XPK Start: Mon Apr 20 16:25:44 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-20 16:26:09.013514: 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) I0420 16:26:09.191365 139815446923072 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 16:26:18.232504 139815446923072 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 16:26:18.234757 139815446923072 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-2aww1-slice-job-0-0.mt-10-shardy-false-2aww1:8482 I0420 16:26:19.121065 139815446923072 max_utils.py:284] Jax distributed system initialized! I0420 16:26:24.294909 139815446923072 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 16:26:24.295011 139815446923072 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 16:26:24.295051 139815446923072 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 I0420 16:26:24.295105 139815446923072 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 16:26:24.981045 139815446923072 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:26:24.981640 139815446923072 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:26:24.981812 139815446923072 checkpointing.py:688] Setting up checkpoint logger... I0420 16:26:24.981862 139815446923072 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 16:26:24.981905 139815446923072 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 16:26:24.982298 139815446923072 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 0x7f289a4449b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 16:26:27.914407 139815446923072 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 16:26:27.914641 139815446923072 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 0x7f1508179b50>}, handler_registry=None I0420 16:26:27.914891 139815446923072 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f1508179b50>` 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`. I0420 16:26:27.914940 139815446923072 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f1508179760>` 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`. I0420 16:26:27.914976 139815446923072 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 0x7f1508179b50>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f1508179b50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f1508179760>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f1508179760>}). I0420 16:26:27.915297 139815446923072 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 16:26:27.915369 139815446923072 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f14c448b740> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 16:26:29.199644 139815446923072 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints I0420 16:26:29.632958 139815446923072 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=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/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f15081783e0> I0420 16:26:29.633153 139815446923072 checkpointing.py:302] Checkpoint manager created! I0420 16:26:30.555217 139815446923072 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 16:26:30.555320 139815446923072 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:26:30.932065 139815446923072 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:26:30.932172 139815446923072 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:26:30.948483 139815446923072 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:26:30.948542 139815446923072 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:26:30.972037 139815446923072 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 16:26:30.972117 139815446923072 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:26:30.988377 139815446923072 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 16:26:30.988437 139815446923072 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:26:31.004692 139815446923072 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 16:26:31.004751 139815446923072 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:26:31.031349 139815446923072 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 16:26:31.031434 139815446923072 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:26:31.052794 139815446923072 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 16:26:31.052862 139815446923072 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 16:26:31.268589 139815446923072 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 16:26:31.268692 139815446923072 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 16:26:32.662586 139815446923072 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 16:26:32.662708 139815446923072 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) I0420 16:26:32.662778 139815446923072 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) I0420 16:26:32.662841 139815446923072 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') I0420 16:26:32.662897 139815446923072 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) I0420 16:26:32.662938 139815446923072 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) I0420 16:26:32.662992 139815446923072 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) I0420 16:26:32.663045 139815446923072 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') I0420 16:26:32.663101 139815446923072 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) I0420 16:26:32.663163 139815446923072 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) I0420 16:26:32.663243 139815446923072 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 16:26:32.663317 139815446923072 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 16:26:33.155002 139815446923072 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 16:26:33.155105 139815446923072 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 16:26:33.170363 139815446923072 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 16:26:33.170427 139815446923072 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 16:26:43.972066 139815446923072 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. I0420 16:26:43.980006 139815446923072 metric_logger.py:301] number parameters: 1.104 billion I0420 16:26:55.399441 139815446923072 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 16:26:55.553499 139815446923072 checkpointing.py:798] Waited 0.1540369987487793 seconds for step 0 to finish before starting checkpointing. I0420 16:26:55.555902 139815446923072 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:26:55.557612 139815446923072 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0420 16:26:55.559209 139815446923072 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0. I0420 16:26:56.303847 139815446923072 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 16:26:56.318853 139815446923072 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 16:26:56.318928 139815446923072 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 I0420 16:26:56.591237 139689288279808 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0 I0420 16:26:56.593638 139815446923072 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275712s I0420 16:26:56.594584 139815446923072 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.462 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28244519233703613 s) (per-host) I0420 16:26:56.594649 139815446923072 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282533s (batch_requests_ready=0.002554s, total_serialization_initiated=0.279068s, others=0.000911s) I0420 16:26:56.594727 139815446923072 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286690s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.286673) I0420 16:26:56.595498 139815446923072 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0. I0420 16:26:56.595825 139689296672512 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 16:46:56.595787 I0420 16:26:56.602070 139815446923072 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:26:56.602368 139684971149056 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:26:56.602521 139815446923072 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776702415.555884, 'wait_for_prev_duration_secs': 5.984306335449219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776702415.5576525, 'checkpointer_blocking_duration_secs': 1.038323163986206, 'get_old_steps_start_time': 1776702416.595999, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776702415.5540376, 'checkpoint_manager_blocking_duration_secs': 1.0484468936920166} I0420 16:26:56.602622 139815446923072 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 16:26:56.602672 139815446923072 max_utils.py:750] Memstats: After params initialized: I0420 16:26:56.602720 139815446923072 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 16:26:56.602756 139815446923072 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 16:26:56.602781 139815446923072 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 16:26:56.602803 139815446923072 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 16:26:56.938766 139815446923072 metric_logger.py:196] completed step: 0, seconds: 11.114, TFLOP/s/device: 1.222, Tokens/s/device: 184.267, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 16:26:56.942288 139815446923072 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/tensorboard/' I0420 16:26:57.366627 139815446923072 metric_logger.py:196] completed step: 1, seconds: 1.537, TFLOP/s/device: 8.837, Tokens/s/device: 1332.070, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 16:26:57.512616 139815446923072 metric_logger.py:196] completed step: 2, seconds: 0.429, TFLOP/s/device: 31.704, Tokens/s/device: 4778.727, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0420 16:26:57.659882 139815446923072 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1436.423, Tokens/s/device: 216513.374, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0420 16:26:58.905036 139689271494400 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/items I0420 16:26:59.089567 139683409426176 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776702418638721354, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776702418638721354, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0420 16:27:00.476618 2779 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 16:27:02.398031 139685020038912 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/items/array_metadatas/process_0 I0420 16:27:19.444850 139815446923072 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.504, Tokens/s/device: 13943.165, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0420 16:27:19.453591 139815446923072 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.261, Tokens/s/device: 13906.525, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0420 16:27:19.599768 139815446923072 metric_logger.py:196] completed step: 6, seconds: 21.784, TFLOP/s/device: 0.624, Tokens/s/device: 94.016, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0420 16:27:19.747161 139815446923072 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2067.427, Tokens/s/device: 311625.076, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0420 16:27:19.894227 139815446923072 metric_logger.py:196] completed step: 8, seconds: 0.148, TFLOP/s/device: 92.028, Tokens/s/device: 13871.392, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0420 16:27:20.040717 139815446923072 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 16:27:20.041409 139815446923072 checkpointing.py:798] Waited 0.0007152557373046875 seconds for step 9 to finish before starting checkpointing. I0420 16:27:20.043690 139815446923072 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:27:37.393426 139685011646208 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.914940s (commit=35.665164s, array_metadata_write=1.249776s) I0420 16:27:37.394605 139689296672512 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 38.449 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.08243250846863 s) (per-host) I0420 16:27:37.394730 139689296672512 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.798788s. I0420 16:27:38.595330 139689296672512 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776702418638721354, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0420 16:27:39.355963 139683409426176 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776702418638721354, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0420 16:27:39.733952 139689296672512 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000251s. I0420 16:27:39.996458 139689296672512 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 16:27:40.939033 139689296672512 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.216331s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/items I0420 16:27:40.940817 139689296672512 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/items I0420 16:27:41.501225 139689296672512 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0 I0420 16:27:42.906315 139689296672512 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0`. I0420 16:27:42.907028 139689296672512 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 47.35 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0 I0420 16:27:42.908433 139689296672512 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.312489s. I0420 16:27:42.908628 139684971149056 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:27:42.908753 139684971149056 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:27:42.908818 139684971149056 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:27:42.911810 139684971149056 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 16:27:42.911981 139815446923072 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. W0420 16:27:42.912135 139815446923072 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.868441 seconds. If this number is high, consider checkpointing less frequently. I0420 16:27:42.914054 139815446923072 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0420 16:27:42.916279 139815446923072 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9. I0420 16:27:43.623913 139815446923072 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 16:27:43.624010 139815446923072 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 I0420 16:27:43.650482 139815446923072 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027315s I0420 16:27:43.651520 139815446923072 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.044 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03210759162902832 s) (per-host) I0420 16:27:43.651597 139815446923072 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032203s (batch_requests_ready=0.001936s, total_serialization_initiated=0.029307s, others=0.000959s) I0420 16:27:43.651688 139815446923072 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036263s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.036248) I0420 16:27:43.652448 139815446923072 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9. I0420 16:27:43.652730 139684971149056 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 16:47:43.652707 I0420 16:27:43.664914 139815446923072 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:27:43.665181 139688205371136 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:27:43.665313 139815446923072 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776702440.0436594, 'wait_for_prev_duration_secs': 22.868441104888916, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776702462.9141068, 'checkpointer_blocking_duration_secs': 0.7387261390686035, 'get_old_steps_start_time': 1776702463.6528573, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776702440.041636, 'checkpoint_manager_blocking_duration_secs': 23.62364101409912} I0420 16:27:43.665430 139815446923072 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 16:27:43.665475 139815446923072 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:27:43.867445 139689296672512 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9 I0420 16:27:45.720479 139682699601664 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9/items I0420 16:27:49.141620 139688736171776 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9/items/array_metadatas/process_0 I0420 16:28:29.057294 139688213763840 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.700941s (commit=40.036605s, array_metadata_write=1.664336s) I0420 16:28:29.058424 139684971149056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.763 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.43898129463196 s) (per-host) I0420 16:28:29.058553 139684971149056 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.405742s. I0420 16:28:30.985745 139684971149056 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000242s. I0420 16:28:31.219044 139684971149056 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 16:28:32.200886 139684971149056 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.245630s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9/items I0420 16:28:32.202510 139684971149056 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9/items I0420 16:28:32.766472 139684971149056 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9 I0420 16:28:34.194585 139684971149056 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9`. I0420 16:28:34.195420 139684971149056 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.28 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9 I0420 16:28:34.196831 139684971149056 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.544018s. I0420 16:28:34.196984 139688205371136 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:28:34.197065 139688205371136 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:28:34.197154 139688205371136 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:28:34.199242 139688205371136 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 16:28:34.199401 139815446923072 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. I0420 16:28:34.199548 139815446923072 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:28:34.200541 139815446923072 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.402, Tokens/s/device: 13927.804, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 16:28:45 UTC 2026 EXIT_CODE=0