XPK Start: Sun Apr 19 01:56:22 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 01:56:47.040824: 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 01:56:47.219144 137893174937408 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 01:56:56.260269 137893174937408 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 01:56:56.262486 137893174937408 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-660fq-slice-job-0-0.mt-10-shardy-false-660fq:8482 I0419 01:56:57.804432 137893174937408 max_utils.py:284] Jax distributed system initialized! I0419 01:57:04.196191 137893174937408 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 01:57:04.196313 137893174937408 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 01:57:04.196355 137893174937408 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 01:57:04.196390 137893174937408 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 01:57:04.895719 137893174937408 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 01:57:05.006940 137893174937408 checkpointing.py:688] Setting up checkpoint logger... I0419 01:57:05.007055 137893174937408 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 01:57:05.007101 137893174937408 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 01:57:05.007314 137893174937408 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 0x7d690a223770>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 01:57:09.192737 137893174937408 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 01:57:09.192972 137893174937408 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 0x7d54dc49a690>}, handler_registry=None I0419 01:57:09.193207 137893174937408 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d54dc49a690>` 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 01:57:09.193269 137893174937408 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d54b462c5f0>` 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 01:57:09.193307 137893174937408 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 0x7d54dc49a690>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d54dc49a690>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d54b462c5f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d54b462c5f0>}). I0419 01:57:09.193626 137893174937408 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 01:57:09.193696 137893174937408 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d54dc1d8fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 01:57:11.117060 137893174937408 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_false/checkpoints I0419 01:57:11.119403 137893174937408 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d54dc498bc0> I0419 01:57:11.119514 137893174937408 checkpointing.py:302] Checkpoint manager created! I0419 01:57:12.082595 137893174937408 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 01:57:12.082712 137893174937408 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 01:57:13.437863 137893174937408 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 01:57:13.437977 137893174937408 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 01:57:13.438029 137893174937408 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 01:57:13.438072 137893174937408 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 01:57:13.438109 137893174937408 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 01:57:13.438143 137893174937408 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 01:57:13.438178 137893174937408 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 01:57:13.438213 137893174937408 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 01:57:13.438262 137893174937408 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 01:57:13.438305 137893174937408 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 01:57:13.438338 137893174937408 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 01:57:13.438371 137893174937408 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 01:57:13.835649 137893174937408 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 01:57:13.835741 137893174937408 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:57:13.840958 137893174937408 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 01:57:13.841014 137893174937408 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:57:13.857161 137893174937408 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 01:57:13.857224 137893174937408 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:57:13.872737 137893174937408 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 01:57:13.872795 137893174937408 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:57:13.897488 137893174937408 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 01:57:13.897558 137893174937408 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:57:13.913147 137893174937408 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 01:57:13.913208 137893174937408 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:57:13.928730 137893174937408 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 01:57:13.928790 137893174937408 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:57:13.958623 137893174937408 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 01:57:13.958695 137893174937408 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:57:13.978375 137893174937408 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 01:57:13.978438 137893174937408 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 01:57:25.342980 137893174937408 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 01:57:25.345385 137893174937408 metric_logger.py:301] number parameters: 1.104 billion I0419 01:57:36.580312 137893174937408 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 01:57:36.733270 137893174937408 checkpointing.py:798] Waited 0.15292119979858398 seconds for step 0 to finish before starting checkpointing. I0419 01:57:36.735692 137893174937408 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 01:57:36.737350 137893174937408 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0419 01:57:36.738923 137893174937408 event_tracking.py:70] [process=6] [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_false/checkpoints/0. I0419 01:57:37.896788 137893174937408 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 01:57:37.897621 137893174937408 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 01:57:37.897682 137893174937408 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 01:57:38.177032 137893174937408 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281036s I0419 01:57:38.177217 137893174937408 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.307 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29065394401550293 s) (per-host) I0419 01:57:38.177306 137893174937408 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.290748s (batch_requests_ready=0.003532s, total_serialization_initiated=0.287113s, others=0.000103s) I0419 01:57:38.177457 137893174937408 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.294821s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.294803) I0419 01:57:38.178329 137893174937408 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.44 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_false/checkpoints/0. I0419 01:57:38.178666 137764095915776 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:17:38.178632 I0419 01:57:38.181764 137893174937408 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:57:38.182060 137762485303040 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:57:38.182217 137893174937408 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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776563856.7356722, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776563856.7373886, 'checkpointer_blocking_duration_secs': 1.441441297531128, 'get_old_steps_start_time': 1776563858.1788554, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776563856.7336793, 'checkpoint_manager_blocking_duration_secs': 1.4484984874725342} I0419 01:57:38.182350 137893174937408 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 01:57:38.182421 137893174937408 max_utils.py:750] Memstats: After params initialized: I0419 01:57:38.182471 137893174937408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 01:57:38.182503 137893174937408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 01:57:38.182529 137893174937408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 01:57:38.182552 137893174937408 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 01:57:38.502324 137893174937408 metric_logger.py:196] completed step: 0, seconds: 11.234, TFLOP/s/device: 1.209, Tokens/s/device: 182.307, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0419 01:57:38.678705 137893174937408 metric_logger.py:196] completed step: 1, seconds: 1.920, TFLOP/s/device: 7.076, Tokens/s/device: 1066.641, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0419 01:57:39.056196 137893174937408 metric_logger.py:196] completed step: 2, seconds: 0.030, TFLOP/s/device: 457.479, Tokens/s/device: 68956.229, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0419 01:57:39.203948 137893174937408 metric_logger.py:196] completed step: 3, seconds: 0.368, TFLOP/s/device: 36.954, Tokens/s/device: 5570.106, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0419 01:57:39.499645 137893174937408 metric_logger.py:196] completed step: 4, seconds: 0.165, TFLOP/s/device: 82.274, Tokens/s/device: 12401.298, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0419 01:57:39.506739 137893174937408 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.665, Tokens/s/device: 13967.509, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0419 01:57:43.180437 2826 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 01:57:46.913270 137762967242496 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/items/array_metadatas/process_6 I0419 01:58:01.743462 137893174937408 metric_logger.py:196] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.884, Tokens/s/device: 6916.115, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0419 01:58:01.891297 137893174937408 metric_logger.py:196] completed step: 7, seconds: 22.090, TFLOP/s/device: 0.615, Tokens/s/device: 92.711, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0419 01:58:02.038696 137893174937408 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.742, Tokens/s/device: 13376.179, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0419 01:58:02.186821 137893174937408 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 01:58:02.187512 137893174937408 checkpointing.py:798] Waited 0.0007038116455078125 seconds for step 9 to finish before starting checkpointing. I0419 01:58:02.189686 137893174937408 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:58:16.498691 137764095915776 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 40.909 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.612077474594116 s) (per-host) I0419 01:58:16.498812 137764095915776 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.320031s. I0419 01:58:26.887557 137764095915776 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.708760s. I0419 01:58:26.887862 137762485303040 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:58:26.887995 137762485303040 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:58:26.888048 137762485303040 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:58:26.890630 137762485303040 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 01:58:26.890814 137893174937408 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. W0419 01:58:26.890954 137893174937408 checkpoint_manager.py:1452] Waiting for previous save to complete took 24.701267 seconds. If this number is high, consider checkpointing less frequently. I0419 01:58:26.892829 137893174937408 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0419 01:58:26.894867 137893174937408 event_tracking.py:70] [process=6] [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_false/checkpoints/9. I0419 01:58:27.618268 137893174937408 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 01:58:27.618359 137893174937408 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 01:58:27.659558 137893174937408 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042903s I0419 01:58:27.659746 137893174937408 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 30.763 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05014324188232422 s) (per-host) I0419 01:58:27.659809 137893174937408 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.050222s (batch_requests_ready=0.002872s, total_serialization_initiated=0.047261s, others=0.000090s) I0419 01:58:27.659932 137893174937408 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.054911s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.054895) I0419 01:58:27.660743 137893174937408 event_tracking.py:125] [process=6] [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_false/checkpoints/9. I0419 01:58:27.661093 137762485303040 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:18:27.661055 I0419 01:58:27.663055 137893174937408 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:58:27.663379 137762967242496 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:58:27.663545 137893174937408 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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776563882.1896555, 'wait_for_prev_duration_secs': 24.70126700401306, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776563906.8928692, 'checkpointer_blocking_duration_secs': 0.768383264541626, 'get_old_steps_start_time': 1776563907.6612751, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776563882.1877446, 'checkpoint_manager_blocking_duration_secs': 25.47576332092285} I0419 01:58:27.663653 137893174937408 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 01:58:27.663699 137893174937408 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:58:34.431335 137741144684288 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9/items/array_metadatas/process_6 I0419 01:59:09.916587 137762485303040 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.337 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.30694890022278 s) (per-host) I0419 01:59:09.916721 137762485303040 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.255510s. I0419 01:59:18.857838 137762485303040 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.196612s. I0419 01:59:18.858112 137762967242496 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:59:18.858230 137762967242496 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:59:18.858305 137762967242496 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:59:18.862891 137762967242496 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 01:59:18.863082 137893174937408 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. I0419 01:59:18.863278 137893174937408 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 01:59:18.864275 137893174937408 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.572, Tokens/s/device: 13802.678, 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 01:59:27 UTC 2026 EXIT_CODE=0