XPK Start: Sat Apr 25 16:41:19 UTC 2026 2026-04-25 16:41:23.080195: 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 00:00:1777135283.093201 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777135283.096983 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777135283.108322 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777135283.108341 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777135283.108343 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777135283.108345 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:41:42.310127: 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 16:41:42.829933 136715696326464 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:41:51,871:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:41:51.871419 136715696326464 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:41:51,873:jax._src.distributed:157: Connecting to JAX distributed service on mt-04-int8-we6og-slice-job-0-0.mt-04-int8-we6og:8482 I0425 16:41:51.873866 136715696326464 distributed.py:157] Connecting to JAX distributed service on mt-04-int8-we6og-slice-job-0-0.mt-04-int8-we6og:8482 I0425 16:41:52.956869 136715696326464 max_utils.py:284] Jax distributed system initialized! I0425 16:41:59.719778 136715696326464 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:41:59.719886 136715696326464 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:41:59.719926 136715696326464 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 16:41:59.719959 136715696326464 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:42:00.343605 136715696326464 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:42:00.900242 136715696326464 checkpointing.py:688] Setting up checkpoint logger... I0425 16:42:00.900359 136715696326464 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 16:42:00.900406 136715696326464 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:42:00.900629 136715696326464 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 0x7c5713b75280>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:42:03.718217 136715696326464 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:42:03.718566 136715696326464 checkpoint_manager.py:709] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c43ba564680>}, handler_registry=None I0425 16:42:03.718816 136715696326464 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c43ba564680>` 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 16:42:03.718865 136715696326464 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c43ba5cc2f0>` 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 16:42:03.718903 136715696326464 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 0x7c43ba564680>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c43ba564680>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c43ba5cc2f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c43ba5cc2f0>}). I0425 16:42:03.719260 136715696326464 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:42:03.719338 136715696326464 async_checkpointer.py:177] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c43bbe65bc0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:42:04.887594 136715696326464 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_04_int8/checkpoints I0425 16:42:05.353770 136715696326464 checkpoint_manager.py:929] [process=5][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_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c43ba5ced80> I0425 16:42:05.353950 136715696326464 checkpointing.py:302] Checkpoint manager created! I0425 16:42:06.806472 136715696326464 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:42:06.806595 136715696326464 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 16:42:09.148726 136715696326464 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 16:42:09.148848 136715696326464 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 16:42:09.148903 136715696326464 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 16:42:09.148949 136715696326464 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 16:42:09.148987 136715696326464 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 16:42:09.149028 136715696326464 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 16:42:09.149067 136715696326464 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:42:09.149107 136715696326464 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 16:42:09.149144 136715696326464 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:42:09.149179 136715696326464 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:42:09.149213 136715696326464 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 16:42:09.149248 136715696326464 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 16:42:09.555339 136715696326464 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:42:09.555486 136715696326464 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:42:09.561005 136715696326464 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:42:09.561099 136715696326464 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:42:09.706584 136715696326464 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:42:09.706730 136715696326464 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:42:09.845208 136715696326464 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:42:09.845335 136715696326464 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:42:10.035091 136715696326464 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 16:42:10.035224 136715696326464 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:42:10.172369 136715696326464 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 16:42:10.172499 136715696326464 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:42:10.310204 136715696326464 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 16:42:10.310335 136715696326464 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:42:10.459519 136715696326464 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:42:10.459679 136715696326464 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:42:10.947076 136715696326464 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:42:10.947197 136715696326464 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 16:42:24.754911 136715696326464 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. I0425 16:42:24.759863 136715696326464 metric_logger.py:289] number parameters: 1.104 billion I0425 16:42:39.046845 136715696326464 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 16:42:39.282057 136715696326464 checkpointing.py:798] Waited 0.2351982593536377 seconds for step 0 to finish before starting checkpointing. I0425 16:42:39.284806 136715696326464 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:42:39.286724 136715696326464 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 0 I0425 16:42:39.289514 136715696326464 async_checkpointer.py:452] [process=5] 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_04_int8/checkpoints/0. I0425 16:42:39.853801 136715696326464 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:42:39.854721 136715696326464 jax_array_handlers.py:358] Scheduling D2H of 111 prioritized jax.Array. I0425 16:42:39.854779 136715696326464 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 16:42:40.104618 136715696326464 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.253278s I0425 16:42:40.104808 136715696326464 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.979 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5178966522216797 s) (per-host) I0425 16:42:40.104860 136715696326464 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.517959s (batch_requests_ready=0.254550s, total_serialization_initiated=0.263338s, others=0.000070s) I0425 16:42:40.104980 136715696326464 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.522073s (all_items=0.000021s, per_item={'items': '0.00002122'}, temp_paths=0.522052) I0425 16:42:40.105899 136585148667648 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0425 16:42:40.106075 136715696326464 async_checkpointer.py:561] Finished blocking save. Time taken: 0.819283s. 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_04_int8/checkpoints/0. I0425 16:42:40.489346 136715696326464 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:42:40.489741 136581964883712 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:42:40.489907 136715696326464 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_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777135359.2847881, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777135359.2867646, 'checkpointer_blocking_duration_secs': 0.8194162845611572, 'get_old_steps_start_time': 1777135360.1061995, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1777135359.2825208, 'checkpoint_manager_blocking_duration_secs': 1.2073462009429932} I0425 16:42:40.490028 136715696326464 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 16:42:40.490114 136715696326464 max_utils.py:750] Memstats: After params initialized: I0425 16:42:40.490175 136715696326464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0425 16:42:40.490211 136715696326464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0425 16:42:40.490242 136715696326464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0425 16:42:40.490268 136715696326464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0425 16:42:40.764221 136715696326464 metric_logger.py:185] completed step: 0, seconds: 14.285, TFLOP/s/device: 0.951, Tokens/s/device: 143.366, total_weights: 65536, loss: 10.872 I0425 16:42:41.010089 136715696326464 metric_logger.py:185] completed step: 1, seconds: 1.715, TFLOP/s/device: 7.923, Tokens/s/device: 1194.252, total_weights: 65536, loss: 10.872 I0425 16:42:41.449105 136715696326464 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 647.808, Tokens/s/device: 97644.703, total_weights: 65536, loss: 10.258 I0425 16:42:41.678244 136715696326464 metric_logger.py:185] completed step: 3, seconds: 0.420, TFLOP/s/device: 32.327, Tokens/s/device: 4872.745, total_weights: 65536, loss: 9.727 I0425 16:42:42.137484 136715696326464 metric_logger.py:185] completed step: 4, seconds: 0.253, TFLOP/s/device: 53.664, Tokens/s/device: 8088.819, total_weights: 65536, loss: 9.270 I0425 16:42:42.146121 136715696326464 metric_logger.py:185] completed step: 5, seconds: 0.229, TFLOP/s/device: 59.312, Tokens/s/device: 8940.186, total_weights: 65536, loss: 8.884 I0425 16:43:03.588118 2762 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:43:03.850037 136715696326464 metric_logger.py:185] completed step: 6, seconds: 0.459, TFLOP/s/device: 29.573, Tokens/s/device: 4457.571, total_weights: 65536, loss: 8.584 I0425 16:43:04.079124 136715696326464 metric_logger.py:185] completed step: 7, seconds: 21.466, TFLOP/s/device: 0.633, Tokens/s/device: 95.405, total_weights: 65536, loss: 8.376 I0425 16:43:04.308220 136715696326464 metric_logger.py:185] completed step: 8, seconds: 0.247, TFLOP/s/device: 55.090, Tokens/s/device: 8303.735, total_weights: 65536, loss: 8.247 I0425 16:43:04.539155 136715696326464 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 16:43:04.540075 136715696326464 checkpointing.py:798] Waited 0.0009310245513916016 seconds for step 9 to finish before starting checkpointing. I0425 16:43:04.542103 136715696326464 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:43:06.073036 136585676830464 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 111 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_04_int8/checkpoints/0/items/array_metadatas/process_5 I0425 16:43:33.697584 136585148667648 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 29.192 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 54.110644578933716 s) (per-host) I0425 16:43:33.697721 136585148667648 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 53.591701s. I0425 16:43:42.222867 136585148667648 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 62.116832s. I0425 16:43:42.223161 136581964883712 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:43:42.223287 136581964883712 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:43:42.223339 136581964883712 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:43:42.224721 136581964883712 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 16:43:42.224898 136715696326464 checkpoint_manager.py:2036] [process=5][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0425 16:43:42.225034 136715696326464 checkpoint_manager.py:1458] Waiting for previous save to complete took 37.682938 seconds. If this number is high, consider checkpointing less frequently. I0425 16:43:42.227144 136715696326464 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 9 I0425 16:43:42.230508 136715696326464 async_checkpointer.py:452] [process=5] 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_04_int8/checkpoints/9. I0425 16:43:42.793141 136715696326464 jax_array_handlers.py:358] Scheduling D2H of 111 prioritized jax.Array. I0425 16:43:42.793244 136715696326464 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 16:43:42.835413 136715696326464 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045621s I0425 16:43:42.835558 136715696326464 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.911 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3141162395477295 s) (per-host) I0425 16:43:42.835607 136715696326464 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314173s (batch_requests_ready=0.260431s, total_serialization_initiated=0.053679s, others=0.000063s) I0425 16:43:42.835745 136715696326464 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318271s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.318255) I0425 16:43:42.836627 136585697801984 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0425 16:43:42.836799 136715696326464 async_checkpointer.py:561] Finished blocking save. Time taken: 0.609583s. 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_04_int8/checkpoints/9. I0425 16:43:42.845407 136715696326464 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:43:42.845668 136581964883712 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:43:42.845830 136715696326464 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_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777135384.5420735, 'wait_for_prev_duration_secs': 37.68293833732605, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777135422.227183, 'checkpointer_blocking_duration_secs': 0.6097245216369629, 'get_old_steps_start_time': 1777135422.8369255, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777135384.5403602, 'checkpoint_manager_blocking_duration_secs': 38.30543565750122} I0425 16:43:42.845956 136715696326464 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 16:43:42.845999 136715696326464 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:43:47.737856 136585706194688 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 111 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_04_int8/checkpoints/9/items/array_metadatas/process_5 I0425 16:44:24.658236 136585697801984 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 37.487 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.13675379753113 s) (per-host) I0425 16:44:24.658366 136585697801984 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.821615s. I0425 16:44:31.731893 136585697801984 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 48.895127s. I0425 16:44:31.732195 136581964883712 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:44:31.732331 136581964883712 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:44:31.732380 136581964883712 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:44:31.733783 136581964883712 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 16:44:31.733961 136715696326464 checkpoint_manager.py:2036] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0425 16:44:31.734138 136715696326464 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:44:31.734801 136715696326464 metric_logger.py:185] completed step: 9, seconds: 0.229, TFLOP/s/device: 59.447, Tokens/s/device: 8960.487, 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: Sat Apr 25 16:44:44 UTC 2026 EXIT_CODE=0