XPK Start: Wed Apr 22 13:40:34 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-22 13:40:58.104884: 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) I0422 13:40:58.654603 135020979578688 max_utils.py:273] Attempting to initialize the jax distributed system... I0422 13:41:07.696753 135020979578688 distributed.py:149] Starting JAX distributed service on [::]:8482 I0422 13:41:07.699142 135020979578688 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-y6cae-slice-job-0-0.mt-15-ocdbt-true-sv-y6cae:8482 I0422 13:41:09.269900 135020979578688 max_utils.py:284] Jax distributed system initialized! I0422 13:41:15.624601 135020979578688 max_utils.py:800] System Information: Jax Version: 0.9.2 I0422 13:41:15.624705 135020979578688 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0422 13:41:15.624744 135020979578688 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 I0422 13:41:15.624777 135020979578688 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 13:41:16.324455 135020979578688 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 13:41:16.325038 135020979578688 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 13:41:16.325227 135020979578688 checkpointing.py:688] Setting up checkpoint logger... I0422 13:41:16.325283 135020979578688 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 13:41:16.325325 135020979578688 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0422 13:41:16.325656 135020979578688 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 0x7acc4df34260>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 13:41:19.251648 135020979578688 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 13:41:19.251881 135020979578688 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab830054b00>}, handler_registry=None I0422 13:41:19.252132 135020979578688 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab830054b00>` 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`. I0422 13:41:19.252182 135020979578688 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ab75c1f7200>` 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`. I0422 13:41:19.252219 135020979578688 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 0x7ab830054b00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab830054b00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ab75c1f7200>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ab75c1f7200>}). I0422 13:41:19.252542 135020979578688 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0422 13:41:19.252614 135020979578688 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ab75c290180> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0422 13:41:20.465539 135020979578688 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints I0422 13:41:20.885292 135020979578688 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ab75c636cf0> I0422 13:41:20.885461 135020979578688 checkpointing.py:302] Checkpoint manager created! I0422 13:41:21.825781 135020979578688 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 13:41:21.825884 135020979578688 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 13:41:22.202445 135020979578688 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 13:41:22.202538 135020979578688 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 13:41:22.219037 135020979578688 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 13:41:22.219106 135020979578688 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 13:41:22.243749 135020979578688 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 13:41:22.243814 135020979578688 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:41:22.260273 135020979578688 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 13:41:22.260333 135020979578688 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:41:22.276744 135020979578688 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 13:41:22.276804 135020979578688 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:41:22.301674 135020979578688 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 13:41:22.301744 135020979578688 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:41:22.322734 135020979578688 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 13:41:22.322798 135020979578688 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 13:41:22.538842 135020979578688 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 13:41:22.538948 135020979578688 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 13:41:23.959287 135020979578688 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0422 13:41:23.959409 135020979578688 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) I0422 13:41:23.959461 135020979578688 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) I0422 13:41:23.959542 135020979578688 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') I0422 13:41:23.959598 135020979578688 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) I0422 13:41:23.959640 135020979578688 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) I0422 13:41:23.959691 135020979578688 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) I0422 13:41:23.959741 135020979578688 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') I0422 13:41:23.959779 135020979578688 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) I0422 13:41:23.959814 135020979578688 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) I0422 13:41:23.959859 135020979578688 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0422 13:41:23.959903 135020979578688 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0422 13:41:24.457603 135020979578688 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 13:41:24.457703 135020979578688 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0422 13:41:24.473340 135020979578688 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 13:41:24.473400 135020979578688 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0422 13:41:35.557663 135020979578688 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. I0422 13:41:35.558480 135020979578688 metric_logger.py:301] number parameters: 1.104 billion I0422 13:41:46.891274 135020979578688 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 13:41:47.045623 135020979578688 checkpointing.py:798] Waited 0.1543281078338623 seconds for step 0 to finish before starting checkpointing. I0422 13:41:47.048040 135020979578688 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 13:41:47.049775 135020979578688 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0422 13:41:47.051196 135020979578688 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/0. I0422 13:41:47.388658 135020979578688 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 13:41:47.389679 135020979578688 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 13:41:47.389735 135020979578688 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 I0422 13:41:47.664887 135020979578688 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276300s I0422 13:41:47.665061 135020979578688 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.476 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2817230224609375 s) (per-host) I0422 13:41:47.665122 135020979578688 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281793s (batch_requests_ready=0.002204s, total_serialization_initiated=0.279510s, others=0.000079s) I0422 13:41:47.665221 135020979578688 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286302s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.286283) I0422 13:41:47.666038 135020979578688 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/0. I0422 13:41:47.666427 134891606202112 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:01:47.666381 I0422 13:41:48.097937 135020979578688 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 13:41:48.098324 134890553075456 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 13:41:48.098468 135020979578688 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776865307.048021, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776865307.0498137, 'checkpointer_blocking_duration_secs': 0.6167311668395996, 'get_old_steps_start_time': 1776865307.6665702, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776865307.0461547, 'checkpoint_manager_blocking_duration_secs': 1.0522711277008057} I0422 13:41:48.098646 135020979578688 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 13:41:48.098700 135020979578688 max_utils.py:750] Memstats: After params initialized: I0422 13:41:48.098757 135020979578688 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 13:41:48.098788 135020979578688 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 13:41:48.098814 135020979578688 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 13:41:48.098837 135020979578688 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 13:41:48.415228 135020979578688 metric_logger.py:196] completed step: 0, seconds: 11.333, TFLOP/s/device: 1.199, Tokens/s/device: 180.716, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0422 13:41:48.596508 135020979578688 metric_logger.py:196] completed step: 1, seconds: 1.522, TFLOP/s/device: 8.924, Tokens/s/device: 1345.182, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0422 13:41:49.021311 135020979578688 metric_logger.py:196] completed step: 2, seconds: 0.034, TFLOP/s/device: 400.163, Tokens/s/device: 60316.899, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0422 13:41:49.168477 135020979578688 metric_logger.py:196] completed step: 3, seconds: 0.425, TFLOP/s/device: 31.968, Tokens/s/device: 4818.563, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0422 13:41:49.172940 135020979578688 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0422 13:41:49.462844 135020979578688 checkpointing.py:798] Waited 0.2898683547973633 seconds for step 5 to finish before starting checkpointing. I0422 13:41:49.465956 135020979578688 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 13:41:51.292827 2786 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 13:41:53.835035 134890561468160 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0422 13:42:25.741139 134891606202112 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 41.181 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.35774493217468 s) (per-host) I0422 13:42:25.741262 134891606202112 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.074191s. I0422 13:42:33.613358 134891606202112 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 45.946270s. I0422 13:42:33.613645 134890553075456 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 13:42:33.613769 134890553075456 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 13:42:33.613820 134890553075456 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 13:42:33.615522 134890553075456 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 13:42:33.615696 135020979578688 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0422 13:42:33.615825 135020979578688 checkpoint_manager.py:1452] Waiting for previous save to complete took 44.149869 seconds. If this number is high, consider checkpointing less frequently. I0422 13:42:33.618772 135020979578688 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0422 13:42:33.621269 135020979578688 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/5. I0422 13:42:33.921727 135020979578688 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 13:42:33.921819 135020979578688 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 I0422 13:42:33.956914 135020979578688 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036384s I0422 13:42:33.957080 135020979578688 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 38.804 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.039752960205078125 s) (per-host) I0422 13:42:33.957144 135020979578688 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039827s (batch_requests_ready=0.001720s, total_serialization_initiated=0.038025s, others=0.000082s) I0422 13:42:33.957236 135020979578688 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045156s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.045140) I0422 13:42:33.957937 135020979578688 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/5. I0422 13:42:33.958221 134890553075456 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:02:33.958190 I0422 13:42:33.965418 135020979578688 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 13:42:33.965713 134867526309632 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 13:42:33.965878 135020979578688 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776865309.4659276, 'wait_for_prev_duration_secs': 44.149869203567505, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776865353.6188116, 'checkpointer_blocking_duration_secs': 0.3395576477050781, 'get_old_steps_start_time': 1776865353.9583924, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776865309.4634886, 'checkpoint_manager_blocking_duration_secs': 44.50235462188721} I0422 13:42:33.965984 135020979578688 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0422 13:42:33.966737 135020979578688 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.920, Tokens/s/device: 13403.054, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0422 13:42:33.972239 135020979578688 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.230, Tokens/s/device: 13901.899, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0422 13:42:39.636472 134890561468160 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0422 13:42:56.724969 135020979578688 metric_logger.py:196] completed step: 6, seconds: 44.798, TFLOP/s/device: 0.303, Tokens/s/device: 45.716, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0422 13:42:56.872177 135020979578688 metric_logger.py:196] completed step: 7, seconds: 22.606, TFLOP/s/device: 0.601, Tokens/s/device: 90.594, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0422 13:42:57.019423 135020979578688 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.326, Tokens/s/device: 13464.118, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0422 13:42:57.166252 135020979578688 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 13:42:57.166850 135020979578688 checkpointing.py:798] Waited 0.0006208419799804688 seconds for step 9 to finish before starting checkpointing. I0422 13:42:57.168963 135020979578688 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 13:43:16.078114 134890553075456 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.466 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.160733461380005 s) (per-host) I0422 13:43:16.078229 134890553075456 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.119895s. I0422 13:43:25.467960 134890553075456 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.509611s. I0422 13:43:25.468255 134867526309632 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 13:43:25.468389 134867526309632 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 13:43:25.468439 134867526309632 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 13:43:25.469952 134867526309632 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 13:43:25.470138 135020979578688 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0422 13:43:25.470259 135020979578688 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.301303 seconds. If this number is high, consider checkpointing less frequently. I0422 13:43:25.471731 135020979578688 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0422 13:43:25.473804 135020979578688 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/9. I0422 13:43:25.874559 135020979578688 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 13:43:25.874652 135020979578688 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 I0422 13:43:25.909203 135020979578688 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035570s I0422 13:43:25.909376 135020979578688 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 39.566 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03898739814758301 s) (per-host) I0422 13:43:25.909425 135020979578688 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039046s (batch_requests_ready=0.001615s, total_serialization_initiated=0.037367s, others=0.000064s) I0422 13:43:25.909516 135020979578688 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043202s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.043190) I0422 13:43:25.910188 135020979578688 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.44 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/9. I0422 13:43:25.910526 134890561468160 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:03:25.910488 I0422 13:43:25.917164 135020979578688 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 13:43:25.917445 134867526309632 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 13:43:25.917599 135020979578688 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776865377.1689339, 'wait_for_prev_duration_secs': 28.301303386688232, 'time_between_consecutive_saves_sec': 23.553365468978882, 'checkpointer_blocking_start_time': 1776865405.4717689, 'checkpointer_blocking_duration_secs': 0.4389007091522217, 'get_old_steps_start_time': 1776865405.9106867, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776865377.1670732, 'checkpoint_manager_blocking_duration_secs': 28.750494241714478} I0422 13:43:25.917709 135020979578688 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 13:43:25.917751 135020979578688 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 13:43:31.077285 134889445762816 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0422 13:44:07.755184 134890561468160 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.713 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.88475227355957 s) (per-host) I0422 13:44:07.755303 134890561468160 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.844663s. I0422 13:44:16.371804 134890561468160 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.461148s. I0422 13:44:16.372071 134867526309632 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 13:44:16.372217 134867526309632 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 13:44:16.372257 134867526309632 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 13:44:16.374725 134867526309632 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 13:44:16.374899 135020979578688 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0422 13:44:16.375046 135020979578688 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 13:44:16.375937 135020979578688 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.264, Tokens/s/device: 13907.091, 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: Wed Apr 22 13:44:29 UTC 2026 EXIT_CODE=0