XPK Start: Sun Apr 19 00:47:06 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 00:47:30.995787: 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) I0419 00:47:31.176610 136031473182528 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 00:47:40.218658 136031473182528 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 00:47:40.221091 136031473182528 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-2raos-slice-job-0-0.mt-15-ocdbt-true-sv-2raos:8482 I0419 00:47:41.635115 136031473182528 max_utils.py:284] Jax distributed system initialized! I0419 00:47:47.719530 136031473182528 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 00:47:47.719631 136031473182528 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 00:47:47.719672 136031473182528 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 I0419 00:47:47.719707 136031473182528 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 00:47:48.757147 136031473182528 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 00:47:48.757837 136031473182528 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 00:47:48.758034 136031473182528 checkpointing.py:688] Setting up checkpoint logger... I0419 00:47:48.758090 136031473182528 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 00:47:48.758154 136031473182528 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 00:47:48.758508 136031473182528 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 0x7bb7943343b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 00:47:52.178356 136031473182528 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 00:47:52.178703 136031473182528 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 0x7bb6882651c0>}, handler_registry=None I0419 00:47:52.178953 136031473182528 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bb6882651c0>` 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`. I0419 00:47:52.179003 136031473182528 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bb688264aa0>` 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`. I0419 00:47:52.179039 136031473182528 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 0x7bb6882651c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bb6882651c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bb688264aa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bb688264aa0>}). I0419 00:47:52.179371 136031473182528 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 00:47:52.179440 136031473182528 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ba314276c00> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 00:47:54.406419 136031473182528 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints I0419 00:47:55.285392 136031473182528 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bb688152780> I0419 00:47:55.285558 136031473182528 checkpointing.py:302] Checkpoint manager created! I0419 00:47:56.205844 136031473182528 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 00:47:56.205948 136031473182528 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:47:56.593055 136031473182528 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 00:47:56.593142 136031473182528 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:47:56.609382 136031473182528 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 00:47:56.609438 136031473182528 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:47:56.633273 136031473182528 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:47:56.633343 136031473182528 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:47:56.649683 136031473182528 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:47:56.649744 136031473182528 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:47:56.666163 136031473182528 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:47:56.666224 136031473182528 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:47:56.691652 136031473182528 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 00:47:56.691718 136031473182528 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:47:56.712582 136031473182528 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 00:47:56.712653 136031473182528 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 00:47:56.929075 136031473182528 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 00:47:56.929187 136031473182528 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 00:47:58.356416 136031473182528 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 00:47:58.356538 136031473182528 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 00:47:58.356593 136031473182528 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 00:47:58.356685 136031473182528 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0419 00:47:58.356755 136031473182528 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 00:47:58.356796 136031473182528 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 00:47:58.356953 136031473182528 maxtext_utils.py:1827] 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) I0419 00:47:58.357032 136031473182528 maxtext_utils.py:1827] 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') I0419 00:47:58.357082 136031473182528 maxtext_utils.py:1827] 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) I0419 00:47:58.357134 136031473182528 maxtext_utils.py:1827] 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) I0419 00:47:58.357189 136031473182528 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 00:47:58.357242 136031473182528 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 00:47:58.857022 136031473182528 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 00:47:58.857114 136031473182528 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 00:47:58.872404 136031473182528 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 00:47:58.872462 136031473182528 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 00:48:09.380680 136031473182528 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. I0419 00:48:09.388662 136031473182528 metric_logger.py:301] number parameters: 1.104 billion I0419 00:48:20.981421 136031473182528 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 00:48:21.215869 136031473182528 checkpointing.py:798] Waited 0.23442935943603516 seconds for step 0 to finish before starting checkpointing. I0419 00:48:21.218153 136031473182528 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 00:48:21.220030 136031473182528 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0419 00:48:21.221702 136031473182528 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0. I0419 00:48:22.369455 136031473182528 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 00:48:22.383671 136031473182528 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:48:22.383739 136031473182528 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 I0419 00:48:22.657206 136031473182528 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274218s I0419 00:48:22.657986 136031473182528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.499 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2805171012878418 s) (per-host) I0419 00:48:22.658059 136031473182528 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.280611s (batch_requests_ready=0.002105s, total_serialization_initiated=0.277809s, others=0.000697s) I0419 00:48:22.658139 136031473182528 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.284569s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.284552) I0419 00:48:22.659011 136031473182528 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.44 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0. I0419 00:48:22.659316 135902496028416 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:08:22.659292 I0419 00:48:22.668410 136031473182528 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:48:22.668698 135901929793280 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:48:22.668873 136031473182528 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776559701.2181349, 'wait_for_prev_duration_secs': 6.175041198730469e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776559701.2201157, 'checkpointer_blocking_duration_secs': 1.4393198490142822, 'get_old_steps_start_time': 1776559702.6594641, 'get_old_steps_duration_secs': 3.4332275390625e-05, 'checkpoint_manager_blocking_start_time': 1776559701.2163818, 'checkpoint_manager_blocking_duration_secs': 1.4524431228637695} I0419 00:48:22.668995 136031473182528 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 00:48:22.669059 136031473182528 max_utils.py:750] Memstats: After params initialized: I0419 00:48:22.669121 136031473182528 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0419 00:48:22.669175 136031473182528 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0419 00:48:22.669214 136031473182528 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0419 00:48:22.669248 136031473182528 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0419 00:48:23.014249 136031473182528 metric_logger.py:196] completed step: 0, seconds: 11.291, TFLOP/s/device: 1.203, Tokens/s/device: 181.383, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0419 00:48:23.017834 136031473182528 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/tensorboard/' I0419 00:48:23.073937 135902487635712 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0 I0419 00:48:23.400562 136031473182528 metric_logger.py:196] completed step: 1, seconds: 2.031, TFLOP/s/device: 6.690, Tokens/s/device: 1008.389, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0419 00:48:23.546666 136031473182528 metric_logger.py:196] completed step: 2, seconds: 0.387, TFLOP/s/device: 35.101, Tokens/s/device: 5290.787, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0419 00:48:23.694038 136031473182528 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1459.569, Tokens/s/device: 220002.148, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0419 00:48:23.703414 136031473182528 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0419 00:48:23.988101 136031473182528 checkpointing.py:798] Waited 0.2846412658691406 seconds for step 5 to finish before starting checkpointing. I0419 00:48:23.991405 136031473182528 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:48:25.654367 135902470850304 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/items I0419 00:48:25.906097 135901921400576 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776559705404793119, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776559705404793119, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0419 00:48:28.074920 2816 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 00:48:29.902962 135902454064896 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_0 I0419 00:49:05.538542 135902445672192 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.469009s (commit=35.809840s, array_metadata_write=1.659168s) I0419 00:49:05.539873 135902496028416 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 36.597 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.162352085113525 s) (per-host) I0419 00:49:05.540046 135902496028416 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.880650s. I0419 00:49:06.753417 135902496028416 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776559705404793119, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0419 00:49:08.216587 135901921400576 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776559705404793119, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0419 00:49:08.606544 135902496028416 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000364s. I0419 00:49:08.834424 135902496028416 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:49:09.735414 135902496028416 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.571368s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/items I0419 00:49:09.737570 135902496028416 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/items I0419 00:49:11.121990 135902496028416 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0 I0419 00:49:13.535065 135902496028416 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0`. I0419 00:49:13.536040 135902496028416 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.32 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0 I0419 00:49:13.537533 135902496028416 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.878134s. I0419 00:49:13.537831 135901929793280 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:49:13.537995 135901929793280 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:49:13.538114 135901929793280 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:49:13.540970 135901929793280 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 00:49:13.541271 136031473182528 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. W0419 00:49:13.541449 136031473182528 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.550046 seconds. If this number is high, consider checkpointing less frequently. I0419 00:49:13.543623 136031473182528 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5 I0419 00:49:13.546050 136031473182528 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5. I0419 00:49:14.267174 136031473182528 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:49:14.267309 136031473182528 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 I0419 00:49:14.307044 136031473182528 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040921s I0419 00:49:14.308098 136031473182528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 30.947 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04984593391418457 s) (per-host) I0419 00:49:14.308212 136031473182528 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.049993s (batch_requests_ready=0.004209s, total_serialization_initiated=0.044831s, others=0.000953s) I0419 00:49:14.308346 136031473182528 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.054964s (all_items=0.000024s, per_item={'items': '0.00002360'}, temp_paths=0.054941) I0419 00:49:14.309340 136031473182528 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5. I0419 00:49:14.309712 135901929793280 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:09:14.309676 I0419 00:49:14.312194 136031473182528 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:49:14.312592 135901393278720 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:49:14.312835 136031473182528 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776559703.9913597, 'wait_for_prev_duration_secs': 49.55004620552063, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776559753.5436943, 'checkpointer_blocking_duration_secs': 0.7661566734313965, 'get_old_steps_start_time': 1776559754.3098772, 'get_old_steps_duration_secs': 4.410743713378906e-05, 'checkpoint_manager_blocking_start_time': 1776559703.988936, 'checkpoint_manager_blocking_duration_secs': 50.323832273483276} I0419 00:49:14.312985 136031473182528 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0419 00:49:14.497438 135902496028416 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5 I0419 00:49:16.637671 135901410064128 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5/items I0419 00:49:20.354229 135901953545984 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_0 I0419 00:49:38.322889 136031473182528 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.268, Tokens/s/device: 13907.658, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0419 00:49:38.332367 136031473182528 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.274, Tokens/s/device: 13908.603, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0419 00:49:38.477865 136031473182528 metric_logger.py:196] completed step: 6, seconds: 74.627, TFLOP/s/device: 0.182, Tokens/s/device: 27.443, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0419 00:49:38.625424 136031473182528 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1944.630, Tokens/s/device: 293115.786, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0419 00:49:38.772799 136031473182528 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.651, Tokens/s/device: 13965.319, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0419 00:49:38.919541 136031473182528 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 00:49:38.920474 136031473182528 checkpointing.py:798] Waited 0.0009169578552246094 seconds for step 9 to finish before starting checkpointing. I0419 00:49:38.922699 136031473182528 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:50:01.324222 135901945153280 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.572453s (commit=41.125455s, array_metadata_write=2.446998s) I0419 00:50:01.325273 135901929793280 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.561 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.06704330444336 s) (per-host) I0419 00:50:01.325341 135901929793280 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.014828s. I0419 00:50:04.738453 135901929793280 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:50:05.317483 135901929793280 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000231s. I0419 00:50:05.796242 135901929793280 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.326952s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5/items I0419 00:50:05.797977 135901929793280 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5/items I0419 00:50:06.823492 135901929793280 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5 I0419 00:50:09.068314 135901929793280 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5`. I0419 00:50:09.069045 135901929793280 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 55.53 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/5 I0419 00:50:09.070428 135901929793280 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 54.759910s. I0419 00:50:09.070655 135901393278720 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:50:09.070777 135901393278720 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:50:09.070856 135901393278720 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:50:09.072839 135901393278720 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0419 00:50:09.073003 136031473182528 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. W0419 00:50:09.073131 136031473182528 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.150437 seconds. If this number is high, consider checkpointing less frequently. I0419 00:50:09.074725 136031473182528 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0419 00:50:09.076691 136031473182528 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9. I0419 00:50:09.798627 136031473182528 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:50:09.798722 136031473182528 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 I0419 00:50:09.825570 136031473182528 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027896s I0419 00:50:09.826828 136031473182528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 47.043 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0327908992767334 s) (per-host) I0419 00:50:09.826911 136031473182528 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032917s (batch_requests_ready=0.001893s, total_serialization_initiated=0.029855s, others=0.001170s) I0419 00:50:09.826998 136031473182528 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036899s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.036883) I0419 00:50:09.827701 136031473182528 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9. I0419 00:50:09.828001 135901401671424 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:10:09.827961 I0419 00:50:09.843862 136031473182528 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:50:09.844153 135901393278720 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:50:09.844314 136031473182528 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776559778.922666, 'wait_for_prev_duration_secs': 30.150437116622925, 'time_between_consecutive_saves_sec': 25.38162589073181, 'checkpointer_blocking_start_time': 1776559809.0747652, 'checkpointer_blocking_duration_secs': 0.7533526420593262, 'get_old_steps_start_time': 1776559809.8281348, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776559778.9208848, 'checkpoint_manager_blocking_duration_secs': 30.923397302627563} I0419 00:50:09.844415 136031473182528 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 00:50:09.844459 136031473182528 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:50:10.042672 135901929793280 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9 I0419 00:50:11.756026 135901410064128 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9/items I0419 00:50:16.754284 135901953545984 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_0 I0419 00:50:57.122146 135901945153280 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.824797s (commit=41.314485s, array_metadata_write=2.510312s) I0419 00:50:57.123256 135901401671424 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.375 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.32921743392944 s) (per-host) I0419 00:50:57.123389 135901401671424 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.294879s. I0419 00:50:58.783310 135901401671424 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:50:59.001516 135901401671424 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000245s. I0419 00:50:59.848412 135901401671424 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.888613s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9/items I0419 00:50:59.850076 135901401671424 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9/items I0419 00:51:00.371377 135901401671424 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9 I0419 00:51:02.621181 135901401671424 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_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9`. I0419 00:51:02.622020 135901401671424 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.55 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9 I0419 00:51:02.623581 135901401671424 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.795069s. I0419 00:51:02.623737 135901393278720 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:51:02.623827 135901393278720 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:51:02.623878 135901393278720 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:51:02.625841 135901393278720 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 00:51:02.626001 136031473182528 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. I0419 00:51:02.626145 136031473182528 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 00:51:02.627019 136031473182528 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.974, Tokens/s/device: 13863.316, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 00:51:12 UTC 2026 EXIT_CODE=0