XPK Start: Fri Apr 17 00:59:40 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-17 01:00:04.322645: 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) I0417 01:00:04.501131 132609569302336 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 01:00:13.542745 132609569302336 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 01:00:13.545085 132609569302336 distributed.py:172] Connecting to JAX distributed service on mt-07-eval-wz5eo-slice-job-0-0.mt-07-eval-wz5eo:8482 I0417 01:00:14.843063 132609569302336 max_utils.py:284] Jax distributed system initialized! I0417 01:00:20.937658 132609569302336 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 01:00:20.937758 132609569302336 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 01:00:20.937800 132609569302336 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 I0417 01:00:20.937851 132609569302336 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 01:00:21.617935 132609569302336 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 01:00:21.730664 132609569302336 checkpointing.py:688] Setting up checkpoint logger... I0417 01:00:21.730776 132609569302336 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 01:00:21.730832 132609569302336 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 01:00:21.731035 132609569302336 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 0x789ada84c0b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 01:00:25.075857 132609569302336 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 01:00:25.076097 132609569302336 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 0x7886dc10bb30>}, handler_registry=None I0417 01:00:25.076336 132609569302336 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7886dc10bb30>` 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`. I0417 01:00:25.076385 132609569302336 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78870415b6b0>` 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`. I0417 01:00:25.076420 132609569302336 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 0x7886dc10bb30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7886dc10bb30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78870415b6b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78870415b6b0>}). I0417 01:00:25.076751 132609569302336 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 01:00:25.076838 132609569302336 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7886dc77e020> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 01:00:26.628124 132609569302336 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints I0417 01:00:27.077466 132609569302336 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_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7886dc71bf80> I0417 01:00:27.077648 132609569302336 checkpointing.py:302] Checkpoint manager created! I0417 01:00:28.041659 132609569302336 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 01:00:28.041773 132609569302336 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 01:00:29.398923 132609569302336 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0417 01:00:29.399038 132609569302336 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 01:00:29.399089 132609569302336 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 01:00:29.399132 132609569302336 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0417 01:00:29.399170 132609569302336 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 01:00:29.399205 132609569302336 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 01:00:29.399241 132609569302336 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:00:29.399277 132609569302336 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0417 01:00:29.399311 132609569302336 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:00:29.399343 132609569302336 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:00:29.399375 132609569302336 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0417 01:00:29.399406 132609569302336 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0417 01:00:29.800064 132609569302336 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 01:00:29.800161 132609569302336 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:00:29.805419 132609569302336 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 01:00:29.805475 132609569302336 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:00:29.821873 132609569302336 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 01:00:29.821939 132609569302336 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:00:29.838037 132609569302336 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 01:00:29.838105 132609569302336 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:00:29.862203 132609569302336 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 01:00:29.862268 132609569302336 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:00:29.878222 132609569302336 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 01:00:29.878281 132609569302336 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:00:29.894064 132609569302336 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 01:00:29.894134 132609569302336 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:00:29.924600 132609569302336 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 01:00:29.924673 132609569302336 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:00:29.944483 132609569302336 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 01:00:29.944553 132609569302336 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 01:00:41.230422 132609569302336 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. I0417 01:00:41.232772 132609569302336 metric_logger.py:289] number parameters: 1.104 billion I0417 01:00:52.365467 132609569302336 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 01:00:52.517675 132609569302336 checkpointing.py:798] Waited 0.1521909236907959 seconds for step 0 to finish before starting checkpointing. I0417 01:00:52.520471 132609569302336 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 01:00:52.522173 132609569302336 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0417 01:00:52.523627 132609569302336 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints/0. I0417 01:00:53.260710 132609569302336 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 01:00:53.261703 132609569302336 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 01:00:53.261762 132609569302336 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 I0417 01:00:53.541926 132609569302336 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281985s I0417 01:00:53.542100 132609569302336 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.299 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29108691215515137 s) (per-host) I0417 01:00:53.542152 132609569302336 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.291150s (batch_requests_ready=0.002987s, total_serialization_initiated=0.288092s, others=0.000072s) I0417 01:00:53.542240 132609569302336 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.295226s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.295208) I0417 01:00:53.543061 132609569302336 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints/0. I0417 01:00:53.543360 132482091448064 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 01:20:53.543325 I0417 01:00:53.546907 132609569302336 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 01:00:53.547160 132481546184448 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0417 01:00:53.547291 132609569302336 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776387652.5204518, 'wait_for_prev_duration_secs': 5.984306335449219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776387652.5222137, 'checkpointer_blocking_duration_secs': 1.0212907791137695, 'get_old_steps_start_time': 1776387653.5435305, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776387652.5181544, 'checkpoint_manager_blocking_duration_secs': 1.0290968418121338} I0417 01:00:53.547405 132609569302336 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 01:00:53.547475 132609569302336 max_utils.py:750] Memstats: After params initialized: I0417 01:00:53.547525 132609569302336 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0417 01:00:53.547561 132609569302336 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0417 01:00:53.547605 132609569302336 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0417 01:00:53.547633 132609569302336 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0417 01:00:53.864892 132609569302336 metric_logger.py:185] completed step: 0, seconds: 11.132, TFLOP/s/device: 1.221, Tokens/s/device: 183.981, total_weights: 65536, loss: 10.874 I0417 01:00:54.030404 132609569302336 metric_logger.py:185] completed step: 1, seconds: 1.498, TFLOP/s/device: 9.072, Tokens/s/device: 1367.427, total_weights: 65536, loss: 10.874 I0417 01:00:54.465145 132609569302336 metric_logger.py:185] completed step: 2, seconds: 0.020, TFLOP/s/device: 695.777, Tokens/s/device: 104875.051, total_weights: 65536, loss: 10.262 I0417 01:00:54.612720 132609569302336 metric_logger.py:185] completed step: 3, seconds: 0.425, TFLOP/s/device: 31.984, Tokens/s/device: 4820.990, total_weights: 65536, loss: 9.732 I0417 01:00:54.909089 132609569302336 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.844, Tokens/s/device: 12487.120, total_weights: 65536, loss: 9.274 I0417 01:00:54.917001 132609569302336 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.280, Tokens/s/device: 13909.453, total_weights: 65536, loss: 8.888 I0417 01:00:56.441266 2776 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 01:00:58.852479 132481554577152 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_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints/0/items/array_metadatas/process_6 I0417 01:01:17.374871 132609569302336 metric_logger.py:185] completed step: 6, seconds: 0.297, TFLOP/s/device: 45.702, Tokens/s/device: 6888.665, total_weights: 65536, loss: 8.589 I0417 01:01:17.522342 132609569302336 metric_logger.py:185] completed step: 7, seconds: 22.312, TFLOP/s/device: 0.609, Tokens/s/device: 91.791, total_weights: 65536, loss: 8.380 I0417 01:01:17.669808 132609569302336 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.665, Tokens/s/device: 13364.482, total_weights: 65536, loss: 8.251 I0417 01:01:17.818496 132609569302336 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 01:01:17.819211 132609569302336 checkpointing.py:798] Waited 0.0007259845733642578 seconds for step 9 to finish before starting checkpointing. I0417 01:01:17.821164 132609569302336 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 01:01:29.942031 132482091448064 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 43.051 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.69098114967346 s) (per-host) I0417 01:01:29.942153 132482091448064 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.398685s. I0417 01:01:39.039814 132482091448064 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 45.496328s. I0417 01:01:39.040129 132481546184448 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 01:01:39.040257 132481546184448 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0417 01:01:39.040308 132481546184448 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 01:01:39.043299 132481546184448 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 01:01:39.043479 132609569302336 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. W0417 01:01:39.043621 132609569302336 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.222455 seconds. If this number is high, consider checkpointing less frequently. I0417 01:01:39.045479 132609569302336 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0417 01:01:39.047640 132609569302336 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints/9. I0417 01:01:39.760931 132609569302336 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 01:01:39.761023 132609569302336 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 I0417 01:01:39.798002 132609569302336 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038724s I0417 01:01:39.798172 132609569302336 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.727 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.045737266540527344 s) (per-host) I0417 01:01:39.798225 132609569302336 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045799s (batch_requests_ready=0.002686s, total_serialization_initiated=0.043043s, others=0.000070s) I0417 01:01:39.798356 132609569302336 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049978s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.049961) I0417 01:01:39.799104 132609569302336 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints/9. I0417 01:01:39.799383 132481546184448 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 01:21:39.799351 I0417 01:01:39.801298 132609569302336 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 01:01:39.801529 132456044816128 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0417 01:01:39.801674 132609569302336 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776387677.8211343, 'wait_for_prev_duration_secs': 21.22245502471924, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776387699.0455167, 'checkpointer_blocking_duration_secs': 0.7540035247802734, 'get_old_steps_start_time': 1776387699.7995477, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776387677.81945, 'checkpoint_manager_blocking_duration_secs': 21.982191801071167} I0417 01:01:39.801789 132609569302336 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 01:01:39.801846 132609569302336 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 01:01:46.200587 132481554577152 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_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_07_eval/checkpoints/9/items/array_metadatas/process_6 I0417 01:02:22.071752 132481546184448 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.326 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.31929683685303 s) (per-host) I0417 01:02:22.071869 132481546184448 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.272377s. I0417 01:02:30.855798 132481546184448 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.056288s. I0417 01:02:30.856114 132456044816128 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 01:02:30.856237 132456044816128 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0417 01:02:30.856286 132456044816128 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 01:02:30.858021 132456044816128 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 01:02:30.858202 132609569302336 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. I0417 01:02:30.858374 132609569302336 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 01:02:30.858974 132609569302336 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.968, Tokens/s/device: 13862.378, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 17 01:02:41 UTC 2026 EXIT_CODE=0