XPK Start: Sun Apr 19 01:52:42 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 01:53:06.074494: 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) I0419 01:53:06.251777 136996464691008 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 01:53:15.291761 136996464691008 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 01:53:15.294084 136996464691008 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-true-hqm9a-slice-job-0-0.mt-10-shardy-true-hqm9a:8482 I0419 01:53:17.456552 136996464691008 max_utils.py:284] Jax distributed system initialized! I0419 01:53:23.505993 136996464691008 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 01:53:23.506097 136996464691008 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 01:53:23.506143 136996464691008 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 I0419 01:53:23.506185 136996464691008 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 01:53:24.573001 136996464691008 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 01:53:24.686623 136996464691008 checkpointing.py:688] Setting up checkpoint logger... I0419 01:53:24.686743 136996464691008 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 01:53:24.686789 136996464691008 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 01:53:24.686992 136996464691008 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 0x7c9858b18e90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 01:53:28.845077 136996464691008 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 01:53:28.845329 136996464691008 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c84244e9130>}, handler_registry=None I0419 01:53:28.845566 136996464691008 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c84244e9130>` 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`. I0419 01:53:28.845613 136996464691008 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c843c158890>` 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`. I0419 01:53:28.845649 136996464691008 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 0x7c84244e9130>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c84244e9130>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c843c158890>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c843c158890>}). I0419 01:53:28.845969 136996464691008 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 01:53:28.846040 136996464691008 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c84246984a0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 01:53:30.591730 136996464691008 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints I0419 01:53:31.034814 136996464691008 checkpoint_manager.py:929] [process=3][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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c9735a192e0> I0419 01:53:31.034996 136996464691008 checkpointing.py:302] Checkpoint manager created! I0419 01:53:32.001124 136996464691008 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 01:53:32.001256 136996464691008 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 01:53:33.408772 136996464691008 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 01:53:33.408886 136996464691008 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 01:53:33.408936 136996464691008 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 01:53:33.408978 136996464691008 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 01:53:33.409014 136996464691008 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 01:53:33.409050 136996464691008 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 01:53:33.409085 136996464691008 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 01:53:33.409119 136996464691008 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 01:53:33.409156 136996464691008 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 01:53:33.409189 136996464691008 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 01:53:33.409219 136996464691008 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 01:53:33.409261 136996464691008 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 01:53:33.812933 136996464691008 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 01:53:33.813023 136996464691008 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:53:33.819115 136996464691008 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 01:53:33.819178 136996464691008 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:53:33.835461 136996464691008 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 01:53:33.835521 136996464691008 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:53:33.851171 136996464691008 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 01:53:33.851228 136996464691008 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:53:33.875037 136996464691008 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 01:53:33.875103 136996464691008 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:53:33.890792 136996464691008 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 01:53:33.890852 136996464691008 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:53:33.906521 136996464691008 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 01:53:33.906581 136996464691008 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:53:33.937063 136996464691008 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 01:53:33.937155 136996464691008 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:53:33.958890 136996464691008 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 01:53:33.958963 136996464691008 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 01:53:45.368170 136996464691008 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. I0419 01:53:45.370510 136996464691008 metric_logger.py:301] number parameters: 1.104 billion I0419 01:53:56.586768 136996464691008 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 01:53:56.738766 136996464691008 checkpointing.py:798] Waited 0.15198397636413574 seconds for step 0 to finish before starting checkpointing. I0419 01:53:56.741043 136996464691008 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 01:53:56.742868 136996464691008 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0 I0419 01:53:56.744207 136996464691008 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints/0. I0419 01:53:57.510251 136996464691008 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 01:53:57.511269 136996464691008 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 01:53:57.511327 136996464691008 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 I0419 01:53:57.797540 136996464691008 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.288065s I0419 01:53:57.797714 136996464691008 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.191 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2971479892730713 s) (per-host) I0419 01:53:57.797765 136996464691008 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.297209s (batch_requests_ready=0.002978s, total_serialization_initiated=0.294162s, others=0.000068s) I0419 01:53:57.797860 136996464691008 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.301391s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.301374) I0419 01:53:57.798643 136996464691008 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.06 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints/0. I0419 01:53:57.798948 136868454905600 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:13:57.798916 I0419 01:53:57.800907 136996464691008 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:53:57.801180 136867874277120 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:53:57.801347 136996464691008 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776563636.7410238, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776563636.742907, 'checkpointer_blocking_duration_secs': 1.0561888217926025, 'get_old_steps_start_time': 1776563637.799121, 'get_old_steps_duration_secs': 3.62396240234375e-05, 'checkpoint_manager_blocking_start_time': 1776563636.7391994, 'checkpoint_manager_blocking_duration_secs': 1.062110185623169} I0419 01:53:57.801451 136996464691008 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 01:53:57.801519 136996464691008 max_utils.py:750] Memstats: After params initialized: I0419 01:53:57.801566 136996464691008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0419 01:53:57.801599 136996464691008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0419 01:53:57.801626 136996464691008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0419 01:53:57.801651 136996464691008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0419 01:53:58.119478 136996464691008 metric_logger.py:196] completed step: 0, seconds: 11.215, TFLOP/s/device: 1.211, Tokens/s/device: 182.609, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0419 01:53:58.306661 136996464691008 metric_logger.py:196] completed step: 1, seconds: 1.531, TFLOP/s/device: 8.875, Tokens/s/device: 1337.801, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0419 01:53:59.144932 136996464691008 metric_logger.py:196] completed step: 2, seconds: 0.041, TFLOP/s/device: 334.173, Tokens/s/device: 50370.152, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0419 01:53:59.292656 136996464691008 metric_logger.py:196] completed step: 3, seconds: 0.828, TFLOP/s/device: 16.413, Tokens/s/device: 2473.995, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0419 01:53:59.588476 136996464691008 metric_logger.py:196] completed step: 4, seconds: 0.165, TFLOP/s/device: 82.468, Tokens/s/device: 12430.428, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0419 01:53:59.595687 136996464691008 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.258, Tokens/s/device: 13906.147, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0419 01:54:01.538809 2787 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 01:54:04.830122 136867882669824 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints/0/items/array_metadatas/process_3 I0419 01:54:22.031399 136996464691008 metric_logger.py:196] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.836, Tokens/s/device: 6908.976, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0419 01:54:22.179497 136996464691008 metric_logger.py:196] completed step: 7, seconds: 22.289, TFLOP/s/device: 0.610, Tokens/s/device: 91.883, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0419 01:54:22.327279 136996464691008 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.685, Tokens/s/device: 13367.535, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0419 01:54:22.475578 136996464691008 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 01:54:22.476254 136996464691008 checkpointing.py:798] Waited 0.0006768703460693359 seconds for step 9 to finish before starting checkpointing. I0419 01:54:22.478305 136996464691008 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:54:35.041015 136868454905600 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 42.077 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.540422439575195 s) (per-host) I0419 01:54:35.041128 136868454905600 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.242073s. I0419 01:54:45.981251 136868454905600 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.182160s. I0419 01:54:45.981527 136867874277120 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:54:45.981653 136867874277120 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:54:45.981704 136867874277120 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:54:45.984303 136867874277120 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 01:54:45.984484 136996464691008 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 01:54:45.984620 136996464691008 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.506313 seconds. If this number is high, consider checkpointing less frequently. I0419 01:54:45.986588 136996464691008 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9 I0419 01:54:45.988586 136996464691008 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints/9. I0419 01:54:46.715371 136996464691008 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 01:54:46.715462 136996464691008 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 I0419 01:54:46.752481 136996464691008 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038760s I0419 01:54:46.752650 136996464691008 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 33.688 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.045789480209350586 s) (per-host) I0419 01:54:46.752702 136996464691008 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045851s (batch_requests_ready=0.002684s, total_serialization_initiated=0.043096s, others=0.000071s) I0419 01:54:46.752794 136996464691008 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049907s (all_items=0.000015s, per_item={'items': '0.00001478'}, temp_paths=0.049893) I0419 01:54:46.753550 136996464691008 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints/9. I0419 01:54:46.753902 136867874277120 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:14:46.753865 I0419 01:54:46.757277 136996464691008 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:54:46.757508 136846126331648 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:54:46.757665 136996464691008 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776563662.478275, 'wait_for_prev_duration_secs': 23.50631308555603, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776563685.986627, 'checkpointer_blocking_duration_secs': 0.7674160003662109, 'get_old_steps_start_time': 1776563686.7540658, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776563662.4764936, 'checkpoint_manager_blocking_duration_secs': 24.281137704849243} I0419 01:54:46.757773 136996464691008 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 01:54:46.757817 136996464691008 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:54:55.698571 136867882669824 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_true/checkpoints/9/items/array_metadatas/process_3 I0419 01:55:31.224397 136867874277120 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 35.483 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.517505168914795 s) (per-host) I0419 01:55:31.224509 136867874277120 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.470494s. I0419 01:55:39.996971 136867874277120 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 53.242941s. I0419 01:55:39.997313 136846126331648 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:55:39.997430 136846126331648 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:55:39.997476 136846126331648 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:55:40.000398 136846126331648 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 01:55:40.000601 136996464691008 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 01:55:40.000788 136996464691008 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 01:55:40.001781 136996464691008 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.503, Tokens/s/device: 13792.360, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3551.438 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 01:55:52 UTC 2026 EXIT_CODE=0