XPK Start: Mon Apr 20 16:57:22 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 16:57:47.261547: 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 16:57:47.440580 139557104953152 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 16:57:56.479996 139557104953152 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 16:57:56.482321 139557104953152 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-0ju46-slice-job-0-0.mt-15-ocdbt-true-sv-0ju46:8482 I0420 16:57:57.553808 139557104953152 max_utils.py:284] Jax distributed system initialized! I0420 16:58:03.903660 139557104953152 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 16:58:03.903761 139557104953152 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 16:58:03.903805 139557104953152 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 16:58:03.903841 139557104953152 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 16:58:04.599591 139557104953152 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:58:04.600168 139557104953152 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 16:58:04.600347 139557104953152 checkpointing.py:688] Setting up checkpoint logger... I0420 16:58:04.600395 139557104953152 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 16:58:04.600436 139557104953152 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 16:58:04.600786 139557104953152 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 0x7eec8ab171d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 16:58:08.028019 139557104953152 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 16:58:08.028257 139557104953152 checkpoint_manager.py:708] [process=2][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ed83c367e00>}, handler_registry=None I0420 16:58:08.028492 139557104953152 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ed83c367e00>` 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 16:58:08.028554 139557104953152 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ed83c36ca40>` 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 16:58:08.028589 139557104953152 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 0x7ed83c367e00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ed83c367e00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ed83c36ca40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ed83c36ca40>}). I0420 16:58:08.028912 139557104953152 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 16:58:08.028980 139557104953152 async_checkpointer.py:192] [process=2][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ed74458b740> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 16:58:09.166001 139557104953152 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_15_ocdbt_true/checkpoints I0420 16:58:09.603001 139557104953152 checkpoint_manager.py:929] [process=2][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_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ed83c36ab70> I0420 16:58:09.603170 139557104953152 checkpointing.py:302] Checkpoint manager created! I0420 16:58:10.539229 139557104953152 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 16:58:10.539336 139557104953152 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:58:10.929833 139557104953152 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:58:10.929926 139557104953152 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:58:10.946210 139557104953152 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 16:58:10.946265 139557104953152 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 16:58:10.969848 139557104953152 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 16:58:10.969915 139557104953152 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:58:10.986227 139557104953152 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 16:58:10.986288 139557104953152 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:58:11.002349 139557104953152 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 16:58:11.002409 139557104953152 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:58:11.026995 139557104953152 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 16:58:11.027065 139557104953152 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 16:58:11.047910 139557104953152 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 16:58:11.047973 139557104953152 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 16:58:11.263082 139557104953152 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 16:58:11.263189 139557104953152 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 16:58:12.683816 139557104953152 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 16:58:12.683945 139557104953152 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 16:58:12.684013 139557104953152 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 16:58:12.684080 139557104953152 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 16:58:12.684139 139557104953152 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 16:58:12.684181 139557104953152 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 16:58:12.684236 139557104953152 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 16:58:12.684290 139557104953152 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 16:58:12.684331 139557104953152 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 16:58:12.684368 139557104953152 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 16:58:12.684416 139557104953152 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 16:58:12.684464 139557104953152 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 16:58:13.182845 139557104953152 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 16:58:13.182942 139557104953152 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 16:58:13.198173 139557104953152 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 16:58:13.198234 139557104953152 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 16:58:24.168784 139557104953152 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 16:58:24.169575 139557104953152 metric_logger.py:301] number parameters: 1.104 billion I0420 16:58:35.411649 139557104953152 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 16:58:35.565598 139557104953152 checkpointing.py:798] Waited 0.15393376350402832 seconds for step 0 to finish before starting checkpointing. I0420 16:58:35.568073 139557104953152 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 16:58:35.569736 139557104953152 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 0 I0420 16:58:35.571358 139557104953152 event_tracking.py:70] [process=2] [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_15_ocdbt_true/checkpoints/0. I0420 16:58:35.917483 139557104953152 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 16:58:35.918366 139557104953152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 16:58:35.918422 139557104953152 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 16:58:36.191956 139557104953152 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274545s I0420 16:58:36.192131 139557104953152 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 5.506 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2801821231842041 s) (per-host) I0420 16:58:36.192185 139557104953152 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.280246s (batch_requests_ready=0.002486s, total_serialization_initiated=0.277689s, others=0.000071s) I0420 16:58:36.192277 139557104953152 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.284586s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.284569) I0420 16:58:36.193143 139557104953152 event_tracking.py:125] [process=2] [async] Finished blocking save in 0.62 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_15_ocdbt_true/checkpoints/0. I0420 16:58:36.193477 139428748736256 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:18:36.193440 I0420 16:58:36.201994 139557104953152 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:58:36.202234 139426065151744 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:58:36.202354 139557104953152 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776704315.5680537, 'wait_for_prev_duration_secs': 6.29425048828125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776704315.569775, 'checkpointer_blocking_duration_secs': 0.6238753795623779, 'get_old_steps_start_time': 1776704316.1936758, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776704315.5661123, 'checkpoint_manager_blocking_duration_secs': 0.6362049579620361} I0420 16:58:36.202459 139557104953152 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 16:58:36.202510 139557104953152 max_utils.py:750] Memstats: After params initialized: I0420 16:58:36.202571 139557104953152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_8(process=2,(0,2,0,0)) I0420 16:58:36.202602 139557104953152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_9(process=2,(1,2,0,0)) I0420 16:58:36.202630 139557104953152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_12(process=2,(0,3,0,0)) I0420 16:58:36.202655 139557104953152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_13(process=2,(1,3,0,0)) I0420 16:58:36.516307 139557104953152 metric_logger.py:196] completed step: 0, seconds: 11.242, TFLOP/s/device: 1.209, Tokens/s/device: 182.175, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 16:58:36.708100 139557104953152 metric_logger.py:196] completed step: 1, seconds: 1.103, TFLOP/s/device: 12.317, Tokens/s/device: 1856.483, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 16:58:37.128074 139557104953152 metric_logger.py:196] completed step: 2, seconds: 0.044, TFLOP/s/device: 305.796, Tokens/s/device: 46092.906, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0420 16:58:37.275234 139557104953152 metric_logger.py:196] completed step: 3, seconds: 0.420, TFLOP/s/device: 32.351, Tokens/s/device: 4876.318, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0420 16:58:37.279725 139557104953152 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0420 16:58:37.569275 139557104953152 checkpointing.py:798] Waited 0.28951430320739746 seconds for step 5 to finish before starting checkpointing. I0420 16:58:37.572076 139557104953152 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 16:58:40.034665 2825 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 16:58:42.095986 139426102572800 array_metadata_store.py:203] [process=2][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_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_2 I0420 16:59:12.530110 139428748736256 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 43.137 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.61812090873718 s) (per-host) I0420 16:59:12.530234 139428748736256 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.336623s. I0420 16:59:23.397594 139428748736256 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 47.203967s. I0420 16:59:23.397902 139426065151744 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 16:59:23.397959 139426065151744 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save. I0420 16:59:23.398008 139426065151744 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 16:59:23.400723 139426065151744 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 16:59:23.400903 139557104953152 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 16:59:23.401038 139557104953152 checkpoint_manager.py:1452] Waiting for previous save to complete took 45.828960 seconds. If this number is high, consider checkpointing less frequently. I0420 16:59:23.403953 139557104953152 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 5 I0420 16:59:23.406371 139557104953152 event_tracking.py:70] [process=2] [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_15_ocdbt_true/checkpoints/5. I0420 16:59:23.711892 139557104953152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 16:59:23.711987 139557104953152 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 16:59:23.754032 139557104953152 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043253s I0420 16:59:23.754190 139557104953152 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 33.089 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04661917686462402 s) (per-host) I0420 16:59:23.754240 139557104953152 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046679s (batch_requests_ready=0.001749s, total_serialization_initiated=0.044863s, others=0.000068s) I0420 16:59:23.754324 139557104953152 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051702s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.051687) I0420 16:59:23.755043 139557104953152 event_tracking.py:125] [process=2] [async] Finished blocking save in 0.35 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_15_ocdbt_true/checkpoints/5. I0420 16:59:23.755388 139426065151744 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:19:23.755352 I0420 16:59:23.757503 139557104953152 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0420 16:59:23.757832 139403196348160 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save. I0420 16:59:23.758023 139557104953152 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776704317.5720475, 'wait_for_prev_duration_secs': 45.828959941864014, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776704363.4039931, 'checkpointer_blocking_duration_secs': 0.35155415534973145, 'get_old_steps_start_time': 1776704363.7555714, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776704317.5699773, 'checkpoint_manager_blocking_duration_secs': 46.18800401687622} I0420 16:59:23.758160 139557104953152 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0420 16:59:23.759126 139557104953152 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.581, Tokens/s/device: 13351.849, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0420 16:59:23.765275 139557104953152 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.427, Tokens/s/device: 13931.594, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0420 16:59:30.808274 139426102572800 array_metadata_store.py:203] [process=2][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_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_2 I0420 16:59:47.249044 139557104953152 metric_logger.py:196] completed step: 6, seconds: 46.484, TFLOP/s/device: 0.292, Tokens/s/device: 44.058, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0420 16:59:47.396329 139557104953152 metric_logger.py:196] completed step: 7, seconds: 23.337, TFLOP/s/device: 0.582, Tokens/s/device: 87.756, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0420 16:59:47.543447 139557104953152 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.105, Tokens/s/device: 13430.917, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0420 16:59:47.690208 139557104953152 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 16:59:47.690874 139557104953152 checkpointing.py:798] Waited 0.0006873607635498047 seconds for step 9 to finish before starting checkpointing. I0420 16:59:47.692926 139557104953152 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:00:06.634564 139426065151744 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 36.797 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.92694687843323 s) (per-host) I0420 17:00:06.634671 139426065151744 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.879167s. I0420 17:00:15.141296 139426065151744 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 51.385776s. I0420 17:00:15.141590 139403196348160 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:00:15.141733 139403196348160 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:00:15.141802 139403196348160 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:00:15.143389 139403196348160 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0420 17:00:15.143593 139557104953152 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0420 17:00:15.143719 139557104953152 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.450798 seconds. If this number is high, consider checkpointing less frequently. I0420 17:00:15.145207 139557104953152 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 9 I0420 17:00:15.147376 139557104953152 event_tracking.py:70] [process=2] [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_15_ocdbt_true/checkpoints/9. I0420 17:00:15.864154 139557104953152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:00:15.864251 139557104953152 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 17:00:15.905174 139557104953152 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041957s I0420 17:00:15.905339 139557104953152 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 33.963 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04541921615600586 s) (per-host) I0420 17:00:15.905390 139557104953152 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045479s (batch_requests_ready=0.001660s, total_serialization_initiated=0.043754s, others=0.000065s) I0420 17:00:15.905478 139557104953152 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049675s (all_items=0.000010s, per_item={'items': '0.00001049'}, temp_paths=0.049664) I0420 17:00:15.906229 139557104953152 event_tracking.py:125] [process=2] [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_15_ocdbt_true/checkpoints/9. I0420 17:00:15.906584 139403196348160 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:20:15.906546 I0420 17:00:15.908507 139557104953152 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:00:15.908725 139428748736256 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:00:15.908845 139557104953152 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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776704387.6928954, 'wait_for_prev_duration_secs': 27.45079779624939, 'time_between_consecutive_saves_sec': 24.292130947113037, 'checkpointer_blocking_start_time': 1776704415.1452458, 'checkpointer_blocking_duration_secs': 0.7614889144897461, 'get_old_steps_start_time': 1776704415.9067533, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776704387.691127, 'checkpoint_manager_blocking_duration_secs': 28.21768832206726} I0420 17:00:15.908954 139557104953152 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 17:00:15.908996 139557104953152 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:00:21.984064 139427162605312 array_metadata_store.py:203] [process=2][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_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_2 I0420 17:00:58.912787 139403196348160 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 36.690 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.05283188819885 s) (per-host) I0420 17:00:58.912892 139403196348160 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.006192s. I0420 17:01:07.410420 139403196348160 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 51.503703s. I0420 17:01:07.410710 139428748736256 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:01:07.410840 139428748736256 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:01:07.410881 139428748736256 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:01:07.412234 139428748736256 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 17:01:07.412393 139557104953152 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 17:01:07.412559 139557104953152 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 17:01:07.413657 139557104953152 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.353, Tokens/s/device: 13920.514, 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 17:01:19 UTC 2026 EXIT_CODE=0