XPK Start: Mon Apr 20 17:24:41 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 17:25:06.092527: 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 17:25:06.272110 136515988789056 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 17:25:15.311972 136515988789056 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 17:25:15.314506 136515988789056 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-nkpy7-slice-job-0-0.mt-16-zarr3-false-sv-nkpy7:8482 I0420 17:25:16.672247 136515988789056 max_utils.py:284] Jax distributed system initialized! I0420 17:25:22.862390 136515988789056 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 17:25:22.862490 136515988789056 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 17:25:22.862529 136515988789056 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 17:25:22.862562 136515988789056 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 17:25:23.556356 136515988789056 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 17:25:23.556932 136515988789056 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 17:25:23.557120 136515988789056 checkpointing.py:688] Setting up checkpoint logger... I0420 17:25:23.557170 136515988789056 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0420 17:25:23.557212 136515988789056 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 17:25:23.557527 136515988789056 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7c28634501a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 17:25:26.539031 136515988789056 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 17:25:26.539311 136515988789056 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 0x7c14542b8650>}, handler_registry=None I0420 17:25:26.539587 136515988789056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c14542b8650>` 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 17:25:26.539638 136515988789056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c14541726c0>` 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 17:25:26.539674 136515988789056 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 0x7c14542b8650>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c14542b8650>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c14541726c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c14541726c0>}). I0420 17:25:26.539989 136515988789056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 17:25:26.540059 136515988789056 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c143c10f740> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 17:25:27.748656 136515988789056 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_16_zarr3_false/checkpoints I0420 17:25:27.791809 136515988789056 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c1454172150> I0420 17:25:27.791952 136515988789056 checkpointing.py:302] Checkpoint manager created! I0420 17:25:28.721225 136515988789056 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 17:25:28.721328 136515988789056 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 17:25:29.100669 136515988789056 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 17:25:29.100758 136515988789056 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 17:25:29.117022 136515988789056 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 17:25:29.117089 136515988789056 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 17:25:29.140502 136515988789056 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 17:25:29.140575 136515988789056 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:25:29.156831 136515988789056 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 17:25:29.156893 136515988789056 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:25:29.173292 136515988789056 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 17:25:29.173350 136515988789056 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:25:29.198429 136515988789056 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 17:25:29.198500 136515988789056 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:25:29.219476 136515988789056 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 17:25:29.219542 136515988789056 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 17:25:29.439750 136515988789056 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 17:25:29.439874 136515988789056 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 17:25:30.853041 136515988789056 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 17:25:30.853246 136515988789056 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 17:25:30.853301 136515988789056 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 17:25:30.853360 136515988789056 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 17:25:30.853414 136515988789056 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 17:25:30.853453 136515988789056 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 17:25:30.853514 136515988789056 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 17:25:30.853567 136515988789056 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 17:25:30.853607 136515988789056 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 17:25:30.853642 136515988789056 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 17:25:30.853696 136515988789056 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 17:25:30.853744 136515988789056 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 17:25:31.347236 136515988789056 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 17:25:31.347333 136515988789056 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 17:25:31.362850 136515988789056 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 17:25:31.362916 136515988789056 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 17:25:41.921734 136515988789056 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 17:25:41.929653 136515988789056 metric_logger.py:301] number parameters: 1.104 billion I0420 17:25:53.558881 136515988789056 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 17:25:53.711524 136515988789056 checkpointing.py:798] Waited 0.15262436866760254 seconds for step 0 to finish before starting checkpointing. I0420 17:25:53.713857 136515988789056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 17:25:53.715550 136515988789056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0420 17:25:53.717185 136515988789056 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_16_zarr3_false/checkpoints/0. I0420 17:25:54.434073 136515988789056 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 17:25:54.448826 136515988789056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:25:54.448895 136515988789056 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 17:25:54.716524 136515988789056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.268466s I0420 17:25:54.717440 136515988789056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.613 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2748391628265381 s) (per-host) I0420 17:25:54.717506 136515988789056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.274923s (batch_requests_ready=0.002167s, total_serialization_initiated=0.271931s, others=0.000825s) I0420 17:25:54.717579 136515988789056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.278843s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.278826) I0420 17:25:54.718348 136515988789056 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.00 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_16_zarr3_false/checkpoints/0. I0420 17:25:54.718616 136386979215104 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:45:54.718594 I0420 17:25:54.733932 136515988789056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:25:54.734230 136384296404736 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:25:54.734386 136515988789056 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776705953.7138383, 'wait_for_prev_duration_secs': 6.532669067382812e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776705953.715588, 'checkpointer_blocking_duration_secs': 1.0031201839447021, 'get_old_steps_start_time': 1776705954.7187326, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776705953.7119393, 'checkpoint_manager_blocking_duration_secs': 1.0224096775054932} I0420 17:25:54.734486 136515988789056 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 17:25:54.734534 136515988789056 max_utils.py:750] Memstats: After params initialized: I0420 17:25:54.734581 136515988789056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 17:25:54.734614 136515988789056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 17:25:54.734646 136515988789056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 17:25:54.734670 136515988789056 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 17:25:54.762701 136386970822400 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_16_zarr3_false/checkpoints/0 I0420 17:25:55.081916 136515988789056 metric_logger.py:196] completed step: 0, seconds: 11.318, TFLOP/s/device: 1.200, Tokens/s/device: 180.943, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 17:25:55.085521 136515988789056 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_16_zarr3_false/tensorboard/' I0420 17:25:55.514848 136515988789056 metric_logger.py:196] completed step: 1, seconds: 1.521, TFLOP/s/device: 8.932, Tokens/s/device: 1346.365, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 17:25:55.660965 136515988789056 metric_logger.py:196] completed step: 2, seconds: 0.434, TFLOP/s/device: 31.329, Tokens/s/device: 4722.278, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0420 17:25:55.808295 136515988789056 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1441.452, Tokens/s/device: 217271.377, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0420 17:25:55.817564 136515988789056 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0420 17:25:56.103487 136515988789056 checkpointing.py:798] Waited 0.28588223457336426 seconds for step 5 to finish before starting checkpointing. I0420 17:25:56.106473 136515988789056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:25:57.280977 136382190188288 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776705956802718081, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776705956802718081, "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_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0420 17:25:57.374385 136386476762880 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_16_zarr3_false/checkpoints/0/items I0420 17:25:58.543886 2863 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 17:26:00.757746 136384828090112 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_16_zarr3_false/checkpoints/0/items/array_metadatas/process_0 I0420 17:26:40.874228 136384819697408 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.342810s (commit=41.125350s, array_metadata_write=1.217460s) I0420 17:26:40.875567 136386979215104 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.019 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.43294906616211 s) (per-host) I0420 17:26:40.875666 136386979215104 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.156981s. I0420 17:26:41.601792 136386979215104 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776705956802718081, '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_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0420 17:26:42.360837 136382190188288 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776705956802718081, '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_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0420 17:26:42.472875 136386979215104 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000384s. I0420 17:26:44.067897 136386979215104 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.338692s. use_zarr3=False, 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_16_zarr3_false/checkpoints/0/items I0420 17:26:44.070265 136386979215104 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_16_zarr3_false/checkpoints/0/items I0420 17:26:44.652263 136386979215104 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_16_zarr3_false/checkpoints/0 I0420 17:26:46.492779 136386979215104 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_16_zarr3_false/checkpoints/0`. I0420 17:26:46.493899 136386979215104 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.78 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_16_zarr3_false/checkpoints/0 I0420 17:26:46.495556 136386979215104 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.776866s. I0420 17:26:46.495844 136384296404736 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:26:46.496000 136384296404736 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:26:46.496124 136384296404736 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:26:46.497850 136384296404736 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 17:26:46.498062 136515988789056 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 17:26:46.498200 136515988789056 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.391736 seconds. If this number is high, consider checkpointing less frequently. I0420 17:26:46.500105 136515988789056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5 I0420 17:26:46.502498 136515988789056 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_16_zarr3_false/checkpoints/5. I0420 17:26:47.222951 136515988789056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:26:47.223099 136515988789056 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 17:26:47.259612 136515988789056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037732s I0420 17:26:47.260514 136515988789056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 33.219 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04643607139587402 s) (per-host) I0420 17:26:47.260609 136515988789056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046551s (batch_requests_ready=0.004273s, total_serialization_initiated=0.041489s, others=0.000789s) I0420 17:26:47.260730 136515988789056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051524s (all_items=0.000024s, per_item={'items': '0.00002408'}, temp_paths=0.051500) I0420 17:26:47.261785 136515988789056 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_16_zarr3_false/checkpoints/5. I0420 17:26:47.262257 136384296404736 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:46:47.262230 I0420 17:26:47.264636 136515988789056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:26:47.264949 136385357960960 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:26:47.265203 136515988789056 standard_logger.py:34] {'step': 5, '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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776705956.1064277, 'wait_for_prev_duration_secs': 50.391735553741455, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776706006.5001698, 'checkpointer_blocking_duration_secs': 0.7622265815734863, 'get_old_steps_start_time': 1776706007.2624254, 'get_old_steps_duration_secs': 4.1961669921875e-05, 'checkpoint_manager_blocking_start_time': 1776705956.1042109, 'checkpoint_manager_blocking_duration_secs': 51.16094183921814} I0420 17:26:47.265348 136515988789056 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0420 17:26:47.479170 136386979215104 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_16_zarr3_false/checkpoints/5 I0420 17:26:49.660399 136386459977472 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_16_zarr3_false/checkpoints/5/items I0420 17:26:52.602949 136386447095552 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_16_zarr3_false/checkpoints/5/items/array_metadatas/process_0 I0420 17:27:10.321208 136515988789056 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.483, Tokens/s/device: 13940.033, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0420 17:27:10.330146 136515988789056 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.185, Tokens/s/device: 13895.108, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0420 17:27:10.475720 136515988789056 metric_logger.py:196] completed step: 6, seconds: 74.511, TFLOP/s/device: 0.182, Tokens/s/device: 27.486, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0420 17:27:10.623435 136515988789056 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1994.587, Tokens/s/device: 300645.919, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0420 17:27:10.770384 136515988789056 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.601, Tokens/s/device: 13957.799, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0420 17:27:10.917053 136515988789056 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 17:27:10.917804 136515988789056 checkpointing.py:798] Waited 0.0007727146148681641 seconds for step 9 to finish before starting checkpointing. I0420 17:27:10.919746 136515988789056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:27:33.483149 136385374746368 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.695305s (commit=41.394926s, array_metadata_write=1.300379s) I0420 17:27:33.484252 136384296404736 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.139 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.270182609558105 s) (per-host) I0420 17:27:33.484323 136384296404736 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.221961s. I0420 17:27:35.201100 136384296404736 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000191s. I0420 17:27:36.507775 136384296404736 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.952259s. use_zarr3=False, 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_16_zarr3_false/checkpoints/5/items I0420 17:27:36.509439 136384296404736 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_16_zarr3_false/checkpoints/5/items I0420 17:27:37.091033 136384296404736 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_16_zarr3_false/checkpoints/5 I0420 17:27:38.591944 136384296404736 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_16_zarr3_false/checkpoints/5`. I0420 17:27:38.592819 136384296404736 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.09 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_16_zarr3_false/checkpoints/5 I0420 17:27:38.594228 136384296404736 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.331863s. I0420 17:27:38.594420 136385357960960 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:27:38.594542 136385357960960 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:27:38.594608 136385357960960 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:27:38.596309 136385357960960 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0420 17:27:38.596488 136515988789056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0420 17:27:38.596613 136515988789056 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.676873 seconds. If this number is high, consider checkpointing less frequently. I0420 17:27:38.598137 136515988789056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0420 17:27:38.600301 136515988789056 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_16_zarr3_false/checkpoints/9. I0420 17:27:38.902658 136515988789056 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:27:38.902751 136515988789056 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 17:27:38.929461 136515988789056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027744s I0420 17:27:38.930333 136515988789056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 47.792 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03227663040161133 s) (per-host) I0420 17:27:38.930405 136515988789056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032368s (batch_requests_ready=0.001910s, total_serialization_initiated=0.029673s, others=0.000786s) I0420 17:27:38.930483 136515988789056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036327s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.036312) I0420 17:27:38.931247 136515988789056 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.33 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_16_zarr3_false/checkpoints/9. I0420 17:27:38.931548 136386476762880 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:47:38.931513 I0420 17:27:38.946693 136515988789056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:27:38.946956 136385357960960 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:27:38.947153 136515988789056 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776706030.9197145, 'wait_for_prev_duration_secs': 27.676873445510864, 'time_between_consecutive_saves_sec': 24.421794891357422, 'checkpointer_blocking_start_time': 1776706058.5981772, 'checkpointer_blocking_duration_secs': 0.3335127830505371, 'get_old_steps_start_time': 1776706058.9317074, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776706030.9180248, 'checkpoint_manager_blocking_duration_secs': 28.02909278869629} I0420 17:27:38.947273 136515988789056 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 17:27:38.947319 136515988789056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:27:39.149303 136384296404736 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_16_zarr3_false/checkpoints/9 I0420 17:27:41.437281 136386447095552 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_16_zarr3_false/checkpoints/9/items I0420 17:27:44.402434 136387013842688 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_16_zarr3_false/checkpoints/9/items/array_metadatas/process_0 I0420 17:28:37.951930 136386979215104 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 55.368644s (commit=54.080397s, array_metadata_write=1.288246s) I0420 17:28:37.956237 136386476762880 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 26.746 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.05813694000244 s) (per-host) I0420 17:28:37.956356 136386476762880 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.024690s. I0420 17:28:39.467792 136386476762880 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000231s. I0420 17:28:40.595013 136386476762880 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.781179s. use_zarr3=False, 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_16_zarr3_false/checkpoints/9/items I0420 17:28:40.596922 136386476762880 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_16_zarr3_false/checkpoints/9/items I0420 17:28:41.173706 136386476762880 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_16_zarr3_false/checkpoints/9 I0420 17:28:42.690595 136386476762880 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_16_zarr3_false/checkpoints/9`. I0420 17:28:42.691382 136386476762880 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 64.09 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_16_zarr3_false/checkpoints/9 I0420 17:28:42.692705 136386476762880 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 63.761037s. I0420 17:28:42.692843 136385357960960 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:28:42.692914 136385357960960 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:28:42.692968 136385357960960 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:28:42.694687 136385357960960 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 17:28:42.694889 136515988789056 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 17:28:42.695043 136515988789056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 17:28:42.696039 136515988789056 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.100, Tokens/s/device: 13882.299, 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: Mon Apr 20 17:28:53 UTC 2026 EXIT_CODE=0