XPK Start: Fri Apr 17 01:18:30 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-17 01:18:53.741857: 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) I0417 01:18:54.271690 136193301362496 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 01:19:03.311096 136193301362496 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 01:19:03.313400 136193301362496 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-true-etm58-slice-job-0-0.mt-10-shardy-true-etm58:8482 I0417 01:19:04.570809 136193301362496 max_utils.py:284] Jax distributed system initialized! I0417 01:19:10.716707 136193301362496 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 01:19:10.716808 136193301362496 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 01:19:10.716876 136193301362496 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 I0417 01:19:10.716915 136193301362496 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 01:19:11.400599 136193301362496 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 01:19:11.514712 136193301362496 checkpointing.py:688] Setting up checkpoint logger... I0417 01:19:11.514862 136193301362496 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 01:19:11.514913 136193301362496 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 01:19:11.515125 136193301362496 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 0x7bdd67d1c140>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 01:19:14.390645 136193301362496 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 01:19:14.390891 136193301362496 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc91028d880>}, handler_registry=None I0417 01:19:14.391137 136193301362496 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc91028d880>` 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`. I0417 01:19:14.391184 136193301362496 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc9107aea20>` 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`. I0417 01:19:14.391219 136193301362496 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 0x7bc91028d880>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc91028d880>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc9107aea20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc9107aea20>}). I0417 01:19:14.391531 136193301362496 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 01:19:14.391600 136193301362496 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bc9102c9ee0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 01:19:15.555881 136193301362496 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints I0417 01:19:15.983944 136193301362496 checkpoint_manager.py:929] [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_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_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bc91023a0c0> I0417 01:19:15.984121 136193301362496 checkpointing.py:302] Checkpoint manager created! I0417 01:19:16.926025 136193301362496 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 01:19:16.926149 136193301362496 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 01:19:18.288282 136193301362496 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0417 01:19:18.288395 136193301362496 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 01:19:18.288445 136193301362496 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 01:19:18.288486 136193301362496 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0417 01:19:18.288522 136193301362496 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 01:19:18.288556 136193301362496 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 01:19:18.288590 136193301362496 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:19:18.288624 136193301362496 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0417 01:19:18.288658 136193301362496 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:19:18.288689 136193301362496 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:19:18.288720 136193301362496 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0417 01:19:18.288751 136193301362496 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0417 01:19:18.686957 136193301362496 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 01:19:18.687049 136193301362496 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:19:18.692353 136193301362496 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 01:19:18.692407 136193301362496 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:19:18.708517 136193301362496 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 01:19:18.708584 136193301362496 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:19:18.724028 136193301362496 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 01:19:18.724088 136193301362496 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:19:18.747850 136193301362496 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 01:19:18.747918 136193301362496 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:19:18.763515 136193301362496 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 01:19:18.763580 136193301362496 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:19:18.779155 136193301362496 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 01:19:18.779213 136193301362496 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:19:18.808665 136193301362496 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 01:19:18.808734 136193301362496 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:19:18.828277 136193301362496 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 01:19:18.828342 136193301362496 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 01:19:30.066022 136193301362496 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. I0417 01:19:30.075093 136193301362496 metric_logger.py:289] number parameters: 1.104 billion I0417 01:19:41.640932 136193301362496 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 01:19:41.793274 136193301362496 checkpointing.py:798] Waited 0.15232467651367188 seconds for step 0 to finish before starting checkpointing. I0417 01:19:41.796513 136193301362496 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 01:19:41.798266 136193301362496 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 01:19:41.799632 136193301362496 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0. I0417 01:19:42.519068 136193301362496 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 01:19:42.538220 136193301362496 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 01:19:42.538290 136193301362496 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 I0417 01:19:42.804302 136065089357568 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0 I0417 01:19:42.818696 136193301362496 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281907s I0417 01:19:42.819505 136193301362496 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.270 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29268407821655273 s) (per-host) I0417 01:19:42.819572 136193301362496 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.292764s (batch_requests_ready=0.003062s, total_serialization_initiated=0.289017s, others=0.000685s) I0417 01:19:42.819660 136193301362496 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.296687s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.296670) I0417 01:19:42.820465 136193301362496 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0. I0417 01:19:42.820743 136065097750272 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 01:39:42.820712 I0417 01:19:42.825178 136193301362496 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 01:19:42.825423 136064018671360 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 01:19:42.825548 136193301362496 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776388781.796494, 'wait_for_prev_duration_secs': 5.984306335449219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776388781.798304, 'checkpointer_blocking_duration_secs': 1.0225539207458496, 'get_old_steps_start_time': 1776388782.820881, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776388781.7936678, 'checkpoint_manager_blocking_duration_secs': 1.031844139099121} I0417 01:19:42.825652 136193301362496 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 01:19:42.825717 136193301362496 max_utils.py:750] Memstats: After params initialized: I0417 01:19:42.825764 136193301362496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 01:19:42.825795 136193301362496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 01:19:42.825822 136193301362496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 01:19:42.825862 136193301362496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 01:19:43.157279 136193301362496 metric_logger.py:185] completed step: 0, seconds: 11.253, TFLOP/s/device: 1.207, Tokens/s/device: 181.998, total_weights: 65536, loss: 10.874 I0417 01:19:43.160050 136193301362496 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/tensorboard/' I0417 01:19:43.583113 136193301362496 metric_logger.py:185] completed step: 1, seconds: 1.513, TFLOP/s/device: 8.980, Tokens/s/device: 1353.619, total_weights: 65536, loss: 10.874 I0417 01:19:43.724689 136193301362496 metric_logger.py:185] completed step: 2, seconds: 0.427, TFLOP/s/device: 31.839, Tokens/s/device: 4799.153, total_weights: 65536, loss: 10.262 I0417 01:19:43.872387 136193301362496 metric_logger.py:185] completed step: 3, seconds: 0.012, TFLOP/s/device: 1126.908, Tokens/s/device: 169859.832, total_weights: 65536, loss: 9.732 I0417 01:19:45.124317 136065072572160 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/items I0417 01:19:45.324579 136060748973824 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776388784873283055, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776388784873283055, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0417 01:19:46.284989 2778 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 01:19:48.980367 136064507102976 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/items/array_metadatas/process_0 I0417 01:20:07.279664 136193301362496 metric_logger.py:185] completed step: 4, seconds: 0.143, TFLOP/s/device: 94.719, Tokens/s/device: 14277.150, total_weights: 65536, loss: 9.274 I0417 01:20:07.289702 136193301362496 metric_logger.py:185] completed step: 5, seconds: 0.146, TFLOP/s/device: 92.755, Tokens/s/device: 13981.049, total_weights: 65536, loss: 8.888 I0417 01:20:07.434270 136193301362496 metric_logger.py:185] completed step: 6, seconds: 23.406, TFLOP/s/device: 0.580, Tokens/s/device: 87.499, total_weights: 65536, loss: 8.589 I0417 01:20:07.581567 136193301362496 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 1759.307, Tokens/s/device: 265181.924, total_weights: 65536, loss: 8.380 I0417 01:20:07.729472 136193301362496 metric_logger.py:185] completed step: 8, seconds: 0.146, TFLOP/s/device: 93.011, Tokens/s/device: 14019.619, total_weights: 65536, loss: 8.251 I0417 01:20:07.878056 136193301362496 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 01:20:07.878714 136193301362496 checkpointing.py:798] Waited 0.0006718635559082031 seconds for step 9 to finish before starting checkpointing. I0417 01:20:07.880674 136193301362496 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 01:20:23.098578 136064027064064 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.811265s (commit=35.638318s, array_metadata_write=1.172946s) I0417 01:20:23.099695 136065097750272 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 38.932 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.572834491729736 s) (per-host) I0417 01:20:23.099828 136065097750272 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.278491s. I0417 01:20:23.760024 136065097750272 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776388784873283055, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0417 01:20:24.535966 136060748973824 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776388784873283055, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0417 01:20:24.777719 136065097750272 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000417s. I0417 01:20:25.167698 136065097750272 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 01:20:26.103776 136065097750272 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.224606s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/items I0417 01:20:26.105447 136065097750272 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0/items I0417 01:20:26.675907 136065097750272 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0 I0417 01:20:28.139186 136065097750272 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0`. I0417 01:20:28.139888 136065097750272 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 46.34 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/0 I0417 01:20:28.141332 136065097750272 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 45.319994s. I0417 01:20:28.141501 136064018671360 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 01:20:28.141620 136064018671360 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 01:20:28.141684 136064018671360 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 01:20:28.143676 136064018671360 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 01:20:28.143865 136193301362496 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0417 01:20:28.144005 136193301362496 checkpoint_manager.py:1452] Waiting for previous save to complete took 20.263329 seconds. If this number is high, consider checkpointing less frequently. I0417 01:20:28.145872 136193301362496 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 01:20:28.147971 136193301362496 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9. I0417 01:20:28.861596 136193301362496 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 01:20:28.861691 136193301362496 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 I0417 01:20:28.888391 136193301362496 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028307s I0417 01:20:28.889162 136193301362496 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.738 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03695869445800781 s) (per-host) I0417 01:20:28.889230 136193301362496 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037045s (batch_requests_ready=0.002616s, total_serialization_initiated=0.033705s, others=0.000723s) I0417 01:20:28.889302 136193301362496 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041147s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.041131) I0417 01:20:28.890040 136193301362496 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9. I0417 01:20:28.890401 136064018671360 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 01:40:28.890358 I0417 01:20:28.904009 136193301362496 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 01:20:28.904254 136063473407744 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 01:20:28.904389 136193301362496 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776388807.8806431, 'wait_for_prev_duration_secs': 20.263329029083252, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776388828.1459105, 'checkpointer_blocking_duration_secs': 0.7446489334106445, 'get_old_steps_start_time': 1776388828.8905833, 'get_old_steps_duration_secs': 3.314018249511719e-05, 'checkpoint_manager_blocking_start_time': 1776388807.8789575, 'checkpoint_manager_blocking_duration_secs': 21.025397300720215} I0417 01:20:28.904501 136193301362496 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 01:20:28.904546 136193301362496 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 01:20:29.099785 136065097750272 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9 I0417 01:20:31.273900 136064547149568 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9/items I0417 01:20:34.127674 136063490193152 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9/items/array_metadatas/process_0 I0417 01:21:12.860522 136063481800448 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.482042s (commit=39.269105s, array_metadata_write=1.212937s) I0417 01:21:12.861971 136064018671360 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.892 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.009737730026245 s) (per-host) I0417 01:21:12.862102 136064018671360 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.971568s. I0417 01:21:14.749631 136064018671360 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000439s. I0417 01:21:15.046979 136064018671360 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 01:21:15.692562 136064018671360 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.544540s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9/items I0417 01:21:15.694194 136064018671360 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9/items I0417 01:21:16.224418 136064018671360 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9 I0417 01:21:17.627582 136064018671360 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9`. I0417 01:21:17.628455 136064018671360 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.48 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_10_shardy_true/checkpoints/9 I0417 01:21:17.630073 136064018671360 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.739538s. I0417 01:21:17.630289 136063473407744 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 01:21:17.630415 136063473407744 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 01:21:17.630479 136063473407744 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 01:21:17.632656 136063473407744 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 01:21:17.632850 136193301362496 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0417 01:21:17.633015 136193301362496 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 01:21:17.633627 136193301362496 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.333, Tokens/s/device: 13917.487, 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: Fri Apr 17 01:21:25 UTC 2026 EXIT_CODE=0