XPK Start: Thu Apr 16 22:45:55 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-16 22:46:19.305168: 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) I0416 22:46:19.481642 138543412606784 max_utils.py:273] Attempting to initialize the jax distributed system... I0416 22:46:28.522697 138543412606784 distributed.py:149] Starting JAX distributed service on [::]:8482 I0416 22:46:28.525084 138543412606784 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-true-d9qh9-slice-job-0-0.mt-10-shardy-true-d9qh9:8482 I0416 22:46:30.396930 138543412606784 max_utils.py:284] Jax distributed system initialized! I0416 22:46:36.427772 138543412606784 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 22:46:36.427870 138543412606784 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 22:46:36.427910 138543412606784 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0416 22:46:36.427944 138543412606784 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 22:46:37.465821 138543412606784 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 22:46:37.579724 138543412606784 checkpointing.py:688] Setting up checkpoint logger... I0416 22:46:37.579831 138543412606784 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 22:46:37.579877 138543412606784 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0416 22:46:37.580096 138543412606784 base_pytree_checkpoint_handler.py:441] 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 0x7e0092f1c260>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 22:46:40.133327 138543412606784 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0416 22:46:40.133576 138543412606784 checkpoint_manager.py:708] [process=1][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dff634f5c70>}, handler_registry=None I0416 22:46:40.133819 138543412606784 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dff634f5c70>` 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`. I0416 22:46:40.133867 138543412606784 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dff635e7470>` 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`. I0416 22:46:40.133915 138543412606784 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 0x7dff634f5c70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dff634f5c70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dff635e7470>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dff635e7470>}). I0416 22:46:40.134294 138543412606784 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0416 22:46:40.134390 138543412606784 async_checkpointer.py:192] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7deb40186340> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0416 22:46:42.101182 138543412606784 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints I0416 22:46:42.666862 138543412606784 checkpoint_manager.py:929] [process=1][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_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7dff63648860> I0416 22:46:42.667034 138543412606784 checkpointing.py:302] Checkpoint manager created! I0416 22:46:43.650137 138543412606784 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0416 22:46:43.650253 138543412606784 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0416 22:46:45.030233 138543412606784 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0416 22:46:45.030352 138543412606784 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0416 22:46:45.030429 138543412606784 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0416 22:46:45.030508 138543412606784 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0416 22:46:45.030567 138543412606784 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0416 22:46:45.030620 138543412606784 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0416 22:46:45.030685 138543412606784 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 22:46:45.030741 138543412606784 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0416 22:46:45.030800 138543412606784 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 22:46:45.030856 138543412606784 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 22:46:45.030917 138543412606784 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0416 22:46:45.030967 138543412606784 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0416 22:46:45.430618 138543412606784 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 22:46:45.430705 138543412606784 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:46:45.436049 138543412606784 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 22:46:45.436104 138543412606784 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:46:45.452177 138543412606784 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 22:46:45.452233 138543412606784 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:46:45.467703 138543412606784 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 22:46:45.467759 138543412606784 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:46:45.491500 138543412606784 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 22:46:45.491564 138543412606784 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:46:45.507173 138543412606784 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 22:46:45.507229 138543412606784 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:46:45.522812 138543412606784 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 22:46:45.522878 138543412606784 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:46:45.552541 138543412606784 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 22:46:45.552608 138543412606784 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:46:45.572104 138543412606784 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 22:46:45.572167 138543412606784 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0416 22:46:56.813372 138543412606784 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. I0416 22:46:56.815734 138543412606784 metric_logger.py:289] number parameters: 1.104 billion I0416 22:47:08.155912 138543412606784 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0416 22:47:08.306901 138543412606784 checkpointing.py:798] Waited 0.15097498893737793 seconds for step 0 to finish before starting checkpointing. I0416 22:47:08.309148 138543412606784 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 22:47:08.310824 138543412606784 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 0 I0416 22:47:08.312201 138543412606784 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints/0. I0416 22:47:09.056274 138543412606784 signaling_client.py:364] Using JaxDistributedSignalingClient I0416 22:47:09.057165 138543412606784 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0416 22:47:09.057223 138543412606784 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 I0416 22:47:09.339392 138543412606784 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.283874s I0416 22:47:09.339593 138543412606784 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 5.263 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2930903434753418 s) (per-host) I0416 22:47:09.339647 138543412606784 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.293157s (batch_requests_ready=0.003161s, total_serialization_initiated=0.289921s, others=0.000075s) I0416 22:47:09.339743 138543412606784 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.297134s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.297117) I0416 22:47:09.340447 138543412606784 event_tracking.py:125] [process=1] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints/0. I0416 22:47:09.341900 138412920350464 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-16 23:07:09.341858 I0416 22:47:09.343870 138543412606784 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 22:47:09.344182 138414566856448 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0416 22:47:09.344306 138543412606784 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776379628.309129, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776379628.3108633, 'checkpointer_blocking_duration_secs': 1.0311951637268066, 'get_old_steps_start_time': 1776379629.3420825, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776379628.3073075, 'checkpoint_manager_blocking_duration_secs': 1.0369617938995361} I0416 22:47:09.344408 138543412606784 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0416 22:47:09.344482 138543412606784 max_utils.py:750] Memstats: After params initialized: I0416 22:47:09.344530 138543412606784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_2(process=1,(2,0,0,0)) I0416 22:47:09.344560 138543412606784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_3(process=1,(3,0,0,0)) I0416 22:47:09.344587 138543412606784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_6(process=1,(2,1,0,0)) I0416 22:47:09.344610 138543412606784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_7(process=1,(3,1,0,0)) I0416 22:47:09.669804 138543412606784 metric_logger.py:185] completed step: 0, seconds: 11.339, TFLOP/s/device: 1.198, Tokens/s/device: 180.614, total_weights: 65536, loss: 10.874 I0416 22:47:09.829594 138543412606784 metric_logger.py:185] completed step: 1, seconds: 1.512, TFLOP/s/device: 8.985, Tokens/s/device: 1354.350, total_weights: 65536, loss: 10.874 I0416 22:47:10.691219 138543412606784 metric_logger.py:185] completed step: 2, seconds: 0.014, TFLOP/s/device: 992.993, Tokens/s/device: 149674.779, total_weights: 65536, loss: 10.262 I0416 22:47:10.838654 138543412606784 metric_logger.py:185] completed step: 3, seconds: 0.852, TFLOP/s/device: 15.950, Tokens/s/device: 2404.176, total_weights: 65536, loss: 9.732 I0416 22:47:11.134388 138543412606784 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.777, Tokens/s/device: 12477.002, total_weights: 65536, loss: 9.274 I0416 22:47:11.142126 138543412606784 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.536, Tokens/s/device: 13948.008, total_weights: 65536, loss: 8.888 I0416 22:47:14.088337 2771 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0416 22:47:16.451256 138412971521792 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints/0/items/array_metadatas/process_1 I0416 22:47:34.820943 138543412606784 metric_logger.py:185] completed step: 6, seconds: 0.297, TFLOP/s/device: 45.823, Tokens/s/device: 6906.902, total_weights: 65536, loss: 8.589 I0416 22:47:34.968639 138543412606784 metric_logger.py:185] completed step: 7, seconds: 23.533, TFLOP/s/device: 0.577, Tokens/s/device: 87.027, total_weights: 65536, loss: 8.380 I0416 22:47:35.115889 138543412606784 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.804, Tokens/s/device: 13385.533, total_weights: 65536, loss: 8.251 I0416 22:47:35.264411 138543412606784 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0416 22:47:35.265106 138543412606784 checkpointing.py:798] Waited 0.000705718994140625 seconds for step 9 to finish before starting checkpointing. I0416 22:47:35.267159 138543412606784 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 22:47:46.434796 138412920350464 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 42.248 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.38825488090515 s) (per-host) I0416 22:47:46.434921 138412920350464 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.092904s. I0416 22:47:55.996256 138412920350464 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 46.654221s. I0416 22:47:55.996486 138414566856448 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 22:47:55.996544 138414566856448 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0416 22:47:55.996593 138414566856448 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 22:47:55.998466 138414566856448 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 22:47:55.998640 138543412606784 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0416 22:47:55.998775 138543412606784 checkpoint_manager.py:1452] Waiting for previous save to complete took 20.731614 seconds. If this number is high, consider checkpointing less frequently. I0416 22:47:56.000751 138543412606784 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 9 I0416 22:47:56.002825 138543412606784 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints/9. I0416 22:47:56.719084 138543412606784 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0416 22:47:56.719179 138543412606784 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 I0416 22:47:56.753946 138543412606784 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036465s I0416 22:47:56.754122 138543412606784 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 35.283 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04372000694274902 s) (per-host) I0416 22:47:56.754174 138543412606784 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.043781s (batch_requests_ready=0.002859s, total_serialization_initiated=0.040853s, others=0.000069s) I0416 22:47:56.754271 138543412606784 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048076s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.048060) I0416 22:47:56.754999 138543412606784 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints/9. I0416 22:47:56.755288 138412971521792 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-16 23:07:56.755254 I0416 22:47:56.759108 138543412606784 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 22:47:56.759396 138414566856448 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0416 22:47:56.759526 138543412606784 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776379655.2671301, 'wait_for_prev_duration_secs': 20.731614112854004, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776379676.0007906, 'checkpointer_blocking_duration_secs': 0.7546331882476807, 'get_old_steps_start_time': 1776379676.7554488, 'get_old_steps_duration_secs': 3.7670135498046875e-05, 'checkpoint_manager_blocking_start_time': 1776379655.2653372, 'checkpoint_manager_blocking_duration_secs': 21.494155645370483} I0416 22:47:56.759639 138543412606784 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0416 22:47:56.759684 138543412606784 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 22:48:03.053349 138392515872512 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_true/checkpoints/9/items/array_metadatas/process_1 I0416 22:48:38.540203 138412971521792 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 37.763 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.82976198196411 s) (per-host) I0416 22:48:38.540335 138412971521792 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.784936s. I0416 22:48:47.226743 138412971521792 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 50.471330s. I0416 22:48:47.226975 138414566856448 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 22:48:47.227037 138414566856448 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0416 22:48:47.227085 138414566856448 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 22:48:47.229528 138414566856448 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 22:48:47.229685 138543412606784 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0416 22:48:47.229849 138543412606784 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 22:48:47.230560 138543412606784 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.742, Tokens/s/device: 13828.401, 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: Thu Apr 16 22:48:56 UTC 2026 EXIT_CODE=0