XPK Start: Mon Apr 20 16:28:48 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 16:29:12.833846: 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 16:29:13.011966 139362144098112 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 16:29:22.053135 139362144098112 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 16:29:22.055550 139362144098112 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-za8ny-slice-job-0-0.mt-11-optimizer-offload-za8ny:8482 I0420 16:29:22.800649 139362144098112 max_utils.py:284] Jax distributed system initialized! I0420 16:29:27.986669 139362144098112 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 16:29:27.986767 139362144098112 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 16:29:27.986809 139362144098112 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 16:29:27.986846 139362144098112 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 16:29:29.033628 139362144098112 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:29:29.034223 139362144098112 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:29:29.034405 139362144098112 checkpointing.py:688] Setting up checkpoint logger... I0420 16:29:29.034455 139362144098112 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 16:29:29.034498 139362144098112 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 16:29:29.034823 139362144098112 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 0x7ebf0f93cb60>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 16:29:31.713730 139362144098112 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 16:29:31.713969 139362144098112 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 0x7eab4462aed0>}, handler_registry=None I0420 16:29:31.714222 139362144098112 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eab4462aed0>` 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 16:29:31.714273 139362144098112 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ebe031ae090>` 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 16:29:31.714310 139362144098112 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 0x7eab4462aed0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eab4462aed0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ebe031ae090>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ebe031ae090>}). I0420 16:29:31.714630 139362144098112 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 16:29:31.714703 139362144098112 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7eab4429ac00> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 16:29:32.837624 139362144098112 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_11_optimizer_offload_false/checkpoints I0420 16:29:33.288240 139362144098112 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_11_optimizer_offload_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ebe031ae930> I0420 16:29:33.288409 139362144098112 checkpointing.py:302] Checkpoint manager created! I0420 16:29:34.205021 139362144098112 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 16:29:34.205163 139362144098112 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:29:34.581098 139362144098112 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:29:34.581190 139362144098112 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:29:34.597528 139362144098112 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:29:34.597600 139362144098112 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:29:34.621056 139362144098112 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 16:29:34.621139 139362144098112 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:29:34.637452 139362144098112 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 16:29:34.637515 139362144098112 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:29:34.653865 139362144098112 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 16:29:34.653934 139362144098112 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:29:34.678490 139362144098112 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 16:29:34.678563 139362144098112 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:29:34.700182 139362144098112 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 16:29:34.700253 139362144098112 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 16:29:34.911345 139362144098112 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 16:29:34.911451 139362144098112 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 16:29:36.330145 139362144098112 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 16:29:36.330268 139362144098112 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 16:29:36.330322 139362144098112 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 16:29:36.330403 139362144098112 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 16:29:36.330463 139362144098112 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 16:29:36.330504 139362144098112 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 16:29:36.330558 139362144098112 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 16:29:36.330609 139362144098112 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 16:29:36.330649 139362144098112 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 16:29:36.330692 139362144098112 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 16:29:36.330741 139362144098112 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 16:29:36.330790 139362144098112 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 16:29:36.819889 139362144098112 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 16:29:36.819980 139362144098112 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 16:29:36.835356 139362144098112 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 16:29:36.835419 139362144098112 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 16:29:47.865150 139362144098112 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 16:29:47.872886 139362144098112 metric_logger.py:301] number parameters: 1.104 billion I0420 16:29:59.684237 139362144098112 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 16:29:59.840408 139362144098112 checkpointing.py:798] Waited 0.1561579704284668 seconds for step 0 to finish before starting checkpointing. I0420 16:29:59.842724 139362144098112 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:29:59.844439 139362144098112 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0420 16:29:59.845892 139362144098112 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_11_optimizer_offload_false/checkpoints/0. I0420 16:30:00.590131 139362144098112 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 16:30:00.604136 139362144098112 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 16:30:00.604206 139362144098112 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 16:30:00.876763 139234568558336 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_11_optimizer_offload_false/checkpoints/0 I0420 16:30:00.878808 139362144098112 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275484s I0420 16:30:00.879709 139362144098112 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.474 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2817845344543457 s) (per-host) I0420 16:30:00.879769 139362144098112 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281862s (batch_requests_ready=0.002060s, total_serialization_initiated=0.278941s, others=0.000862s) I0420 16:30:00.879841 139362144098112 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285863s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.285845) I0420 16:30:00.880549 139362144098112 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_11_optimizer_offload_false/checkpoints/0. I0420 16:30:00.880847 139234585343744 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 16:50:00.880814 I0420 16:30:00.896499 139362144098112 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:30:00.896747 139233511601920 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:30:00.896879 139362144098112 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_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776702599.8427057, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776702599.8444793, 'checkpointer_blocking_duration_secs': 1.036515474319458, 'get_old_steps_start_time': 1776702600.8810189, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776702599.840884, 'checkpoint_manager_blocking_duration_secs': 1.0559513568878174} I0420 16:30:00.896986 139362144098112 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 16:30:00.897035 139362144098112 max_utils.py:750] Memstats: After params initialized: I0420 16:30:00.897098 139362144098112 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 16:30:00.897133 139362144098112 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 16:30:00.897160 139362144098112 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 16:30:00.897185 139362144098112 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 16:30:01.243294 139362144098112 metric_logger.py:196] completed step: 0, seconds: 11.499, TFLOP/s/device: 1.182, Tokens/s/device: 178.107, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 16:30:01.246890 139362144098112 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_11_optimizer_offload_false/tensorboard/' I0420 16:30:01.655124 139362144098112 metric_logger.py:196] completed step: 1, seconds: 1.557, TFLOP/s/device: 8.725, Tokens/s/device: 1315.157, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 16:30:01.801175 139362144098112 metric_logger.py:196] completed step: 2, seconds: 0.412, TFLOP/s/device: 32.940, Tokens/s/device: 4965.149, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0420 16:30:01.948472 139362144098112 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 1396.273, Tokens/s/device: 210461.412, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0420 16:30:02.710192 139234551772928 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_11_optimizer_offload_false/checkpoints/0/items I0420 16:30:02.831861 139232793175808 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776702602420841844, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776702602420841844, "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_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA I0420 16:30:03.865530 2902 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 16:30:05.672984 139234006083328 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_11_optimizer_offload_false/checkpoints/0/items/array_metadatas/process_0 I0420 16:30:24.023638 139362144098112 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.567, Tokens/s/device: 13952.760, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0420 16:30:24.032526 139362144098112 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.340, Tokens/s/device: 13918.433, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0420 16:30:24.178640 139362144098112 metric_logger.py:196] completed step: 6, seconds: 22.074, TFLOP/s/device: 0.616, Tokens/s/device: 92.779, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0420 16:30:24.326285 139362144098112 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2037.662, Tokens/s/device: 307138.572, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0420 16:30:24.473351 139362144098112 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.291, Tokens/s/device: 13911.059, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0420 16:30:24.619868 139362144098112 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 16:30:24.620544 139362144098112 checkpointing.py:798] Waited 0.0006997585296630859 seconds for step 9 to finish before starting checkpointing. I0420 16:30:24.622588 139362144098112 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:30:40.610096 139233997690624 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.751482s (commit=35.532968s, array_metadata_write=1.218515s) I0420 16:30:40.611158 139234585343744 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 39.477 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.01321005821228 s) (per-host) I0420 16:30:40.611269 139234585343744 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.730312s. I0420 16:30:41.355239 139234585343744 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776702602420841844, '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_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA I0420 16:30:42.089818 139232793175808 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776702602420841844, '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_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA I0420 16:30:42.120163 139234585343744 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000178s. I0420 16:30:42.929254 139234585343744 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 16:30:43.850882 139234585343744 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.368931s. 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_11_optimizer_offload_false/checkpoints/0/items I0420 16:30:43.852632 139234585343744 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_11_optimizer_offload_false/checkpoints/0/items I0420 16:30:44.384882 139234585343744 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_11_optimizer_offload_false/checkpoints/0 I0420 16:30:45.849178 139234585343744 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_11_optimizer_offload_false/checkpoints/0`. I0420 16:30:45.849861 139234585343744 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 46.01 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_11_optimizer_offload_false/checkpoints/0 I0420 16:30:45.851263 139234585343744 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 44.970303s. I0420 16:30:45.851450 139233511601920 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:30:45.851573 139233511601920 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:30:45.851638 139233511601920 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:30:45.853525 139233511601920 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 16:30:45.853710 139362144098112 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 16:30:45.853852 139362144098112 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.231262 seconds. If this number is high, consider checkpointing less frequently. I0420 16:30:45.855726 139362144098112 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0420 16:30:45.857728 139362144098112 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_11_optimizer_offload_false/checkpoints/9. I0420 16:30:46.587604 139362144098112 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 16:30:46.587697 139362144098112 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 16:30:46.613365 139362144098112 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026516s I0420 16:30:46.614168 139362144098112 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.854 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03094196319580078 s) (per-host) I0420 16:30:46.614230 139362144098112 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031020s (batch_requests_ready=0.001787s, total_serialization_initiated=0.028516s, others=0.000717s) I0420 16:30:46.614305 139362144098112 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035116s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.035099) I0420 16:30:46.615039 139362144098112 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_11_optimizer_offload_false/checkpoints/9. I0420 16:30:46.615347 139233511601920 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 16:50:46.615315 I0420 16:30:46.628400 139362144098112 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:30:46.628710 139232806942464 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:30:46.628875 139362144098112 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_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776702624.6225564, 'wait_for_prev_duration_secs': 21.23126220703125, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776702645.8557653, 'checkpointer_blocking_duration_secs': 0.759716272354126, 'get_old_steps_start_time': 1776702646.615507, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776702624.6207774, 'checkpoint_manager_blocking_duration_secs': 22.008065223693848} I0420 16:30:46.628980 139362144098112 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 16:30:46.629023 139362144098112 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:30:46.843513 139234585343744 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_11_optimizer_offload_false/checkpoints/9 I0420 16:30:48.653412 139234006083328 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_11_optimizer_offload_false/checkpoints/9/items I0420 16:30:51.841343 139232823727872 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_11_optimizer_offload_false/checkpoints/9/items/array_metadatas/process_0 I0420 16:31:31.706412 139232815335168 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.892277s (commit=40.644311s, array_metadata_write=1.247966s) I0420 16:31:31.707459 139233511601920 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.006 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.12420153617859 s) (per-host) I0420 16:31:31.707588 139233511601920 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.092131s. I0420 16:31:33.447340 139233511601920 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000245s. I0420 16:31:33.635407 139233511601920 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 16:31:34.601786 139233511601920 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.977940s. 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_11_optimizer_offload_false/checkpoints/9/items I0420 16:31:34.603612 139233511601920 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_11_optimizer_offload_false/checkpoints/9/items I0420 16:31:35.159039 139233511601920 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_11_optimizer_offload_false/checkpoints/9 I0420 16:31:36.619363 139233511601920 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_11_optimizer_offload_false/checkpoints/9`. I0420 16:31:36.620100 139233511601920 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.76 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_11_optimizer_offload_false/checkpoints/9 I0420 16:31:36.621551 139233511601920 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.006093s. I0420 16:31:36.621769 139232806942464 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:31:36.621887 139232806942464 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:31:36.621961 139232806942464 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:31:36.624456 139232806942464 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 16:31:36.624632 139362144098112 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 16:31:36.624782 139362144098112 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:31:36.625811 139362144098112 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.119, Tokens/s/device: 13885.216, 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 16:31:49 UTC 2026 EXIT_CODE=0