I0425 17:03:31.877968 132949805041472 max_utils.py:284] Jax distributed system initialized! I0425 17:03:38.688952 132949805041472 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 17:03:38.689057 132949805041472 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 17:03:38.689121 132949805041472 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 17:03:38.689158 132949805041472 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 17:03:39.310506 132949805041472 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 8, 1, 1, 1, 1, 4, 1, 1, 1, 1) I0425 17:03:39.424725 132949805041472 checkpointing.py:688] Setting up checkpoint logger... I0425 17:03:39.424853 132949805041472 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 17:03:39.424909 132949805041472 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 17:03:39.425155 132949805041472 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 0x78ea459794f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 17:03:42.289715 132949805041472 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 17:03:42.290062 132949805041472 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 0x78e9049e9c40>}, handler_registry=None I0425 17:03:42.290323 132949805041472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78e9049e9c40>` 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 17:03:42.290371 132949805041472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78d6dd7301a0>` 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 17:03:42.290407 132949805041472 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 0x78e9049e9c40>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78e9049e9c40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78d6dd7301a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78d6dd7301a0>}). I0425 17:03:42.290748 132949805041472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 17:03:42.290827 132949805041472 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78d6d7eccf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 17:03:43.122914 132949805041472 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_09_pdb_lt_1/checkpoints I0425 17:03:43.312670 132949805041472 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=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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_09_pdb_lt_1/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78d5ac6889e0> I0425 17:03:43.312844 132949805041472 checkpointing.py:302] Checkpoint manager created! I0425 17:03:44.246775 132949805041472 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 17:03:44.246892 132949805041472 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 8 tensor: 4 I0425 17:03:45.726905 132949805041472 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: ('tensor',) I0425 17:03:45.727027 132949805041472 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, 'tensor') I0425 17:03:45.727082 132949805041472 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, 'tensor') I0425 17:03:45.727144 132949805041472 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: ('tensor', None, 'fsdp') I0425 17:03:45.727183 132949805041472 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: ('tensor', None) I0425 17:03:45.727218 132949805041472 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: ('tensor', None) I0425 17:03:45.727259 132949805041472 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, 'tensor', None) I0425 17:03:45.727295 132949805041472 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: ('tensor', None, None, 'fsdp') I0425 17:03:45.727331 132949805041472 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, 'tensor', None) I0425 17:03:45.727364 132949805041472 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, 'tensor', None) I0425 17:03:45.727396 132949805041472 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', 'tensor') I0425 17:03:45.727428 132949805041472 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: ('tensor', 'fsdp') I0425 17:03:46.128484 132949805041472 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:03:46.128639 132949805041472 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0425 17:03:46.134162 132949805041472 nnx_decoders.py:616] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:03:46.134265 132949805041472 nnx_decoders.py:616] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0425 17:03:46.297410 132949805041472 attentions.py:1084] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:03:46.297550 132949805041472 attentions.py:1084] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0425 17:03:46.455685 132949805041472 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:03:46.455821 132949805041472 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0425 17:03:46.618961 132949805041472 attentions.py:1150] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:03:46.619120 132949805041472 attentions.py:1150] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0425 17:03:46.768601 132949805041472 attentions.py:1151] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:03:46.768746 132949805041472 attentions.py:1151] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0425 17:03:46.917859 132949805041472 attentions.py:1152] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:03:46.918003 132949805041472 attentions.py:1152] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0425 17:03:47.075174 132949805041472 attentions.py:1193] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 17:03:47.075317 132949805041472 attentions.py:1193] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0425 17:03:47.098112 132949805041472 linears.py:541] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 17:03:47.098212 132949805041472 linears.py:541] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, 'tensor'). I0425 17:03:57.463768 132949805041472 max_utils.py:791] Total memory size: 0.9 GB, Output size: 0.4 GB, Temp size: 0.5 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0425 17:03:57.466862 132949805041472 metric_logger.py:289] number parameters: 1.104 billion I0425 17:03:58.772790 132949805041472 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 17:04:07.029731 132949805041472 checkpointing.py:798] Waited 8.256924629211426 seconds for step 0 to finish before starting checkpointing. I0425 17:04:07.032632 132949805041472 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:04:07.034237 132949805041472 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 17:04:07.036922 132949805041472 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_09_pdb_lt_1/checkpoints/0. I0425 17:04:07.591923 132949805041472 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 17:04:07.592964 132949805041472 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:04:07.593024 132949805041472 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 17:04:07.822349 132949805041472 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.231544s I0425 17:04:07.822522 132949805041472 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.118 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.49469542503356934 s) (per-host) I0425 17:04:07.822575 132949805041472 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.494758s (batch_requests_ready=0.256842s, total_serialization_initiated=0.237846s, others=0.000070s) I0425 17:04:07.822675 132949805041472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.499284s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.499265) I0425 17:04:07.823606 132826577999616 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:04:07.823776 132949805041472 async_checkpointer.py:561] Finished blocking save. Time taken: 0.789468s. 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_09_pdb_lt_1/checkpoints/0. I0425 17:04:07.840332 132949805041472 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:04:07.840613 132823356774144 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:04:07.840767 132949805041472 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_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777136647.032615, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777136647.0342762, 'checkpointer_blocking_duration_secs': 0.7896063327789307, 'get_old_steps_start_time': 1777136647.8239026, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777136647.0301485, 'checkpoint_manager_blocking_duration_secs': 0.810582160949707} I0425 17:04:07.840873 132949805041472 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 17:04:07.840940 132949805041472 max_utils.py:750] Memstats: After params initialized: I0425 17:04:07.840987 132949805041472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 17:04:07.841017 132949805041472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 17:04:07.841043 132949805041472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 17:04:07.841067 132949805041472 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 17:04:08.107284 132949805041472 metric_logger.py:185] completed step: 0, seconds: 1.305, TFLOP/s/device: 2.603, Tokens/s/device: 392.375, total_weights: 16384, loss: 10.864 I0425 17:04:08.181398 132949805041472 metric_logger.py:185] completed step: 1, seconds: 9.333, TFLOP/s/device: 0.364, Tokens/s/device: 54.860, total_weights: 16384, loss: 10.864 I0425 17:04:08.628345 132949805041472 metric_logger.py:185] completed step: 2, seconds: 0.018, TFLOP/s/device: 185.202, Tokens/s/device: 27915.599, total_weights: 16384, loss: 9.760 I0425 17:04:08.686798 132949805041472 metric_logger.py:185] completed step: 3, seconds: 0.437, TFLOP/s/device: 7.777, Tokens/s/device: 1172.295, total_weights: 16384, loss: 8.820 I0425 17:04:08.804620 132949805041472 metric_logger.py:185] completed step: 4, seconds: 0.074, TFLOP/s/device: 46.049, Tokens/s/device: 6940.961, total_weights: 16384, loss: 7.947 I0425 17:04:08.812174 132949805041472 metric_logger.py:185] completed step: 5, seconds: 0.058, TFLOP/s/device: 58.753, Tokens/s/device: 8855.833, total_weights: 16384, loss: 7.209 I0425 17:04:21.505503 132949805041472 metric_logger.py:185] completed step: 6, seconds: 0.119, TFLOP/s/device: 28.600, Tokens/s/device: 4310.889, total_weights: 16384, loss: 6.651 I0425 17:04:21.564041 132949805041472 metric_logger.py:185] completed step: 7, seconds: 12.636, TFLOP/s/device: 0.269, Tokens/s/device: 40.518, total_weights: 16384, loss: 6.267 I0425 17:04:21.622632 132949805041472 metric_logger.py:185] completed step: 8, seconds: 0.064, TFLOP/s/device: 52.967, Tokens/s/device: 7983.783, total_weights: 16384, loss: 6.030 I0425 17:04:21.682078 132949805041472 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 17:04:21.682683 132949805041472 checkpointing.py:798] Waited 0.0006122589111328125 seconds for step 9 to finish before starting checkpointing. I0425 17:04:21.684493 132949805041472 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:04:23.101111 2432 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 17:04:25.375812 132826586392320 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_09_pdb_lt_1/checkpoints/0/items/array_metadatas/process_6 I0425 17:04:54.767272 132826577999616 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 33.297 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.439401149749756 s) (per-host) I0425 17:04:54.767383 132826577999616 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.943664s. I0425 17:05:04.544293 132826577999616 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 56.720556s. I0425 17:05:04.544609 132823356774144 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:05:04.544734 132823356774144 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:05:04.544786 132823356774144 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:05:04.546238 132823356774144 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 17:05:04.546433 132949805041472 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 17:05:04.546588 132949805041472 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.862097 seconds. If this number is high, consider checkpointing less frequently. I0425 17:05:04.548654 132949805041472 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 17:05:04.552134 132949805041472 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_09_pdb_lt_1/checkpoints/9. I0425 17:05:05.101044 132949805041472 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:05:05.101216 132949805041472 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 17:05:05.148033 132949805041472 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048994s I0425 17:05:05.148228 132949805041472 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.911 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31408166885375977 s) (per-host) I0425 17:05:05.148284 132949805041472 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314148s (batch_requests_ready=0.260499s, total_serialization_initiated=0.053574s, others=0.000075s) I0425 17:05:05.148390 132949805041472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318351s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.318336) I0425 17:05:05.149360 132823356774144 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:05:05.149514 132949805041472 async_checkpointer.py:561] Finished blocking save. Time taken: 0.600789s. 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_09_pdb_lt_1/checkpoints/9. I0425 17:05:05.162232 132949805041472 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:05:05.162454 132826586392320 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:05:05.162583 132949805041472 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_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777136661.6844656, 'wait_for_prev_duration_secs': 42.862096548080444, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777136704.5486922, 'checkpointer_blocking_duration_secs': 0.600928783416748, 'get_old_steps_start_time': 1777136705.1496382, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1777136661.6829066, 'checkpoint_manager_blocking_duration_secs': 43.47964358329773} I0425 17:05:05.162695 132949805041472 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 17:05:05.162736 132949805041472 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:05:09.855586 132815354066688 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_09_pdb_lt_1/checkpoints/9/items/array_metadatas/process_6 I0425 17:05:44.934533 132823356774144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 39.391 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.10034513473511 s) (per-host) I0425 17:05:44.934648 132823356774144 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.785182s. I0425 17:05:54.027197 132823356774144 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.877714s. I0425 17:05:54.027458 132826586392320 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:05:54.027523 132826586392320 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:05:54.027570 132826586392320 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:05:54.029009 132826586392320 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 17:05:54.029140 132949805041472 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 17:05:54.029224 132949805041472 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:05:54.029897 132949805041472 metric_logger.py:185] completed step: 9, seconds: 0.058, TFLOP/s/device: 58.640, Tokens/s/device: 8838.863, total_weights: 16384, loss: 5.890 Per train step: Total TFLOPs: 3.40 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 17:06:04 UTC 2026 EXIT_CODE=0