XPK Start: Wed Apr 22 14:47:13 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-22 14:47:37.928685: 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) I0422 14:47:38.463421 133855923664704 max_utils.py:273] Attempting to initialize the jax distributed system... I0422 14:47:47.504782 133855923664704 distributed.py:149] Starting JAX distributed service on [::]:8482 I0422 14:47:47.507113 133855923664704 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-lver8-slice-job-0-0.mt-10-shardy-false-lver8:8482 I0422 14:47:48.873081 133855923664704 max_utils.py:284] Jax distributed system initialized! I0422 14:47:54.058656 133855923664704 max_utils.py:800] System Information: Jax Version: 0.9.2 I0422 14:47:54.058756 133855923664704 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0422 14:47:54.058795 133855923664704 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 I0422 14:47:54.058829 133855923664704 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 14:47:54.754010 133855923664704 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 14:47:54.867989 133855923664704 checkpointing.py:688] Setting up checkpoint logger... I0422 14:47:54.868130 133855923664704 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 14:47:54.868178 133855923664704 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0422 14:47:54.868527 133855923664704 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 0x79bd0b1501a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 14:47:57.844371 133855923664704 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 14:47:57.844613 133855923664704 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a878619850>}, handler_registry=None I0422 14:47:57.844847 133855923664704 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a878619850>` 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`. I0422 14:47:57.844893 133855923664704 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79a878642870>` 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`. I0422 14:47:57.844927 133855923664704 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 0x79a878619850>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a878619850>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79a878642870>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79a878642870>}). I0422 14:47:57.845251 133855923664704 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0422 14:47:57.845320 133855923664704 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79a8786b99e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0422 14:47:58.858717 133855923664704 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints I0422 14:47:59.073020 133855923664704 checkpoint_manager.py:929] [process=6][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_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79a9606f23c0> I0422 14:47:59.073218 133855923664704 checkpointing.py:302] Checkpoint manager created! I0422 14:48:00.039140 133855923664704 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 14:48:00.039253 133855923664704 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 14:48:01.418255 133855923664704 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 14:48:01.418370 133855923664704 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 14:48:01.418421 133855923664704 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 14:48:01.418463 133855923664704 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 14:48:01.418500 133855923664704 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 14:48:01.418533 133855923664704 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 14:48:01.418573 133855923664704 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 14:48:01.418608 133855923664704 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 14:48:01.418641 133855923664704 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 14:48:01.418673 133855923664704 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 14:48:01.418705 133855923664704 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 14:48:01.418736 133855923664704 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 14:48:01.817846 133855923664704 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 14:48:01.817936 133855923664704 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 14:48:01.823217 133855923664704 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 14:48:01.823281 133855923664704 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 14:48:01.839970 133855923664704 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 14:48:01.840028 133855923664704 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 14:48:01.856121 133855923664704 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 14:48:01.856179 133855923664704 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 14:48:01.880393 133855923664704 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 14:48:01.880462 133855923664704 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 14:48:01.896453 133855923664704 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 14:48:01.896514 133855923664704 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 14:48:01.912558 133855923664704 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 14:48:01.912616 133855923664704 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 14:48:01.943124 133855923664704 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 14:48:01.943197 133855923664704 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 14:48:01.963327 133855923664704 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 14:48:01.963393 133855923664704 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 14:48:13.461728 133855923664704 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. I0422 14:48:13.464174 133855923664704 metric_logger.py:301] number parameters: 1.104 billion I0422 14:48:24.595174 133855923664704 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 14:48:24.749945 133855923664704 checkpointing.py:798] Waited 0.1547558307647705 seconds for step 0 to finish before starting checkpointing. I0422 14:48:24.752270 133855923664704 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 14:48:24.753770 133855923664704 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0422 14:48:24.755260 133855923664704 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints/0. I0422 14:48:25.074216 133855923664704 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 14:48:25.075244 133855923664704 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0422 14:48:25.075326 133855923664704 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 I0422 14:48:25.345378 133855923664704 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.272019s I0422 14:48:25.345559 133855923664704 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.477 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28163647651672363 s) (per-host) I0422 14:48:25.345613 133855923664704 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281701s (batch_requests_ready=0.003161s, total_serialization_initiated=0.278468s, others=0.000072s) I0422 14:48:25.345709 133855923664704 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285967s (all_items=0.000023s, per_item={'items': '0.00002313'}, temp_paths=0.285944) I0422 14:48:25.346524 133855923664704 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.59 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints/0. I0422 14:48:25.346824 133730246153984 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:08:25.346787 I0422 14:48:25.361556 133855923664704 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 14:48:25.361834 133729654613760 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 14:48:25.361968 133855923664704 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776869304.7522507, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776869304.7538092, 'checkpointer_blocking_duration_secs': 0.593127965927124, 'get_old_steps_start_time': 1776869305.346964, 'get_old_steps_duration_secs': 3.4332275390625e-05, 'checkpoint_manager_blocking_start_time': 1776869304.7503579, 'checkpoint_manager_blocking_duration_secs': 0.6115705966949463} I0422 14:48:25.362081 133855923664704 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 14:48:25.362161 133855923664704 max_utils.py:750] Memstats: After params initialized: I0422 14:48:25.362210 133855923664704 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 14:48:25.362241 133855923664704 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 14:48:25.362268 133855923664704 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 14:48:25.362292 133855923664704 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 14:48:25.676612 133855923664704 metric_logger.py:196] completed step: 0, seconds: 11.130, TFLOP/s/device: 1.221, Tokens/s/device: 184.009, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0422 14:48:25.846708 133855923664704 metric_logger.py:196] completed step: 1, seconds: 1.079, TFLOP/s/device: 12.587, Tokens/s/device: 1897.268, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0422 14:48:26.323597 133855923664704 metric_logger.py:196] completed step: 2, seconds: 0.024, TFLOP/s/device: 572.234, Tokens/s/device: 86253.369, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0422 14:48:26.471067 133855923664704 metric_logger.py:196] completed step: 3, seconds: 0.467, TFLOP/s/device: 29.107, Tokens/s/device: 4387.290, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0422 14:48:26.767004 133855923664704 metric_logger.py:196] completed step: 4, seconds: 0.165, TFLOP/s/device: 82.499, Tokens/s/device: 12435.107, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0422 14:48:26.774624 133855923664704 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.387, Tokens/s/device: 13925.531, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0422 14:48:28.806764 2753 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 14:48:30.818603 133729663006464 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints/0/items/array_metadatas/process_6 I0422 14:48:48.759488 133855923664704 metric_logger.py:196] completed step: 6, seconds: 0.297, TFLOP/s/device: 45.824, Tokens/s/device: 6907.042, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0422 14:48:48.907077 133855923664704 metric_logger.py:196] completed step: 7, seconds: 21.839, TFLOP/s/device: 0.622, Tokens/s/device: 93.778, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0422 14:48:49.054694 133855923664704 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.736, Tokens/s/device: 13375.305, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0422 14:48:49.202920 133855923664704 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 14:48:49.203609 133855923664704 checkpointing.py:798] Waited 0.0007026195526123047 seconds for step 9 to finish before starting checkpointing. I0422 14:48:49.205507 133855923664704 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 14:49:02.124770 133730246153984 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 42.622 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.06080770492554 s) (per-host) I0422 14:49:02.124900 133730246153984 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.777401s. I0422 14:49:09.023542 133730246153984 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 43.676023s. I0422 14:49:09.023826 133729654613760 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 14:49:09.023947 133729654613760 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 14:49:09.023997 133729654613760 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 14:49:09.025467 133729654613760 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 14:49:09.025649 133855923664704 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0422 14:49:09.025777 133855923664704 checkpoint_manager.py:1452] Waiting for previous save to complete took 19.820276 seconds. If this number is high, consider checkpointing less frequently. I0422 14:49:09.027776 133855923664704 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0422 14:49:09.029826 133855923664704 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints/9. I0422 14:49:09.335762 133855923664704 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0422 14:49:09.335857 133855923664704 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 I0422 14:49:09.378110 133855923664704 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043861s I0422 14:49:09.378278 133855923664704 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 30.283 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05093812942504883 s) (per-host) I0422 14:49:09.378327 133855923664704 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.050997s (batch_requests_ready=0.002639s, total_serialization_initiated=0.048289s, others=0.000069s) I0422 14:49:09.378418 133855923664704 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.055122s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.055106) I0422 14:49:09.379168 133855923664704 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints/9. I0422 14:49:09.379444 133729654613760 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:09:09.379423 I0422 14:49:09.381338 133855923664704 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 14:49:09.381640 133729663006464 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 14:49:09.381776 133855923664704 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776869329.205471, 'wait_for_prev_duration_secs': 19.820276260375977, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776869349.0278156, 'checkpointer_blocking_duration_secs': 0.35172247886657715, 'get_old_steps_start_time': 1776869349.3795607, 'get_old_steps_duration_secs': 3.361701965332031e-05, 'checkpoint_manager_blocking_start_time': 1776869329.2038462, 'checkpoint_manager_blocking_duration_secs': 20.17789602279663} I0422 14:49:09.381937 133855923664704 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 14:49:09.381982 133855923664704 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 14:49:14.511952 133702248027904 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_10_shardy_false/checkpoints/9/items/array_metadatas/process_6 I0422 14:49:51.539353 133729654613760 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.421 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.21199417114258 s) (per-host) I0422 14:49:51.539463 133729654613760 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.159948s. I0422 14:49:56.958029 133729654613760 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 47.578493s. I0422 14:49:56.958390 133729663006464 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 14:49:56.958512 133729663006464 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 14:49:56.958570 133729663006464 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 14:49:56.961088 133729663006464 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 14:49:56.961294 133855923664704 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0422 14:49:56.961454 133855923664704 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 14:49:56.962497 133855923664704 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.551, Tokens/s/device: 13799.609, 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: Wed Apr 22 14:50:09 UTC 2026 EXIT_CODE=0