2026-04-20 16:09:04.765276: 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) I0420 16:09:04.886421 135152898849920 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0420 16:09:58.814142 135152898849920 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 16:09:58.814259 135152898849920 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 16:09:58.814294 135152898849920 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Apr 6 2026 20:48:10 (1775533690) cl/895581894 I0420 16:09:58.814320 135152898849920 train_utils.py:365] WARNING: 'dataset_path' might be pointing your local file system I0420 16:09:58.814410 135152898849920 train.py:806] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0420 16:09:59.320364 135152898849920 maxtext_utils.py:1718] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:09:59.795351 135152898849920 checkpointing.py:688] Setting up checkpoint logger... I0420 16:09:59.795454 135152898849920 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 16:09:59.795493 135152898849920 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0420 16:09:59.795678 135152898849920 base_pytree_checkpoint_handler.py:411] 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 0x7aeb259f8e60>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 16:10:02.316385 135152898849920 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 16:10:02.316585 135152898849920 checkpoint_manager.py:702] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae4a8f923c0>}, handler_registry=None I0420 16:10:02.316800 135152898849920 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae4a8f923c0>` 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`. I0420 16:10:02.316850 135152898849920 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae4a8f938f0>` 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`. I0420 16:10:02.316880 135152898849920 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 0x7ae4a8f923c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae4a8f923c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae4a8f938f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae4a8f938f0>}). I0420 16:10:02.317188 135152898849920 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0420 16:10:02.317238 135152898849920 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7ae4a9b4d6c0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 16:10:02.447964 135152898849920 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints I0420 16:10:02.448199 135152898849920 checkpoint_manager.py:921] [process=0][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_hns=False, 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), root_directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ae4a8f937a0> I0420 16:10:02.448276 135152898849920 checkpointing.py:302] Checkpoint manager created! I0420 16:10:02.524687 135152898849920 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0420 16:10:02.528035 135152898849920 reader.py:262] Creating a tf.data.Dataset reading 8 files located in folders: tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0. I0420 16:10:02.585655 135152898849920 logging_logger.py:49] Constructing tf.data.Dataset __local_c4_builder for split train, from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0420 16:10:02.609733 135152898849920 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0420 16:10:02.609799 135152898849920 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0420 16:10:03.261422 135152898849920 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 16:10:03.261530 135152898849920 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. W0420 16:10:03.373150 1157395 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. [DECOUPLED NO-OP] gcs_storage: using stubs. [DECOUPLED NO-OP] mldiagnostics: using stub. [DECOUPLED NO-OP] mldiagnostics: using stub. [DECOUPLED NO-OP] mldiagnostics: using stub. [DECOUPLED NO-OP] workload_monitor: using stub. [DECOUPLED NO-OP] vertex_tensorboard: using stub. fsdp: 8 I0420 16:10:03.581903 135152898849920 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 16:10:03.581993 135152898849920 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0420 16:10:03.587343 135152898849920 nnx_decoders.py:465] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 16:10:03.587393 135152898849920 nnx_decoders.py:465] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0420 16:10:03.602510 135152898849920 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:10:03.602557 135152898849920 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0420 16:10:03.617087 135152898849920 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:10:03.617136 135152898849920 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0420 16:10:03.639195 135152898849920 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 16:10:03.639249 135152898849920 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0420 16:10:03.653848 135152898849920 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 16:10:03.653900 135152898849920 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0420 16:10:03.668475 135152898849920 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 16:10:03.668524 135152898849920 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0420 16:10:03.696507 135152898849920 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 16:10:03.696567 135152898849920 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0420 16:10:03.714693 135152898849920 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 16:10:03.714751 135152898849920 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). W0420 16:10:04.100621 1157395 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0420 16:10:04.216469 135152898849920 max_utils.py:791] Total memory size: 3.6 GB, Output size: 1.5 GB, Temp size: 2.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB. I0420 16:10:04.217020 135152898849920 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0420 16:10:04.218373 135152898849920 metric_logger.py:301] number parameters: 1.104 billion W0420 16:10:07.076785 1157395 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0420 16:10:07.217070 135152898849920 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 16:10:07.336652 135152898849920 checkpointing.py:798] Waited 0.1195826530456543 seconds for step 0 to finish before starting checkpointing. I0420 16:10:07.337106 135152898849920 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:10:07.337257 135152898849920 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0420 16:10:07.337641 135152898849920 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0. I0420 16:10:07.421940 135152898849920 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0420 16:10:07.510272 135039742379584 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0 I0420 16:10:07.527803 135152898849920 jax_array_handlers.py:347] Scheduling D2H of 69 prioritized jax.Array. I0420 16:10:07.527898 135152898849920 replica_slices.py:410] 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 I0420 16:10:08.545967 135039731893824 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/items I0420 16:10:08.701613 135039683659328 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776701408454305519, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776701408454305519, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0420 16:10:09.106296 1160063 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0420 16:10:09.106360 1160063 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0420 16:10:09.720205 135152898849920 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.193429s I0420 16:10:09.720780 135152898849920 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.372 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0420 16:10:09.727281 135152898849920 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.303810s (batch_requests_ready=0.095631s, total_serialization_initiated=2.201338s, others=0.006842s) I0420 16:10:09.727399 135152898849920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.304964s (all_items=0.000021s, per_item={'items': '0.00002050'}, temp_paths=2.304944) I0420 16:10:09.729173 135039610259008 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 16:10:09.729333 135152898849920 async_checkpointer.py:561] Finished blocking save. Time taken: 2.392031s. Continuing background save to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0. I0420 16:10:09.729667 135152898849920 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:10:09.729891 135039752865344 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:10:09.730031 135152898849920 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776701407.3370936, 'wait_for_prev_duration_secs': 4.172325134277344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776701407.337278, 'checkpointer_blocking_duration_secs': 2.3922903537750244, 'get_old_steps_start_time': 1776701409.729595, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776701407.336996, 'checkpoint_manager_blocking_duration_secs': 2.392998695373535} I0420 16:10:09.730193 135152898849920 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 16:10:09.730260 135152898849920 max_utils.py:750] Memstats: After params initialized: I0420 16:10:09.730311 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0420 16:10:09.730336 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0420 16:10:09.730355 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0420 16:10:09.730375 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0420 16:10:09.730394 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0420 16:10:09.730412 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0420 16:10:09.730429 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0420 16:10:09.730446 135152898849920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0420 16:10:10.136059 135152898849920 metric_logger.py:196] completed step: 0, seconds: 2.997, TFLOP/s/device: 4.533, Tokens/s/device: 683.251, total_weights: 13328, loss: 10.880, lm_loss: 10.880, perplexity: 53108.402 I0420 16:10:10.137008 135152898849920 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/tensorboard/' I0420 16:10:10.248837 135152898849920 metric_logger.py:196] completed step: 1, seconds: 2.915, TFLOP/s/device: 4.661, Tokens/s/device: 702.546, total_weights: 12332, loss: 10.862, lm_loss: 10.862, perplexity: 52152.828 I0420 16:10:10.366096 135152898849920 metric_logger.py:196] completed step: 2, seconds: 0.022, TFLOP/s/device: 615.610, Tokens/s/device: 92791.446, total_weights: 15161, loss: 9.926, lm_loss: 9.926, perplexity: 20464.070 I0420 16:10:10.483484 135152898849920 metric_logger.py:196] completed step: 3, seconds: 0.105, TFLOP/s/device: 129.701, Tokens/s/device: 19550.006, total_weights: 13327, loss: 9.396, lm_loss: 9.396, perplexity: 12037.986 I0420 16:10:10.484301 135039710922304 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/items/array_metadatas/process_0 I0420 16:10:30.587540 135152898849920 metric_logger.py:196] completed step: 4, seconds: 0.117, TFLOP/s/device: 115.827, Tokens/s/device: 17458.761, total_weights: 11939, loss: 9.005, lm_loss: 9.005, perplexity: 8141.998 I0420 16:10:30.599173 135152898849920 metric_logger.py:196] completed step: 5, seconds: 0.118, TFLOP/s/device: 115.253, Tokens/s/device: 17372.127, total_weights: 15502, loss: 8.861, lm_loss: 8.861, perplexity: 7054.059 I0420 16:10:30.714074 135152898849920 metric_logger.py:196] completed step: 6, seconds: 20.103, TFLOP/s/device: 0.676, Tokens/s/device: 101.875, total_weights: 13864, loss: 8.738, lm_loss: 8.738, perplexity: 6233.198 I0420 16:10:30.831035 135152898849920 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1664.477, Tokens/s/device: 250888.154, total_weights: 12988, loss: 8.651, lm_loss: 8.651, perplexity: 5718.487 I0420 16:10:30.948406 135152898849920 metric_logger.py:196] completed step: 8, seconds: 0.116, TFLOP/s/device: 116.652, Tokens/s/device: 17583.021, total_weights: 13820, loss: 8.671, lm_loss: 8.671, perplexity: 5832.209 I0420 16:10:31.066345 135152898849920 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 16:10:31.067541 135152898849920 checkpointing.py:798] Waited 0.0012102127075195312 seconds for step 9 to finish before starting checkpointing. I0420 16:10:31.068398 135152898849920 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:10:40.823813 135039631230528 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 31.095744s (commit=30.629090s, array_metadata_write=0.466654s) I0420 16:10:40.825033 135039610259008 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 378.330 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 33 seconds) (per-host) I0420 16:10:40.825158 135039610259008 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 31.095767s. I0420 16:10:41.055174 135039610259008 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776701408454305519, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0420 16:10:41.233860 135039610259008 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0420 16:10:41.438424 135039683659328 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776701408454305519, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0420 16:10:41.650262 135039610259008 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0420 16:10:41.650970 135039610259008 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.554985s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/items I0420 16:10:41.651670 135039610259008 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0/items I0420 16:10:41.897426 135039610259008 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0 I0420 16:10:42.565692 135039610259008 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0`. I0420 16:10:42.566375 135039610259008 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 35.229085s. directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/0 I0420 16:10:42.566459 135039610259008 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 32.837071s. I0420 16:10:42.566675 135039752865344 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:10:42.566789 135039752865344 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:10:42.566850 135039752865344 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:10:42.566896 135039752865344 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 16:10:42.567149 135152898849920 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 16:10:42.567279 135152898849920 checkpoint_manager.py:1441] Waiting for previous save to complete took 11.498887 seconds. If this number is high, consider checkpointing less frequently. I0420 16:10:42.568110 135152898849920 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0420 16:10:42.568401 135152898849920 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9. I0420 16:10:42.759995 135152898849920 jax_array_handlers.py:347] Scheduling D2H of 69 prioritized jax.Array. I0420 16:10:42.760163 135152898849920 replica_slices.py:410] 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 I0420 16:10:42.766516 135039752865344 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9 I0420 16:10:43.500901 135039599773248 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9/items I0420 16:10:44.876844 135152898849920 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.117866s I0420 16:10:44.883105 135152898849920 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.543 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0420 16:10:44.883169 135152898849920 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.226282s (batch_requests_ready=0.094279s, total_serialization_initiated=2.125859s, others=0.006145s) I0420 16:10:44.883244 135152898849920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.227072s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=2.227055) I0420 16:10:44.884623 135039662687808 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 16:10:44.884724 135152898849920 async_checkpointer.py:561] Finished blocking save. Time taken: 2.316566s. Continuing background save to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9. I0420 16:10:44.884927 135152898849920 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:10:44.885156 135039641716288 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:10:44.885410 135152898849920 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776701431.0683634, 'wait_for_prev_duration_secs': 11.498887062072754, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776701442.5681376, 'checkpointer_blocking_duration_secs': 2.316697835922241, 'get_old_steps_start_time': 1776701444.8848534, 'get_old_steps_duration_secs': 3.504753112792969e-05, 'checkpoint_manager_blocking_start_time': 1776701431.0683174, 'checkpoint_manager_blocking_duration_secs': 13.817065000534058} I0420 16:10:44.885507 135152898849920 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 16:10:44.885541 135152898849920 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:10:45.650661 135039710922304 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9/items/array_metadatas/process_0 I0420 16:11:19.625883 135039752865344 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 34.742126s (commit=34.313383s, array_metadata_write=0.428742s) I0420 16:11:19.627240 135039662687808 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 341.809 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 36 seconds) (per-host) I0420 16:11:19.627364 135039662687808 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.742594s. I0420 16:11:20.056123 135039662687808 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0420 16:11:20.498334 135039662687808 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0420 16:11:20.498522 135039662687808 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.586101s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9/items I0420 16:11:20.499250 135039662687808 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9/items I0420 16:11:20.742716 135039662687808 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9 I0420 16:11:21.413116 135039662687808 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9`. I0420 16:11:21.413777 135039662687808 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 38.845623s. directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_trainstate_and_training_loop_20260420_142345/nnx_feat_nnx_trainstate_and_training_loop_20260420_142345_10_shardy_true/checkpoints/9 I0420 16:11:21.413858 135039662687808 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.529091s. I0420 16:11:21.414001 135039641716288 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:11:21.414065 135039641716288 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:11:21.414134 135039641716288 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:11:21.414178 135039641716288 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 16:11:21.414748 135152898849920 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 16:11:21.414947 135152898849920 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:11:21.416002 135152898849920 metric_logger.py:196] completed step: 9, seconds: 0.116, TFLOP/s/device: 116.718, Tokens/s/device: 17592.990, total_weights: 12300, loss: 8.668, lm_loss: 8.668, perplexity: 5810.989 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops