XPK Start: Fri Apr 24 13:37:39 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. `rope_parameters`'s factor field must be a float >= 1, got 40 `rope_parameters`'s beta_fast field must be a float, got 32 `rope_parameters`'s beta_slow field must be a float, got 1 DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 2026-04-24 13:38:03.948592: 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) I0424 13:38:04.159093 134473351382848 max_utils.py:273] Attempting to initialize the jax distributed system... I0424 13:38:13.200778 134473351382848 distributed.py:149] Starting JAX distributed service on [::]:8482 I0424 13:38:13.203642 134473351382848 distributed.py:172] Connecting to JAX distributed service on mt-03-dropout-7j2cz-slice-job-0-0.mt-03-dropout-7j2cz:8482 I0424 13:38:14.206371 134473351382848 max_utils.py:284] Jax distributed system initialized! I0424 13:38:19.224672 134473351382848 max_utils.py:800] System Information: Jax Version: 0.9.2 I0424 13:38:19.224779 134473351382848 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0424 13:38:19.224820 134473351382848 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 I0424 13:38:19.224854 134473351382848 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 13:38:19.918712 134473351382848 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 13:38:20.032389 134473351382848 checkpointing.py:688] Setting up checkpoint logger... I0424 13:38:20.032512 134473351382848 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 13:38:20.032557 134473351382848 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0424 13:38:20.032763 134473351382848 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 0x7a4cccb381d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 13:38:23.028967 134473351382848 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 13:38:23.029211 134473351382848 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 0x7a3974363bf0>}, handler_registry=None I0424 13:38:23.029519 134473351382848 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3974363bf0>` 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`. I0424 13:38:23.029575 134473351382848 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a39544a7a40>` 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`. I0424 13:38:23.029613 134473351382848 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 0x7a3974363bf0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3974363bf0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a39544a7a40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a39544a7a40>}). I0424 13:38:23.029968 134473351382848 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0424 13:38:23.030042 134473351382848 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a395444f6a0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0424 13:38:23.787098 134473351382848 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints I0424 13:38:23.829063 134473351382848 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_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a3954462b70> I0424 13:38:23.829217 134473351382848 checkpointing.py:302] Checkpoint manager created! I0424 13:38:24.206759 134473351382848 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 13:38:24.206876 134473351382848 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 13:38:25.598876 134473351382848 maxtext_utils.py:1889] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0424 13:38:25.599006 134473351382848 maxtext_utils.py:1889] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 13:38:25.599060 134473351382848 maxtext_utils.py:1889] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0424 13:38:25.599104 134473351382848 maxtext_utils.py:1889] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0424 13:38:25.599142 134473351382848 maxtext_utils.py:1889] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 13:38:25.599177 134473351382848 maxtext_utils.py:1889] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0424 13:38:25.599213 134473351382848 maxtext_utils.py:1889] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 13:38:25.599247 134473351382848 maxtext_utils.py:1889] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0424 13:38:25.599280 134473351382848 maxtext_utils.py:1889] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 13:38:25.599313 134473351382848 maxtext_utils.py:1889] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0424 13:38:25.599345 134473351382848 maxtext_utils.py:1889] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0424 13:38:25.599377 134473351382848 maxtext_utils.py:1889] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0424 13:38:26.006907 134473351382848 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 13:38:26.007016 134473351382848 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0424 13:38:26.012840 134473351382848 nnx_decoders.py:465] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 13:38:26.012898 134473351382848 nnx_decoders.py:465] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0424 13:38:26.029320 134473351382848 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 13:38:26.029380 134473351382848 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0424 13:38:26.045274 134473351382848 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 13:38:26.045334 134473351382848 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0424 13:38:26.069487 134473351382848 attentions.py:1154] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 13:38:26.069555 134473351382848 attentions.py:1154] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0424 13:38:26.085603 134473351382848 attentions.py:1155] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 13:38:26.085665 134473351382848 attentions.py:1155] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0424 13:38:26.101686 134473351382848 attentions.py:1156] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 13:38:26.101749 134473351382848 attentions.py:1156] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0424 13:38:26.131406 134473351382848 attentions.py:1198] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0424 13:38:26.131476 134473351382848 attentions.py:1198] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0424 13:38:26.157373 134473351382848 linears.py:525] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 13:38:26.157438 134473351382848 linears.py:525] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None). I0424 13:38:29.975105 134473351382848 max_utils.py:791] Total memory size: 0.8 GB, Output size: 0.4 GB, Temp size: 0.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0424 13:38:29.977382 134473351382848 metric_logger.py:301] number parameters: 1.104 billion I0424 13:38:34.377963 134473351382848 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 13:38:34.452277 134473351382848 checkpointing.py:798] Waited 0.07431888580322266 seconds for step 0 to finish before starting checkpointing. I0424 13:38:34.454755 134473351382848 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:38:34.456974 134473351382848 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0 I0424 13:38:34.458447 134473351382848 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints/0. I0424 13:38:34.754289 134473351382848 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 13:38:34.755163 134473351382848 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0424 13:38:34.755223 134473351382848 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 I0424 13:38:35.034003 134473351382848 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.280490s I0424 13:38:35.034188 134473351382848 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.325 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2897012233734131 s) (per-host) I0424 13:38:35.034260 134473351382848 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.289780s (batch_requests_ready=0.003040s, total_serialization_initiated=0.286653s, others=0.000087s) I0424 13:38:35.034385 134473351382848 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.294019s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.294001) I0424 13:38:35.035249 134473351382848 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.58 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints/0. I0424 13:38:35.035586 134347717867264 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:58:35.035547 I0424 13:38:35.044974 134473351382848 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:38:35.045221 134346644125440 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:38:35.045375 134473351382848 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777037914.4547362, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777037914.4570124, 'checkpointer_blocking_duration_secs': 0.5787355899810791, 'get_old_steps_start_time': 1777037915.0357723, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1777037914.452679, 'checkpoint_manager_blocking_duration_secs': 0.5926601886749268} I0424 13:38:35.045489 134473351382848 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 13:38:35.045560 134473351382848 max_utils.py:750] Memstats: After params initialized: I0424 13:38:35.045608 134473351382848 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_10(process=3,(2,2,0,0)) I0424 13:38:35.045640 134473351382848 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_11(process=3,(3,2,0,0)) I0424 13:38:35.045667 134473351382848 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_14(process=3,(2,3,0,0)) I0424 13:38:35.045691 134473351382848 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_15(process=3,(3,3,0,0)) I0424 13:38:35.363244 134473351382848 metric_logger.py:196] completed step: 0, seconds: 4.400, TFLOP/s/device: 0.182, Tokens/s/device: 29.094, total_weights: 4096, loss: 10.874, lm_loss: 10.874, perplexity: 52805.141 I0424 13:38:35.464768 134473351382848 metric_logger.py:196] completed step: 1, seconds: 0.983, TFLOP/s/device: 0.814, Tokens/s/device: 130.169, total_weights: 4096, loss: 10.892, lm_loss: 10.892, perplexity: 53764.488 I0424 13:38:35.876653 134473351382848 metric_logger.py:196] completed step: 2, seconds: 0.032, TFLOP/s/device: 24.891, Tokens/s/device: 3978.120, total_weights: 4096, loss: 9.969, lm_loss: 9.969, perplexity: 21350.100 I0424 13:38:35.947105 134473351382848 metric_logger.py:196] completed step: 3, seconds: 0.402, TFLOP/s/device: 1.994, Tokens/s/device: 318.639, total_weights: 4096, loss: 9.125, lm_loss: 9.125, perplexity: 9186.262 I0424 13:38:36.089014 134473351382848 metric_logger.py:196] completed step: 4, seconds: 0.088, TFLOP/s/device: 9.147, Tokens/s/device: 1461.922, total_weights: 4096, loss: 8.469, lm_loss: 8.469, perplexity: 4765.735 I0424 13:38:36.096390 134473351382848 metric_logger.py:196] completed step: 5, seconds: 0.070, TFLOP/s/device: 11.447, Tokens/s/device: 1829.565, total_weights: 4096, loss: 7.908, lm_loss: 7.908, perplexity: 2719.938 I0424 13:38:38.239210 2741 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 13:38:40.260019 134347659372288 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_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints/0/items/array_metadatas/process_3 I0424 13:38:45.815497 134473351382848 metric_logger.py:196] completed step: 6, seconds: 0.143, TFLOP/s/device: 5.620, Tokens/s/device: 898.176, total_weights: 4096, loss: 7.554, lm_loss: 7.554, perplexity: 1909.063 I0424 13:38:45.885949 134473351382848 metric_logger.py:196] completed step: 7, seconds: 9.650, TFLOP/s/device: 0.083, Tokens/s/device: 13.265, total_weights: 4096, loss: 7.313, lm_loss: 7.313, perplexity: 1500.206 I0424 13:38:45.956644 134473351382848 metric_logger.py:196] completed step: 8, seconds: 0.076, TFLOP/s/device: 10.521, Tokens/s/device: 1681.489, total_weights: 4096, loss: 7.091, lm_loss: 7.091, perplexity: 1200.912 I0424 13:38:46.027861 134473351382848 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 13:38:46.028523 134473351382848 checkpointing.py:798] Waited 0.00067138671875 seconds for step 9 to finish before starting checkpointing. I0424 13:38:46.030872 134473351382848 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:39:07.703812 134347717867264 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 47.926 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 32.95928621292114 s) (per-host) I0424 13:39:07.703949 134347717867264 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.668253s. I0424 13:39:16.730671 134347717867264 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 41.694959s. I0424 13:39:16.731026 134346644125440 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:39:16.731159 134346644125440 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:39:16.731216 134346644125440 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:39:16.732994 134346644125440 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 13:39:16.733186 134473351382848 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. W0424 13:39:16.733335 134473351382848 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.702468 seconds. If this number is high, consider checkpointing less frequently. I0424 13:39:16.735266 134473351382848 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9 I0424 13:39:16.737247 134473351382848 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints/9. I0424 13:39:17.424666 134473351382848 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0424 13:39:17.424764 134473351382848 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 I0424 13:39:17.458544 134473351382848 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035417s I0424 13:39:17.458709 134473351382848 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 36.230 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04257774353027344 s) (per-host) I0424 13:39:17.458764 134473351382848 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042643s (batch_requests_ready=0.002749s, total_serialization_initiated=0.039822s, others=0.000073s) I0424 13:39:17.458862 134473351382848 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.046828s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.046813) I0424 13:39:17.459566 134473351382848 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.72 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints/9. I0424 13:39:17.459915 134346644125440 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:59:17.459877 I0424 13:39:17.468621 134473351382848 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:39:17.468894 134347659372288 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:39:17.469067 134473351382848 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777037926.0308387, 'wait_for_prev_duration_secs': 30.702467679977417, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777037956.7353058, 'checkpointer_blocking_duration_secs': 0.7247638702392578, 'get_old_steps_start_time': 1777037957.4600923, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1777037926.028754, 'checkpoint_manager_blocking_duration_secs': 31.440277576446533} I0424 13:39:17.469177 134473351382848 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 13:39:17.469222 134473351382848 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:39:23.105343 134320983320320 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_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_03_dropout/checkpoints/9/items/array_metadatas/process_3 I0424 13:40:00.031383 134346644125440 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.066 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.61521053314209 s) (per-host) I0424 13:40:00.031510 134346644125440 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.571468s. I0424 13:40:06.096348 134346644125440 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.636291s. I0424 13:40:06.096657 134347659372288 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:40:06.096790 134347659372288 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:40:06.096841 134347659372288 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:40:06.098746 134347659372288 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 13:40:06.098917 134473351382848 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. I0424 13:40:06.099111 134473351382848 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:40:06.100084 134473351382848 metric_logger.py:196] completed step: 9, seconds: 0.071, TFLOP/s/device: 11.307, Tokens/s/device: 1807.195, total_weights: 4096, loss: 7.016, lm_loss: 7.016, perplexity: 1113.843 Per train step: Total TFLOPs: 0.80 split as 99.60% learnable weight flops and 0.40% attention flops XPK End: Fri Apr 24 13:40:19 UTC 2026 EXIT_CODE=0