XPK Start: Fri Apr 17 04:28:16 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 04:28:40.621125: 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 04:28:41.154580 139308337518400 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 04:28:50.195096 139308337518400 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 04:28:50.197692 139308337518400 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-22pl2-slice-job-0-0.mt-16-zarr3-false-sv-22pl2:8482 I0417 04:28:51.189390 139308337518400 max_utils.py:284] Jax distributed system initialized! I0417 04:28:57.225482 139308337518400 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 04:28:57.225582 139308337518400 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 04:28:57.225621 139308337518400 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 04:28:57.225659 139308337518400 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 04:28:57.893875 139308337518400 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 04:28:57.894473 139308337518400 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 04:28:57.894648 139308337518400 checkpointing.py:688] Setting up checkpoint logger... I0417 04:28:57.894697 139308337518400 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0417 04:28:57.894741 139308337518400 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 04:28:57.895102 139308337518400 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7eb29ef1ce60>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 04:29:01.248983 139308337518400 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 04:29:01.249216 139308337518400 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e9d941fd010>}, handler_registry=None I0417 04:29:01.249454 139308337518400 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e9d941fd010>` 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 04:29:01.249502 139308337518400 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e9d940a4a10>` 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 04:29:01.249537 139308337518400 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 0x7e9d941fd010>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e9d941fd010>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e9d940a4a10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e9d940a4a10>}). I0417 04:29:01.249866 139308337518400 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 04:29:01.249942 139308337518400 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e9d68790680> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 04:29:02.243407 139308337518400 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints I0417 04:29:03.102644 139308337518400 checkpoint_manager.py:929] [process=0][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=5), 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/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e9d940a0590> I0417 04:29:03.102821 139308337518400 checkpointing.py:302] Checkpoint manager created! I0417 04:29:04.012637 139308337518400 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 04:29:04.012738 139308337518400 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:29:04.397350 139308337518400 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 04:29:04.397442 139308337518400 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:29:04.413866 139308337518400 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 04:29:04.413931 139308337518400 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:29:04.437899 139308337518400 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 04:29:04.437967 139308337518400 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:29:04.454372 139308337518400 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 04:29:04.454431 139308337518400 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:29:04.477261 139308337518400 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 04:29:04.477374 139308337518400 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:29:04.503297 139308337518400 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 04:29:04.503367 139308337518400 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:29:04.524230 139308337518400 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 04:29:04.524293 139308337518400 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 04:29:04.741981 139308337518400 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 04:29:04.742093 139308337518400 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 04:29:06.158368 139308337518400 maxtext_utils.py:1796] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 04:29:06.158502 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 04:29:06.158558 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 04:29:06.158617 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0417 04:29:06.158692 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 04:29:06.158751 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 04:29:06.158811 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0417 04:29:06.158898 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0417 04:29:06.158945 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0417 04:29:06.158982 139308337518400 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0417 04:29:06.159032 139308337518400 maxtext_utils.py:1796] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 04:29:06.159083 139308337518400 maxtext_utils.py:1796] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 04:29:06.652607 139308337518400 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 04:29:06.652702 139308337518400 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0417 04:29:06.668326 139308337518400 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 04:29:06.668388 139308337518400 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0417 04:29:17.674227 139308337518400 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 04:29:17.682712 139308337518400 metric_logger.py:289] number parameters: 1.104 billion I0417 04:29:29.075643 139308337518400 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 04:29:29.228027 139308337518400 checkpointing.py:798] Waited 0.1523580551147461 seconds for step 0 to finish before starting checkpointing. I0417 04:29:29.230633 139308337518400 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 04:29:29.232273 139308337518400 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 04:29:29.233729 139308337518400 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0. I0417 04:29:29.961643 139308337518400 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 04:29:29.976085 139308337518400 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:29:29.976153 139308337518400 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 04:29:30.221582 139179919779584 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0 I0417 04:29:30.248647 139308337518400 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273376s I0417 04:29:30.249560 139308337518400 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.508 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28003931045532227 s) (per-host) I0417 04:29:30.249636 139308337518400 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.280134s (batch_requests_ready=0.002304s, total_serialization_initiated=0.276997s, others=0.000834s) I0417 04:29:30.249716 139308337518400 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283952s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.283934) I0417 04:29:30.250542 139308337518400 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0. I0417 04:29:30.250860 139179928172288 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:49:30.250816 I0417 04:29:30.260944 139308337518400 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:29:30.261171 139178306152192 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:29:30.261282 139308337518400 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776400169.230612, 'wait_for_prev_duration_secs': 7.152557373046875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776400169.2323198, 'checkpointer_blocking_duration_secs': 1.0186846256256104, 'get_old_steps_start_time': 1776400170.251029, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1776400169.2286308, 'checkpoint_manager_blocking_duration_secs': 1.0326132774353027} I0417 04:29:30.261382 139308337518400 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 04:29:30.261431 139308337518400 max_utils.py:750] Memstats: After params initialized: I0417 04:29:30.261478 139308337518400 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 04:29:30.261513 139308337518400 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 04:29:30.261541 139308337518400 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 04:29:30.261567 139308337518400 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 04:29:30.591823 139308337518400 metric_logger.py:185] completed step: 0, seconds: 11.102, TFLOP/s/device: 1.224, Tokens/s/device: 184.477, total_weights: 65536, loss: 10.874 I0417 04:29:30.595359 139308337518400 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/tensorboard/' I0417 04:29:30.969617 139308337518400 metric_logger.py:185] completed step: 1, seconds: 1.515, TFLOP/s/device: 8.970, Tokens/s/device: 1352.095, total_weights: 65536, loss: 10.874 I0417 04:29:31.115615 139308337518400 metric_logger.py:185] completed step: 2, seconds: 0.378, TFLOP/s/device: 35.904, Tokens/s/device: 5411.790, total_weights: 65536, loss: 10.560 I0417 04:29:31.263237 139308337518400 metric_logger.py:185] completed step: 3, seconds: 0.009, TFLOP/s/device: 1562.637, Tokens/s/device: 235537.665, total_weights: 65536, loss: 9.980 I0417 04:29:31.272532 139308337518400 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0417 04:29:31.557424 139308337518400 checkpointing.py:798] Waited 0.284745454788208 seconds for step 5 to finish before starting checkpointing. I0417 04:29:31.560392 139308337518400 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:29:32.059642 139179902994176 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/items I0417 04:29:32.225918 139176708138752 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776400171811953479, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776400171811953479, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0417 04:29:33.605185 2809 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 04:29:36.112475 139178322937600 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/items/array_metadatas/process_0 I0417 04:30:15.103439 139178314544896 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.505083s (commit=40.292884s, array_metadata_write=1.212199s) I0417 04:30:15.105014 139179928172288 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.997 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.13545775413513 s) (per-host) I0417 04:30:15.105176 139179928172288 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.854200s. I0417 04:30:15.855134 139179928172288 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776400171811953479, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0417 04:30:16.629974 139176708138752 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776400171811953479, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0417 04:30:17.002600 139179928172288 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000369s. I0417 04:30:18.148417 139179928172288 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.165261s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/items I0417 04:30:18.150506 139179928172288 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0/items I0417 04:30:18.681138 139179928172288 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0 I0417 04:30:20.115530 139179928172288 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0`. I0417 04:30:20.116690 139179928172288 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.88 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/0 I0417 04:30:20.118040 139179928172288 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.867061s. I0417 04:30:20.118337 139178306152192 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:30:20.118498 139178306152192 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:30:20.118607 139178306152192 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:30:20.120225 139178306152192 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 04:30:20.120489 139308337518400 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0417 04:30:20.120668 139308337518400 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.560283 seconds. If this number is high, consider checkpointing less frequently. I0417 04:30:20.122572 139308337518400 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5 I0417 04:30:20.125006 139308337518400 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5. I0417 04:30:20.831521 139308337518400 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:30:20.831660 139308337518400 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 04:30:20.869053 139308337518400 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038553s I0417 04:30:20.870186 139308337518400 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 32.027 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04816484451293945 s) (per-host) I0417 04:30:20.870289 139308337518400 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.048294s (batch_requests_ready=0.004817s, total_serialization_initiated=0.042454s, others=0.001023s) I0417 04:30:20.870417 139308337518400 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.053137s (all_items=0.000025s, per_item={'items': '0.00002503'}, temp_paths=0.053112) I0417 04:30:20.871492 139308337518400 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5. I0417 04:30:20.871847 139178306152192 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:50:20.871808 I0417 04:30:20.874233 139308337518400 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:30:20.874557 139178314544896 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:30:20.874778 139308337518400 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776400171.560343, 'wait_for_prev_duration_secs': 48.56028342247009, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776400220.122637, 'checkpointer_blocking_duration_secs': 0.7493433952331543, 'get_old_steps_start_time': 1776400220.8720078, 'get_old_steps_duration_secs': 4.2438507080078125e-05, 'checkpoint_manager_blocking_start_time': 1776400171.5581152, 'checkpoint_manager_blocking_duration_secs': 49.316611528396606} I0417 04:30:20.874946 139308337518400 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0417 04:30:21.070103 139179928172288 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5 I0417 04:30:23.043971 139178322937600 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5/items I0417 04:30:27.211490 139178845116160 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5/items/array_metadatas/process_0 I0417 04:30:44.349063 139308337518400 metric_logger.py:185] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.210, Tokens/s/device: 13898.880, total_weights: 65536, loss: 9.460 I0417 04:30:44.357827 139308337518400 metric_logger.py:185] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.047, Tokens/s/device: 13874.305, total_weights: 65536, loss: 8.959 I0417 04:30:44.503291 139308337518400 metric_logger.py:185] completed step: 6, seconds: 73.084, TFLOP/s/device: 0.186, Tokens/s/device: 28.023, total_weights: 65536, loss: 8.469 I0417 04:30:44.650819 139308337518400 metric_logger.py:185] completed step: 7, seconds: 0.007, TFLOP/s/device: 2050.578, Tokens/s/device: 309085.421, total_weights: 65536, loss: 8.003 I0417 04:30:44.798245 139308337518400 metric_logger.py:185] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.571, Tokens/s/device: 13953.330, total_weights: 65536, loss: 7.572 I0417 04:30:44.945225 139308337518400 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 04:30:44.945971 139308337518400 checkpointing.py:798] Waited 0.0007648468017578125 seconds for step 9 to finish before starting checkpointing. I0417 04:30:44.948003 139308337518400 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:31:07.017205 139178836723456 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.985919s (commit=40.363315s, array_metadata_write=1.622604s) I0417 04:31:07.018331 139178306152192 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.193 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.19629454612732 s) (per-host) I0417 04:31:07.018451 139178306152192 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.146502s. I0417 04:31:09.348478 139178306152192 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000222s. I0417 04:31:10.456285 139178306152192 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.160863s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5/items I0417 04:31:10.457946 139178306152192 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5/items I0417 04:31:10.960511 139178306152192 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5 I0417 04:31:12.383453 139178306152192 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5`. I0417 04:31:12.384299 139178306152192 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.26 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/5 I0417 04:31:12.385665 139178306152192 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.513713s. I0417 04:31:12.385869 139178314544896 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:31:12.385988 139178314544896 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:31:12.386055 139178314544896 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:31:12.387867 139178314544896 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0417 04:31:12.388009 139308337518400 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0417 04:31:12.388090 139308337518400 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.440097 seconds. If this number is high, consider checkpointing less frequently. I0417 04:31:12.389629 139308337518400 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 04:31:12.391647 139308337518400 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9. I0417 04:31:13.119485 139308337518400 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:31:13.119580 139308337518400 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 04:31:13.145192 139308337518400 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026650s I0417 04:31:13.145978 139308337518400 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.200 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031352996826171875 s) (per-host) I0417 04:31:13.146049 139308337518400 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031442s (batch_requests_ready=0.002085s, total_serialization_initiated=0.028620s, others=0.000737s) I0417 04:31:13.146127 139308337518400 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035514s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.035497) I0417 04:31:13.146796 139308337518400 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9. I0417 04:31:13.147145 139179371755264 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:51:13.147103 I0417 04:31:13.161984 139308337518400 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:31:13.162271 139178314544896 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:31:13.162431 139308337518400 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776400244.9479692, 'wait_for_prev_duration_secs': 27.440097093582153, 'time_between_consecutive_saves_sec': 24.827675580978394, 'checkpointer_blocking_start_time': 1776400272.3896692, 'checkpointer_blocking_duration_secs': 0.7576143741607666, 'get_old_steps_start_time': 1776400273.1473036, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776400244.9462078, 'checkpoint_manager_blocking_duration_secs': 28.216189861297607} I0417 04:31:13.162548 139308337518400 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 04:31:13.162589 139308337518400 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:31:13.350254 139178306152192 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9 I0417 04:31:15.931756 139178876606208 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9/items I0417 04:31:19.245594 139179928172288 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9/items/array_metadatas/process_0 I0417 04:31:59.171964 139179412276992 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.120868s (commit=40.441928s, array_metadata_write=1.678940s) I0417 04:31:59.173187 139179371755264 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.295 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.058526277542114 s) (per-host) I0417 04:31:59.173306 139179371755264 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.026047s. I0417 04:32:01.301072 139179371755264 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000229s. I0417 04:32:02.846447 139179371755264 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.154515s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9/items I0417 04:32:02.848116 139179371755264 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9/items I0417 04:32:03.380992 139179371755264 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9 I0417 04:32:05.254763 139179371755264 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9`. I0417 04:32:05.255637 139179371755264 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.87 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9 I0417 04:32:05.256925 139179371755264 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.109663s. I0417 04:32:05.257076 139178314544896 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:32:05.257158 139178314544896 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:32:05.257207 139178314544896 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:32:05.259479 139178314544896 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 04:32:05.259640 139308337518400 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0417 04:32:05.259784 139308337518400 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 04:32:05.260439 139308337518400 metric_logger.py:185] completed step: 9, seconds: 0.149, TFLOP/s/device: 91.465, Tokens/s/device: 13786.604, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 17 04:32:18 UTC 2026 EXIT_CODE=0