XPK Start: Thu Apr 16 22:50:03 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-16 22:50:28.004547: 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) I0416 22:50:28.184663 134676050806592 max_utils.py:273] Attempting to initialize the jax distributed system... I0416 22:50:37.225924 134676050806592 distributed.py:149] Starting JAX distributed service on [::]:8482 I0416 22:50:37.228166 134676050806592 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-j8esq-slice-job-0-0.mt-10-shardy-false-j8esq:8482 I0416 22:50:37.968739 134676050806592 max_utils.py:284] Jax distributed system initialized! I0416 22:50:44.213346 134676050806592 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 22:50:44.213448 134676050806592 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 22:50:44.213487 134676050806592 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 I0416 22:50:44.213521 134676050806592 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 22:50:44.891622 134676050806592 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 22:50:45.006880 134676050806592 checkpointing.py:688] Setting up checkpoint logger... I0416 22:50:45.007013 134676050806592 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 22:50:45.007072 134676050806592 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0416 22:50:45.007330 134676050806592 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 0x7a7bfe844530>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 22:50:48.300603 134676050806592 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0416 22:50:48.300854 134676050806592 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 0x7a66d07343b0>}, handler_registry=None I0416 22:50:48.301092 134676050806592 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a66d07343b0>` 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`. I0416 22:50:48.301139 134676050806592 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a66d0736330>` 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`. I0416 22:50:48.301175 134676050806592 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 0x7a66d07343b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a66d07343b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a66d0736330>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a66d0736330>}). I0416 22:50:48.301490 134676050806592 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0416 22:50:48.301560 134676050806592 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a66d07d2020> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0416 22:50:49.290506 134676050806592 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints I0416 22:50:50.082937 134676050806592 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a67b861ec90> I0416 22:50:50.083108 134676050806592 checkpointing.py:302] Checkpoint manager created! I0416 22:50:51.047057 134676050806592 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0416 22:50:51.047170 134676050806592 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0416 22:50:52.370670 134676050806592 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0416 22:50:52.370782 134676050806592 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0416 22:50:52.370853 134676050806592 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0416 22:50:52.370901 134676050806592 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0416 22:50:52.370939 134676050806592 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0416 22:50:52.370975 134676050806592 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0416 22:50:52.371012 134676050806592 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 22:50:52.371048 134676050806592 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0416 22:50:52.371084 134676050806592 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 22:50:52.371117 134676050806592 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 22:50:52.371149 134676050806592 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0416 22:50:52.371180 134676050806592 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0416 22:50:52.771448 134676050806592 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 22:50:52.771537 134676050806592 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:50:52.777708 134676050806592 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 22:50:52.777763 134676050806592 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:50:52.794043 134676050806592 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 22:50:52.794102 134676050806592 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:50:52.809802 134676050806592 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 22:50:52.809886 134676050806592 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0416 22:50:52.834184 134676050806592 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 22:50:52.834253 134676050806592 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:50:52.849974 134676050806592 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 22:50:52.850032 134676050806592 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:50:52.865775 134676050806592 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 22:50:52.865844 134676050806592 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:50:52.896164 134676050806592 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 22:50:52.896234 134676050806592 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0416 22:50:52.916213 134676050806592 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 22:50:52.916277 134676050806592 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0416 22:51:03.811338 134676050806592 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. I0416 22:51:03.813737 134676050806592 metric_logger.py:289] number parameters: 1.104 billion I0416 22:51:14.926139 134676050806592 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0416 22:51:15.078530 134676050806592 checkpointing.py:798] Waited 0.15237784385681152 seconds for step 0 to finish before starting checkpointing. I0416 22:51:15.080891 134676050806592 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 22:51:15.082622 134676050806592 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0416 22:51:15.084267 134676050806592 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints/0. I0416 22:51:15.586050 134676050806592 signaling_client.py:364] Using JaxDistributedSignalingClient I0416 22:51:15.587054 134676050806592 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0416 22:51:15.587110 134676050806592 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 I0416 22:51:15.858999 134676050806592 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273722s I0416 22:51:15.859172 134676050806592 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.456 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.282745361328125 s) (per-host) I0416 22:51:15.859226 134676050806592 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282809s (batch_requests_ready=0.002985s, total_serialization_initiated=0.279752s, others=0.000073s) I0416 22:51:15.859321 134676050806592 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.287312s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.287294) I0416 22:51:15.860138 134676050806592 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.78 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints/0. I0416 22:51:15.860476 134548251301632 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-16 23:11:15.860439 I0416 22:51:15.871807 134676050806592 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 22:51:15.872107 134546657892096 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0416 22:51:15.872258 134676050806592 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776379875.0808668, 'wait_for_prev_duration_secs': 8.392333984375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776379875.0826752, 'checkpointer_blocking_duration_secs': 0.7779579162597656, 'get_old_steps_start_time': 1776379875.8606582, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776379875.0789638, 'checkpoint_manager_blocking_duration_secs': 0.7932572364807129} I0416 22:51:15.872367 134676050806592 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0416 22:51:15.872434 134676050806592 max_utils.py:750] Memstats: After params initialized: I0416 22:51:15.872483 134676050806592 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0416 22:51:15.872515 134676050806592 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0416 22:51:15.872540 134676050806592 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0416 22:51:15.872564 134676050806592 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0416 22:51:16.195910 134676050806592 metric_logger.py:185] completed step: 0, seconds: 11.111, TFLOP/s/device: 1.223, Tokens/s/device: 184.317, total_weights: 65536, loss: 10.874 I0416 22:51:16.357762 134676050806592 metric_logger.py:185] completed step: 1, seconds: 1.268, TFLOP/s/device: 10.713, Tokens/s/device: 1614.824, total_weights: 65536, loss: 10.874 I0416 22:51:16.779450 134676050806592 metric_logger.py:185] completed step: 2, seconds: 0.015, TFLOP/s/device: 880.395, Tokens/s/device: 132702.650, total_weights: 65536, loss: 10.262 I0416 22:51:16.927599 134676050806592 metric_logger.py:185] completed step: 3, seconds: 0.412, TFLOP/s/device: 32.981, Tokens/s/device: 4971.320, total_weights: 65536, loss: 9.732 I0416 22:51:17.223027 134676050806592 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.701, Tokens/s/device: 12465.534, total_weights: 65536, loss: 9.274 I0416 22:51:17.230394 134676050806592 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.209, Tokens/s/device: 13898.692, total_weights: 65536, loss: 8.888 I0416 22:51:19.380167 2793 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0416 22:51:21.365533 134546666284800 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints/0/items/array_metadatas/process_6 I0416 22:51:39.849853 134676050806592 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.901, Tokens/s/device: 6918.638, total_weights: 65536, loss: 8.589 I0416 22:51:39.997394 134676050806592 metric_logger.py:185] completed step: 7, seconds: 22.473, TFLOP/s/device: 0.605, Tokens/s/device: 91.131, total_weights: 65536, loss: 8.380 I0416 22:51:40.144994 134676050806592 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.654, Tokens/s/device: 13362.913, total_weights: 65536, loss: 8.251 I0416 22:51:40.293560 134676050806592 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0416 22:51:40.294234 134676050806592 checkpointing.py:798] Waited 0.0006861686706542969 seconds for step 9 to finish before starting checkpointing. I0416 22:51:40.296245 134676050806592 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 22:51:52.185145 134548251301632 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 43.148 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.60867929458618 s) (per-host) I0416 22:51:52.185267 134548251301632 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.324671s. I0416 22:52:01.337399 134548251301632 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 45.476788s. I0416 22:52:01.337708 134546657892096 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 22:52:01.337850 134546657892096 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0416 22:52:01.337904 134546657892096 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 22:52:01.339756 134546657892096 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 22:52:01.339933 134676050806592 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. W0416 22:52:01.340064 134676050806592 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.043822 seconds. If this number is high, consider checkpointing less frequently. I0416 22:52:01.342256 134676050806592 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0416 22:52:01.344286 134676050806592 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints/9. I0416 22:52:02.054008 134676050806592 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0416 22:52:02.054101 134676050806592 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 I0416 22:52:02.092334 134676050806592 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039966s I0416 22:52:02.092500 134676050806592 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 32.773 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.047067880630493164 s) (per-host) I0416 22:52:02.092552 134676050806592 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.047132s (batch_requests_ready=0.002765s, total_serialization_initiated=0.044295s, others=0.000072s) I0416 22:52:02.092648 134676050806592 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051321s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.051306) I0416 22:52:02.093336 134676050806592 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints/9. I0416 22:52:02.093678 134546657892096 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-16 23:12:02.093640 I0416 22:52:02.095661 134676050806592 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 22:52:02.095942 134546666284800 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0416 22:52:02.096097 134676050806592 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776379900.2962143, 'wait_for_prev_duration_secs': 21.043821811676025, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776379921.3422947, 'checkpointer_blocking_duration_secs': 0.7515487670898438, 'get_old_steps_start_time': 1776379922.0938687, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776379900.2944958, 'checkpoint_manager_blocking_duration_secs': 21.80156922340393} I0416 22:52:02.096202 134676050806592 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0416 22:52:02.096245 134676050806592 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 22:52:07.977235 134525594089216 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_10_shardy_false/checkpoints/9/items/array_metadatas/process_6 I0416 22:52:43.521343 134546657892096 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.085 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.47587513923645 s) (per-host) I0416 22:52:43.521461 134546657892096 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.427672s. I0416 22:52:51.641195 134546657892096 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.547390s. I0416 22:52:51.641473 134546666284800 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 22:52:51.641586 134546666284800 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0416 22:52:51.641636 134546666284800 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 22:52:51.643226 134546666284800 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 22:52:51.643412 134676050806592 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. I0416 22:52:51.643582 134676050806592 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 22:52:51.644253 134676050806592 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.984, Tokens/s/device: 13864.818, 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: Thu Apr 16 22:53:04 UTC 2026 EXIT_CODE=0