XPK Start: Fri Apr 17 05:26:32 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 05:26:56.330011: 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 05:26:56.506527 136598711007040 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 05:27:05.546413 136598711007040 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 05:27:05.548855 136598711007040 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-09bnm-slice-job-0-0.mt-10-shardy-false-09bnm:8482 I0417 05:27:06.608706 136598711007040 max_utils.py:284] Jax distributed system initialized! I0417 05:27:12.620257 136598711007040 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 05:27:12.620357 136598711007040 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 05:27:12.620399 136598711007040 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 05:27:12.620433 136598711007040 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 05:27:13.292001 136598711007040 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 05:27:13.405614 136598711007040 checkpointing.py:688] Setting up checkpoint logger... I0417 05:27:13.405728 136598711007040 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 05:27:13.405772 136598711007040 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 05:27:13.405996 136598711007040 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 0x7c3ba6139be0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 05:27:16.682215 136598711007040 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 05:27:16.682446 136598711007040 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 0x7c26643f4590>}, handler_registry=None I0417 05:27:16.682680 136598711007040 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c26643f4590>` 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 05:27:16.682728 136598711007040 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c2644653aa0>` 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 05:27:16.682763 136598711007040 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 0x7c26643f4590>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c26643f4590>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c2644653aa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c2644653aa0>}). I0417 05:27:16.683113 136598711007040 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 05:27:16.683185 136598711007040 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c26643e6020> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 05:27:17.968019 136598711007040 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints I0417 05:27:18.415555 136598711007040 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c26446adca0> I0417 05:27:18.415725 136598711007040 checkpointing.py:302] Checkpoint manager created! I0417 05:27:19.357523 136598711007040 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 05:27:19.357646 136598711007040 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 05:27:20.739784 136598711007040 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0417 05:27:20.739912 136598711007040 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 05:27:20.739961 136598711007040 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 05:27:20.740004 136598711007040 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0417 05:27:20.740041 136598711007040 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 05:27:20.740076 136598711007040 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 05:27:20.740112 136598711007040 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 05:27:20.740147 136598711007040 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0417 05:27:20.740181 136598711007040 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 05:27:20.740211 136598711007040 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 05:27:20.740243 136598711007040 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0417 05:27:20.740274 136598711007040 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0417 05:27:21.141254 136598711007040 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 05:27:21.141342 136598711007040 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:27:21.147490 136598711007040 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 05:27:21.147551 136598711007040 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:27:21.164103 136598711007040 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 05:27:21.164164 136598711007040 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:27:21.179650 136598711007040 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 05:27:21.179706 136598711007040 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:27:21.203550 136598711007040 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 05:27:21.203614 136598711007040 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:27:21.219228 136598711007040 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 05:27:21.219292 136598711007040 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:27:21.234735 136598711007040 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 05:27:21.234791 136598711007040 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:27:21.264308 136598711007040 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 05:27:21.264376 136598711007040 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:27:21.283851 136598711007040 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 05:27:21.283914 136598711007040 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 05:27:32.266908 136598711007040 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 05:27:32.275892 136598711007040 metric_logger.py:289] number parameters: 1.104 billion I0417 05:27:43.500504 136598711007040 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 05:27:43.652389 136598711007040 checkpointing.py:798] Waited 0.1518704891204834 seconds for step 0 to finish before starting checkpointing. I0417 05:27:43.654783 136598711007040 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 05:27:43.656577 136598711007040 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 05:27:43.657977 136598711007040 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0. I0417 05:27:44.389990 136598711007040 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 05:27:44.409532 136598711007040 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 05:27:44.409605 136598711007040 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 05:27:44.678078 136598711007040 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.269998s I0417 05:27:44.678928 136598711007040 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.491 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2809321880340576 s) (per-host) I0417 05:27:44.679007 136598711007040 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281031s (batch_requests_ready=0.003045s, total_serialization_initiated=0.277216s, others=0.000769s) I0417 05:27:44.679100 136598711007040 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285060s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.285043) I0417 05:27:44.679859 136598711007040 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0. I0417 05:27:44.680214 136469605537536 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 05:47:44.680176 I0417 05:27:44.693012 136598711007040 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 05:27:44.693315 136466374055680 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 05:27:44.693472 136598711007040 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776403663.6547637, 'wait_for_prev_duration_secs': 8.225440979003906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776403663.656614, 'checkpointer_blocking_duration_secs': 1.0237493515014648, 'get_old_steps_start_time': 1776403664.6803887, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776403663.6527803, 'checkpoint_manager_blocking_duration_secs': 1.0406525135040283} I0417 05:27:44.693578 136598711007040 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 05:27:44.693643 136598711007040 max_utils.py:750] Memstats: After params initialized: I0417 05:27:44.693690 136598711007040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 05:27:44.693723 136598711007040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 05:27:44.693751 136598711007040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 05:27:44.693775 136598711007040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 05:27:45.032458 136598711007040 metric_logger.py:185] completed step: 0, seconds: 10.921, TFLOP/s/device: 1.244, Tokens/s/device: 187.534, total_weights: 65536, loss: 10.874 I0417 05:27:45.035259 136598711007040 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/tensorboard/' I0417 05:27:45.116121 136469597144832 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0 I0417 05:27:45.431100 136598711007040 metric_logger.py:185] completed step: 1, seconds: 1.529, TFLOP/s/device: 8.888, Tokens/s/device: 1339.768, total_weights: 65536, loss: 10.874 I0417 05:27:45.572946 136598711007040 metric_logger.py:185] completed step: 2, seconds: 0.400, TFLOP/s/device: 34.000, Tokens/s/device: 5124.920, total_weights: 65536, loss: 10.262 I0417 05:27:45.720538 136598711007040 metric_logger.py:185] completed step: 3, seconds: 0.013, TFLOP/s/device: 1052.613, Tokens/s/device: 158661.295, total_weights: 65536, loss: 9.732 I0417 05:27:47.086303 136469580359424 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/items I0417 05:27:47.244069 136462346544896 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776403666840042577, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776403666840042577, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0417 05:27:48.244715 2837 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 05:27:49.964384 136467432380160 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/items/array_metadatas/process_0 I0417 05:28:08.733564 136598711007040 metric_logger.py:185] completed step: 4, seconds: 0.142, TFLOP/s/device: 95.372, Tokens/s/device: 14375.461, total_weights: 65536, loss: 9.274 I0417 05:28:08.743873 136598711007040 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.516, Tokens/s/device: 13945.064, total_weights: 65536, loss: 8.888 I0417 05:28:08.888428 136598711007040 metric_logger.py:185] completed step: 6, seconds: 23.012, TFLOP/s/device: 0.590, Tokens/s/device: 88.998, total_weights: 65536, loss: 8.589 I0417 05:28:09.036062 136598711007040 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 1725.131, Tokens/s/device: 260030.472, total_weights: 65536, loss: 8.380 I0417 05:28:09.183517 136598711007040 metric_logger.py:185] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.307, Tokens/s/device: 13913.516, total_weights: 65536, loss: 8.251 I0417 05:28:09.331876 136598711007040 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 05:28:09.332643 136598711007040 checkpointing.py:798] Waited 0.0007772445678710938 seconds for step 9 to finish before starting checkpointing. I0417 05:28:09.334641 136598711007040 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 05:28:24.656035 136466382448384 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.417641s (commit=35.225098s, array_metadata_write=1.192543s) I0417 05:28:24.657162 136469605537536 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 39.236 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.259170055389404 s) (per-host) I0417 05:28:24.657238 136469605537536 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.976913s. I0417 05:28:25.345812 136469605537536 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776403666840042577, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0417 05:28:26.059814 136469605537536 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000418s. I0417 05:28:26.074135 136462346544896 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776403666840042577, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0417 05:28:26.433269 136469605537536 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 05:28:27.516674 136469605537536 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.043133s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/items I0417 05:28:27.518589 136469605537536 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0/items I0417 05:28:28.492522 136469605537536 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0 I0417 05:28:30.348611 136469605537536 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0`. I0417 05:28:30.349428 136469605537536 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 46.69 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/0 I0417 05:28:30.350713 136469605537536 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 45.670384s. I0417 05:28:30.350879 136466374055680 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 05:28:30.350965 136466374055680 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 05:28:30.351032 136466374055680 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 05:28:30.353182 136466374055680 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 05:28:30.353348 136598711007040 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 05:28:30.353483 136598711007040 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.018842 seconds. If this number is high, consider checkpointing less frequently. I0417 05:28:30.355260 136598711007040 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 05:28:30.357236 136598711007040 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9. I0417 05:28:31.062195 136598711007040 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 05:28:31.062293 136598711007040 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 05:28:31.089335 136598711007040 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028579s I0417 05:28:31.090156 136598711007040 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.305 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03734612464904785 s) (per-host) I0417 05:28:31.090226 136598711007040 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037431s (batch_requests_ready=0.002660s, total_serialization_initiated=0.033995s, others=0.000776s) I0417 05:28:31.090301 136598711007040 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041422s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.041404) I0417 05:28:31.091019 136598711007040 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9. I0417 05:28:31.091364 136468525725440 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 05:48:31.091324 I0417 05:28:31.102530 136598711007040 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 05:28:31.102813 136466374055680 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 05:28:31.102993 136598711007040 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776403689.3346088, 'wait_for_prev_duration_secs': 21.018841981887817, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776403710.3552985, 'checkpointer_blocking_duration_secs': 0.7362091541290283, 'get_old_steps_start_time': 1776403711.091533, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776403689.3329122, 'checkpoint_manager_blocking_duration_secs': 21.770044088363647} I0417 05:28:31.103111 136598711007040 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 05:28:31.103157 136598711007040 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 05:28:31.302013 136469605537536 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9 I0417 05:28:33.468887 136468559296256 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9/items I0417 05:28:36.751367 136468550903552 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9/items/array_metadatas/process_0 I0417 05:29:16.890989 136468542510848 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.888296s (commit=40.196622s, array_metadata_write=1.691674s) I0417 05:29:16.892057 136468525725440 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.460 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.83921456336975 s) (per-host) I0417 05:29:16.892184 136468525725440 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.800706s. I0417 05:29:18.778517 136468525725440 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000366s. I0417 05:29:18.979053 136468525725440 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 05:29:20.044963 136468525725440 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.880888s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9/items I0417 05:29:20.046688 136468525725440 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9/items I0417 05:29:20.572308 136468525725440 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9 I0417 05:29:22.445365 136468525725440 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_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9`. I0417 05:29:22.446213 136468525725440 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.09 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_10_shardy_false/checkpoints/9 I0417 05:29:22.447438 136468525725440 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.355958s. I0417 05:29:22.447594 136466374055680 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 05:29:22.447669 136466374055680 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 05:29:22.447731 136466374055680 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 05:29:22.450874 136466374055680 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 05:29:22.451041 136598711007040 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 05:29:22.451207 136598711007040 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 05:29:22.451875 136598711007040 metric_logger.py:185] completed step: 9, seconds: 0.146, TFLOP/s/device: 92.867, Tokens/s/device: 13997.868, 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 05:29:33 UTC 2026 EXIT_CODE=0