XPK Start: Wed Apr 22 17:45:00 UTC 2026 2026-04-22 17:45:04.344601: 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 00:00:1776879904.356874 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776879904.360579 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776879904.372161 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879904.372178 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879904.372181 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879904.372183 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 17:45:23.505951: 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) I0422 17:45:24.036914 134943294441280 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 17:45:33,078:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 17:45:33.078248 134943294441280 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 17:45:33,080:jax._src.distributed:157: Connecting to JAX distributed service on mt-07-eval-f4fcn-slice-job-0-0.mt-07-eval-f4fcn:8482 I0422 17:45:33.080425 134943294441280 distributed.py:157] Connecting to JAX distributed service on mt-07-eval-f4fcn-slice-job-0-0.mt-07-eval-f4fcn:8482 I0422 17:45:34.291229 134943294441280 max_utils.py:284] Jax distributed system initialized! I0422 17:45:41.282382 134943294441280 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 17:45:41.282492 134943294441280 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 17:45:41.282533 134943294441280 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 I0422 17:45:41.282567 134943294441280 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 17:45:41.900656 134943294441280 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:45:42.014341 134943294441280 checkpointing.py:688] Setting up checkpoint logger... I0422 17:45:42.014460 134943294441280 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 17:45:42.014505 134943294441280 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 17:45:42.014718 134943294441280 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 0x7aba6d575460>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 17:45:44.935793 134943294441280 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 17:45:44.936037 134943294441280 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 0x7aa5e8132ae0>}, handler_registry=None I0422 17:45:44.936283 134943294441280 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aa5e8132ae0>` 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`. I0422 17:45:44.936339 134943294441280 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aa5e82330e0>` 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`. I0422 17:45:44.936378 134943294441280 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 0x7aa5e8132ae0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7aa5e8132ae0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aa5e82330e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aa5e82330e0>}). I0422 17:45:44.936703 134943294441280 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 17:45:44.936782 134943294441280 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7aa5e8194f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 17:45:46.126810 134943294441280 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints I0422 17:45:46.549242 134943294441280 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7aa5e8132960> I0422 17:45:46.549429 134943294441280 checkpointing.py:302] Checkpoint manager created! I0422 17:45:47.486645 134943294441280 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 17:45:47.486766 134943294441280 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 17:45:48.905906 134943294441280 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 17:45:48.906030 134943294441280 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 17:45:48.906087 134943294441280 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 17:45:48.906164 134943294441280 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 17:45:48.906208 134943294441280 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 17:45:48.906245 134943294441280 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 17:45:48.906283 134943294441280 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:45:48.906323 134943294441280 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 17:45:48.906361 134943294441280 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:45:48.906396 134943294441280 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:45:48.906430 134943294441280 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 17:45:48.906462 134943294441280 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 17:45:49.301190 134943294441280 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:45:49.301350 134943294441280 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:45:49.306739 134943294441280 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:45:49.306836 134943294441280 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:45:49.464946 134943294441280 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:45:49.465082 134943294441280 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:45:49.620553 134943294441280 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:45:49.620690 134943294441280 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:45:49.786152 134943294441280 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:45:49.786296 134943294441280 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:45:49.936416 134943294441280 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:45:49.936557 134943294441280 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:45:50.085785 134943294441280 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:45:50.085920 134943294441280 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:45:50.244538 134943294441280 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 17:45:50.244680 134943294441280 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:45:50.267297 134943294441280 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 17:45:50.267396 134943294441280 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 17:46:01.415944 134943294441280 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. I0422 17:46:01.418398 134943294441280 metric_logger.py:289] number parameters: 1.104 billion I0422 17:46:12.300839 134943294441280 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 17:46:12.453371 134943294441280 checkpointing.py:798] Waited 0.15251708030700684 seconds for step 0 to finish before starting checkpointing. I0422 17:46:12.455917 134943294441280 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:46:12.457729 134943294441280 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0422 17:46:12.460432 134943294441280 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints/0. I0422 17:46:13.066833 134943294441280 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 17:46:13.067988 134943294441280 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 17:46:13.068046 134943294441280 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 I0422 17:46:13.317173 134943294441280 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251482s I0422 17:46:13.317348 134943294441280 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.962 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5208611488342285 s) (per-host) I0422 17:46:13.317404 134943294441280 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.520927s (batch_requests_ready=0.262851s, total_serialization_initiated=0.258003s, others=0.000073s) I0422 17:46:13.317499 134943294441280 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.525005s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.524985) I0422 17:46:13.318458 134816330671872 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:46:13.318635 134943294441280 async_checkpointer.py:561] Finished blocking save. Time taken: 0.860833s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints/0. I0422 17:46:13.772572 134943294441280 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:46:13.772948 134815223359232 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:46:13.773133 134943294441280 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879972.4558995, 'wait_for_prev_duration_secs': 6.246566772460938e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879972.4577684, 'checkpointer_blocking_duration_secs': 0.8609743118286133, 'get_old_steps_start_time': 1776879973.318764, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776879972.4537878, 'checkpoint_manager_blocking_duration_secs': 1.319286584854126} I0422 17:46:13.773249 134943294441280 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 17:46:13.773327 134943294441280 max_utils.py:750] Memstats: After params initialized: I0422 17:46:13.773385 134943294441280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 17:46:13.773418 134943294441280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 17:46:13.773444 134943294441280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 17:46:13.773468 134943294441280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 17:46:14.043195 134943294441280 metric_logger.py:185] completed step: 0, seconds: 10.881, TFLOP/s/device: 1.249, Tokens/s/device: 188.211, total_weights: 65536, loss: 10.874 I0422 17:46:14.210958 134943294441280 metric_logger.py:185] completed step: 1, seconds: 1.741, TFLOP/s/device: 7.806, Tokens/s/device: 1176.568, total_weights: 65536, loss: 10.874 I0422 17:46:14.633805 134943294441280 metric_logger.py:185] completed step: 2, seconds: 0.022, TFLOP/s/device: 620.332, Tokens/s/device: 93503.173, total_weights: 65536, loss: 10.262 I0422 17:46:14.781247 134943294441280 metric_logger.py:185] completed step: 3, seconds: 0.413, TFLOP/s/device: 32.900, Tokens/s/device: 4959.030, total_weights: 65536, loss: 9.732 I0422 17:46:15.076764 134943294441280 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.942, Tokens/s/device: 12501.984, total_weights: 65536, loss: 9.274 I0422 17:46:15.084141 134943294441280 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.353, Tokens/s/device: 13920.514, total_weights: 65536, loss: 8.888 I0422 17:46:33.994234 2651 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 17:46:34.165362 134943294441280 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.923, Tokens/s/device: 6922.076, total_weights: 65536, loss: 8.589 I0422 17:46:34.312682 134943294441280 metric_logger.py:185] completed step: 7, seconds: 18.929, TFLOP/s/device: 0.718, Tokens/s/device: 108.194, total_weights: 65536, loss: 8.380 I0422 17:46:34.460372 134943294441280 metric_logger.py:185] completed step: 8, seconds: 0.160, TFLOP/s/device: 85.048, Tokens/s/device: 12819.309, total_weights: 65536, loss: 8.251 I0422 17:46:34.608579 134943294441280 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 17:46:34.609213 134943294441280 checkpointing.py:798] Waited 0.0006444454193115234 seconds for step 9 to finish before starting checkpointing. I0422 17:46:34.611210 134943294441280 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:46:36.354990 134816339064576 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints/0/items/array_metadatas/process_6 I0422 17:47:07.029855 134816330671872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 29.126 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 54.23332643508911 s) (per-host) I0422 17:47:07.029978 134816330671872 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 53.711410s. I0422 17:47:15.835052 134816330671872 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 62.516468s. I0422 17:47:15.835407 134815223359232 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:47:15.835550 134815223359232 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:47:15.835602 134815223359232 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:47:15.836984 134815223359232 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 17:47:15.837176 134943294441280 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. W0422 17:47:15.837316 134943294441280 checkpoint_manager.py:1458] Waiting for previous save to complete took 41.226111 seconds. If this number is high, consider checkpointing less frequently. I0422 17:47:15.839255 134943294441280 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0422 17:47:15.842748 134943294441280 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints/9. I0422 17:47:16.397138 134943294441280 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 17:47:16.397301 134943294441280 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 I0422 17:47:16.446546 134943294441280 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051518s I0422 17:47:16.446716 134943294441280 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.062 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30470728874206543 s) (per-host) I0422 17:47:16.446766 134943294441280 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304767s (batch_requests_ready=0.248492s, total_serialization_initiated=0.056207s, others=0.000068s) I0422 17:47:16.446887 134943294441280 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308813s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.308798) I0422 17:47:16.447808 134816280794880 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:47:16.447967 134943294441280 async_checkpointer.py:561] Finished blocking save. Time taken: 0.608641s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints/9. I0422 17:47:16.875118 134943294441280 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:47:16.875491 134815223359232 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:47:16.875664 134943294441280 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879994.6111808, 'wait_for_prev_duration_secs': 41.22611141204834, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776880035.839294, 'checkpointer_blocking_duration_secs': 0.6087779998779297, 'get_old_steps_start_time': 1776880036.4480903, 'get_old_steps_duration_secs': 4.00543212890625e-05, 'checkpoint_manager_blocking_start_time': 1776879994.6094346, 'checkpoint_manager_blocking_duration_secs': 42.266191720962524} I0422 17:47:16.875777 134943294441280 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 17:47:16.875822 134943294441280 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:47:21.955088 134816339064576 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_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_07_eval/checkpoints/9/items/array_metadatas/process_6 I0422 17:47:57.837709 134816280794880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.884 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.6956582069397 s) (per-host) I0422 17:47:57.837829 134816280794880 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.389910s. I0422 17:48:06.693598 134816280794880 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.245663s. I0422 17:48:06.693874 134815223359232 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:48:06.693990 134815223359232 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:48:06.694038 134815223359232 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:48:06.697219 134815223359232 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 17:48:06.697381 134943294441280 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. I0422 17:48:06.697545 134943294441280 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:48:06.698260 134943294441280 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.440, Tokens/s/device: 13933.584, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Wed Apr 22 17:48:15 UTC 2026 EXIT_CODE=0