XPK Start: Sat Apr 25 13:44:25 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. 2026-04-25 13:44:50.198023: 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) I0425 13:44:50.410767 134349341529920 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 13:44:59.451706 134349341529920 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 13:44:59.454424 134349341529920 distributed.py:172] Connecting to JAX distributed service on mt-03-dropout-ukpqm-slice-job-0-0.mt-03-dropout-ukpqm:8482 I0425 13:45:00.330082 134349341529920 max_utils.py:284] Jax distributed system initialized! I0425 13:45:06.569307 134349341529920 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 13:45:06.569412 134349341529920 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 13:45:06.569452 134349341529920 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 I0425 13:45:06.569488 134349341529920 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 13:45:07.269907 134349341529920 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 13:45:07.383685 134349341529920 checkpointing.py:688] Setting up checkpoint logger... I0425 13:45:07.383806 134349341529920 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 13:45:07.383855 134349341529920 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 13:45:07.384056 134349341529920 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 0x7a300df1c230>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 13:45:10.320766 134349341529920 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 13:45:10.321002 134349341529920 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 0x7a1b3056cfb0>}, handler_registry=None I0425 13:45:10.321340 134349341529920 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1b3056cfb0>` 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`. I0425 13:45:10.321398 134349341529920 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1b3056b410>` 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`. I0425 13:45:10.321442 134349341529920 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 0x7a1b3056cfb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1b3056cfb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1b3056b410>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1b3056b410>}). I0425 13:45:10.321805 134349341529920 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 13:45:10.321912 134349341529920 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a1b086a36a0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 13:45:11.425864 134349341529920 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints I0425 13:45:11.621852 134349341529920 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_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a1b3056d340> I0425 13:45:11.622030 134349341529920 checkpointing.py:302] Checkpoint manager created! I0425 13:45:12.006748 134349341529920 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 13:45:12.006873 134349341529920 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 13:45:13.412911 134349341529920 maxtext_utils.py:1889] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 13:45:13.413033 134349341529920 maxtext_utils.py:1889] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 13:45:13.413085 134349341529920 maxtext_utils.py:1889] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 13:45:13.413146 134349341529920 maxtext_utils.py:1889] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 13:45:13.413185 134349341529920 maxtext_utils.py:1889] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 13:45:13.413220 134349341529920 maxtext_utils.py:1889] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 13:45:13.413256 134349341529920 maxtext_utils.py:1889] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 13:45:13.413292 134349341529920 maxtext_utils.py:1889] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 13:45:13.413326 134349341529920 maxtext_utils.py:1889] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 13:45:13.413357 134349341529920 maxtext_utils.py:1889] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 13:45:13.413388 134349341529920 maxtext_utils.py:1889] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 13:45:13.413419 134349341529920 maxtext_utils.py:1889] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 13:45:13.824132 134349341529920 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 13:45:13.824238 134349341529920 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 13:45:13.830172 134349341529920 nnx_decoders.py:465] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 13:45:13.830229 134349341529920 nnx_decoders.py:465] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 13:45:13.847007 134349341529920 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 13:45:13.847070 134349341529920 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 13:45:13.863114 134349341529920 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 13:45:13.863172 134349341529920 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 13:45:13.887596 134349341529920 attentions.py:1154] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 13:45:13.887667 134349341529920 attentions.py:1154] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 13:45:13.903882 134349341529920 attentions.py:1155] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 13:45:13.903943 134349341529920 attentions.py:1155] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 13:45:13.920316 134349341529920 attentions.py:1156] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 13:45:13.920385 134349341529920 attentions.py:1156] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 13:45:13.951034 134349341529920 attentions.py:1198] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0425 13:45:13.951132 134349341529920 attentions.py:1198] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 13:45:13.977564 134349341529920 linears.py:525] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 13:45:13.977636 134349341529920 linears.py:525] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None). I0425 13:45:17.707722 134349341529920 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. I0425 13:45:17.710084 134349341529920 metric_logger.py:301] number parameters: 1.104 billion I0425 13:45:22.150539 134349341529920 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 13:45:22.223249 134349341529920 checkpointing.py:798] Waited 0.07270121574401855 seconds for step 0 to finish before starting checkpointing. I0425 13:45:22.225949 134349341529920 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 13:45:22.227936 134349341529920 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0425 13:45:22.229403 134349341529920 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints/0. I0425 13:45:22.512954 134349341529920 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 13:45:22.513828 134349341529920 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 13:45:22.513887 134349341529920 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 I0425 13:45:22.794270 134349341529920 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.282102s I0425 13:45:22.794447 134349341529920 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.291 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2915503978729248 s) (per-host) I0425 13:45:22.794500 134349341529920 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.291613s (batch_requests_ready=0.003260s, total_serialization_initiated=0.288283s, others=0.000070s) I0425 13:45:22.794600 134349341529920 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.295642s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.295625) I0425 13:45:22.795423 134349341529920 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.57 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints/0. I0425 13:45:22.795759 134222489671424 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:05:22.795721 I0425 13:45:22.798983 134349341529920 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 13:45:22.799231 134220315879168 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 13:45:22.799358 134349341529920 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777124722.225931, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777124722.2279742, 'checkpointer_blocking_duration_secs': 0.567934513092041, 'get_old_steps_start_time': 1777124722.7959344, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1777124722.2236364, 'checkpoint_manager_blocking_duration_secs': 0.5756855010986328} I0425 13:45:22.799468 134349341529920 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 13:45:22.799538 134349341529920 max_utils.py:750] Memstats: After params initialized: I0425 13:45:22.799593 134349341529920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_24(process=6,(0,6,0,0)) I0425 13:45:22.799627 134349341529920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_25(process=6,(1,6,0,0)) I0425 13:45:22.799656 134349341529920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_28(process=6,(0,7,0,0)) I0425 13:45:22.799682 134349341529920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_29(process=6,(1,7,0,0)) I0425 13:45:23.115800 134349341529920 metric_logger.py:196] completed step: 0, seconds: 4.439, TFLOP/s/device: 0.180, Tokens/s/device: 28.833, total_weights: 4096, loss: 10.874, lm_loss: 10.874, perplexity: 52805.141 I0425 13:45:23.291440 134349341529920 metric_logger.py:196] completed step: 1, seconds: 0.963, TFLOP/s/device: 0.831, Tokens/s/device: 132.864, total_weights: 4096, loss: 10.892, lm_loss: 10.892, perplexity: 53764.488 I0425 13:45:23.706500 134349341529920 metric_logger.py:196] completed step: 2, seconds: 0.107, TFLOP/s/device: 7.476, Tokens/s/device: 1194.799, total_weights: 4096, loss: 9.969, lm_loss: 9.969, perplexity: 21350.100 I0425 13:45:23.776988 134349341529920 metric_logger.py:196] completed step: 3, seconds: 0.404, TFLOP/s/device: 1.981, Tokens/s/device: 316.601, total_weights: 4096, loss: 9.125, lm_loss: 9.125, perplexity: 9186.262 I0425 13:45:23.919998 134349341529920 metric_logger.py:196] completed step: 4, seconds: 0.087, TFLOP/s/device: 9.169, Tokens/s/device: 1465.403, total_weights: 4096, loss: 8.469, lm_loss: 8.469, perplexity: 4765.735 I0425 13:45:23.927993 134349341529920 metric_logger.py:196] completed step: 5, seconds: 0.070, TFLOP/s/device: 11.466, Tokens/s/device: 1832.472, total_weights: 4096, loss: 7.908, lm_loss: 7.908, perplexity: 2719.938 I0425 13:45:25.843542 2754 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 13:45:27.866405 134221917193984 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_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints/0/items/array_metadatas/process_6 I0425 13:45:33.606202 134349341529920 metric_logger.py:196] completed step: 6, seconds: 0.144, TFLOP/s/device: 5.553, Tokens/s/device: 887.428, total_weights: 4096, loss: 7.554, lm_loss: 7.554, perplexity: 1909.063 I0425 13:45:33.676947 134349341529920 metric_logger.py:196] completed step: 7, seconds: 9.609, TFLOP/s/device: 0.083, Tokens/s/device: 13.321, total_weights: 4096, loss: 7.313, lm_loss: 7.313, perplexity: 1500.206 I0425 13:45:33.747499 134349341529920 metric_logger.py:196] completed step: 8, seconds: 0.076, TFLOP/s/device: 10.509, Tokens/s/device: 1679.592, total_weights: 4096, loss: 7.091, lm_loss: 7.091, perplexity: 1200.912 I0425 13:45:34.115219 134349341529920 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 13:45:34.115878 134349341529920 checkpointing.py:798] Waited 0.0006709098815917969 seconds for step 9 to finish before starting checkpointing. I0425 13:45:34.117804 134349341529920 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 13:45:54.519158 134222489671424 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 49.337 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 32.016242027282715 s) (per-host) I0425 13:45:54.519262 134222489671424 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 31.723394s. I0425 13:46:03.287935 134222489671424 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 40.492050s. I0425 13:46:03.288194 134220315879168 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 13:46:03.288252 134220315879168 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 13:46:03.288309 134220315879168 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 13:46:03.290039 134220315879168 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 13:46:03.290215 134349341529920 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. W0425 13:46:03.290351 134349341529920 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.172547 seconds. If this number is high, consider checkpointing less frequently. I0425 13:46:03.292237 134349341529920 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0425 13:46:03.294345 134349341529920 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints/9. I0425 13:46:03.581478 134349341529920 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 13:46:03.581577 134349341529920 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 I0425 13:46:03.619714 134349341529920 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039767s I0425 13:46:03.619884 134349341529920 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 32.697 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04717755317687988 s) (per-host) I0425 13:46:03.619931 134349341529920 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.047235s (batch_requests_ready=0.002959s, total_serialization_initiated=0.044212s, others=0.000064s) I0425 13:46:03.620027 134349341529920 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051480s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.051464) I0425 13:46:03.620780 134349341529920 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints/9. I0425 13:46:03.621077 134220315879168 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:06:03.621043 I0425 13:46:03.625722 134349341529920 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 13:46:03.626001 134222458832640 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 13:46:03.626153 134349341529920 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777124734.117774, 'wait_for_prev_duration_secs': 29.172546863555908, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777124763.2922757, 'checkpointer_blocking_duration_secs': 0.32895612716674805, 'get_old_steps_start_time': 1777124763.6212544, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1777124734.1161277, 'checkpoint_manager_blocking_duration_secs': 29.50999140739441} I0425 13:46:03.626315 134349341529920 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 13:46:03.626360 134349341529920 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 13:46:09.050636 134196857009920 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_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_03_dropout/checkpoints/9/items/array_metadatas/process_6 I0425 13:46:43.738799 134220315879168 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 39.327 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.1660532951355 s) (per-host) I0425 13:46:43.738930 134220315879168 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.117725s. I0425 13:46:51.655645 134220315879168 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.034425s. I0425 13:46:51.655959 134222458832640 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 13:46:51.656089 134222458832640 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 13:46:51.656156 134222458832640 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 13:46:51.657557 134222458832640 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 13:46:51.657730 134349341529920 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. I0425 13:46:51.657902 134349341529920 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 13:46:51.658884 134349341529920 metric_logger.py:196] completed step: 9, seconds: 0.071, TFLOP/s/device: 11.273, Tokens/s/device: 1801.700, 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: Sat Apr 25 13:47:00 UTC 2026 EXIT_CODE=0