2026-04-17 14:08:49.252294: 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) I0417 14:08:49.368083 140035665976448 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0417 14:09:46.476234 140035665976448 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 14:09:46.476351 140035665976448 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 14:09:46.476391 140035665976448 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Apr 6 2026 20:48:10 (1775533690) cl/895581894 I0417 14:09:46.476419 140035665976448 train_utils.py:335] WARNING: 'dataset_path' might be pointing your local file system I0417 14:09:46.476441 140035665976448 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 14:09:46.476526 140035665976448 train.py:671] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0417 14:09:46.570311 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 14:09:46.985561 140035665976448 maxtext_utils.py:1548] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 14:09:46.985946 140035665976448 checkpointing.py:677] Setting up checkpoint logger... I0417 14:09:46.985991 140035665976448 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0417 14:09:46.986072 140035665976448 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0417 14:09:46.986761 140035665976448 base_pytree_checkpoint_handler.py:411] 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 0x7f5c01300920>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 14:09:48.645578 140035665976448 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0417 14:09:48.645787 140035665976448 checkpoint_manager.py:702] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f5585536450>}, handler_registry=None I0417 14:09:48.646125 140035665976448 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f5585536450>` 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`. I0417 14:09:48.646169 140035665976448 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f5585535880>` 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`. I0417 14:09:48.646198 140035665976448 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 0x7f5585536450>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f5585536450>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f5585535880>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f5585535880>}). I0417 14:09:48.646489 140035665976448 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0417 14:09:48.646544 140035665976448 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7f5585532200> timeout: 600 secs and primary_host=0 for async checkpoint writes I0417 14:09:49.039274 139929554257472 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776434783813243820, 'commit_timestamp_nsecs': 1776434820898167775, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0417 14:09:49.118897 139929554257472 checkpoint_manager.py:1847] Missing metrics for step 0 I0417 14:09:49.168559 140035665976448 checkpoint_manager.py:1788] Found 3 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints I0417 14:09:49.169066 140035665976448 checkpoint_manager.py:921] [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_hns=False, 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), root_directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f55859439e0> I0417 14:09:49.169162 140035665976448 checkpointing.py:301] Checkpoint manager created! W0417 14:09:49.186124 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 14:09:49.734705 140035665976448 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 14:09:49.734803 140035665976448 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 14:09:49.817645 140035665976448 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 14:09:49.817720 140035665976448 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 14:09:49.831793 140035665976448 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 14:09:49.831842 140035665976448 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 14:09:49.852235 140035665976448 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 14:09:49.852288 140035665976448 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 14:09:49.866441 140035665976448 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 14:09:49.866490 140035665976448 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 14:09:49.880709 140035665976448 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 14:09:49.880759 140035665976448 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 14:09:49.902233 140035665976448 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 14:09:49.902288 140035665976448 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 14:09:49.920446 140035665976448 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 14:09:49.920496 140035665976448 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0417 14:09:50.109471 140035665976448 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0417 14:09:50.109559 140035665976448 checkpointing.py:581] restoring from this run's directory step 9 I0417 14:09:50.275634 140035665976448 checkpointer.py:304] Restoring checkpoint from gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/9. I0417 14:09:50.788448 524192 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0417 14:09:50.788516 524192 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0417 14:09:56.931910 140035665976448 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/read/gbytes_per_sec: 1.950 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 6 seconds) (per-host) I0417 14:09:56.932344 140035665976448 checkpointer.py:318] Finished restoring checkpoint in 6.74 seconds from gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/9. I0417 14:09:56.932475 140035665976448 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776434990.1913595, 'checkpointer_duration_secs': 6.741047382354736, 'checkpoint_manager_start_time': 1776434990.1097653, 'checkpoint_manager_duration_secs': 6.822646856307983} [DECOUPLED NO-OP] gcs_storage: using stubs. [DECOUPLED NO-OP] mldiagnostics: using stub. [DECOUPLED NO-OP] mldiagnostics: using stub. [DECOUPLED NO-OP] mldiagnostics: using stub. [DECOUPLED NO-OP] workload_monitor: using stub. [DECOUPLED NO-OP] vertex_tensorboard: using stub. fsdp: 8 I0417 14:09:57.325594 140035665976448 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 14:09:57.325706 140035665976448 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 14:09:57.325747 140035665976448 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 14:09:57.325792 140035665976448 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0417 14:09:57.325832 140035665976448 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 14:09:57.325860 140035665976448 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 14:09:57.325899 140035665976448 maxtext_utils.py:1651] 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) I0417 14:09:57.325937 140035665976448 maxtext_utils.py:1651] 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') I0417 14:09:57.325966 140035665976448 maxtext_utils.py:1651] 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) I0417 14:09:57.325993 140035665976448 maxtext_utils.py:1651] 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) I0417 14:09:57.326028 140035665976448 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 14:09:57.326080 140035665976448 maxtext_utils.py:1651] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 14:09:57.759894 140035665976448 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 14:09:57.759980 140035665976448 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0417 14:09:57.774482 140035665976448 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 14:09:57.774535 140035665976448 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0417 14:09:58.002125 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 14:09:58.065645 140035665976448 max_utils.py:791] Total memory size: 3.6 GB, Output size: 1.5 GB, Temp size: 2.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB. I0417 14:09:58.066116 140035665976448 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0417 14:09:58.066278 140035665976448 metric_logger.py:301] number parameters: 1.104 billion W0417 14:09:58.879683 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 14:09:58.950109 140035665976448 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0417 14:09:59.066611 140035665976448 checkpointing.py:776] Waited 0.11649298667907715 seconds for step 10 to finish before starting checkpointing. I0417 14:09:59.067126 140035665976448 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 14:09:59.067316 140035665976448 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 10 I0417 14:09:59.067676 140035665976448 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10. I0417 14:09:59.137823 140035665976448 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0417 14:09:59.210645 140035665976448 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 14:09:59.210731 140035665976448 replica_slices.py:410] 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 I0417 14:09:59.220569 139928008656448 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10 W0417 14:09:59.435352 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.452519 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.467967 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.472655 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.477240 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.481606 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.485894 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:09:59.490155 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 14:09:59.927927 139927998170688 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10/items I0417 14:10:00.075500 139927966713408 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776434999832995220, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776434999832995220, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA I0417 14:10:01.543297 140035665976448 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.333065s I0417 14:10:01.547382 140035665976448 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.124 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0417 14:10:01.547446 140035665976448 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.408269s (batch_requests_ready=0.068438s, total_serialization_initiated=2.335842s, others=0.003989s) I0417 14:10:01.547519 140035665976448 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.409235s (all_items=0.000036s, per_item={'items': '0.00003600'}, temp_paths=2.409199) I0417 14:10:01.548799 139928019142208 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 14:10:01.548883 140035665976448 async_checkpointer.py:561] Finished blocking save. Time taken: 2.481522s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10. I0417 14:10:01.549108 140035665976448 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0417 14:10:01.549281 139927987684928 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 14:10:01.549372 140035665976448 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776434999.0671103, 'wait_for_prev_duration_secs': 4.410743713378906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776434999.0673385, 'checkpointer_blocking_duration_secs': 2.481657028198242, 'get_old_steps_start_time': 1776435001.5490158, 'get_old_steps_duration_secs': 4.982948303222656e-05, 'checkpoint_manager_blocking_start_time': 1776434999.0669305, 'checkpoint_manager_blocking_duration_secs': 2.482416868209839} I0417 14:10:01.549487 140035665976448 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0417 14:10:01.549528 140035665976448 max_utils.py:750] Memstats: After params initialized: I0417 14:10:01.549574 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0417 14:10:01.549601 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0417 14:10:01.549621 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0417 14:10:01.549640 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0417 14:10:01.549660 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0417 14:10:01.549678 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0417 14:10:01.549702 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0417 14:10:01.549721 140035665976448 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0417 14:10:01.555630 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:10:01.562632 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 14:10:01.567523 521368 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 14:10:01.782503 140035665976448 metric_logger.py:196] completed step: 10, seconds: 0.883, TFLOP/s/device: 15.382, Tokens/s/device: 2318.507, total_weights: 16384, loss: 3.602, lm_loss: 3.602, perplexity: 36.686 I0417 14:10:01.898811 140035665976448 metric_logger.py:196] completed step: 11, seconds: 2.831, TFLOP/s/device: 4.800, Tokens/s/device: 723.507, total_weights: 16384, loss: 3.101, lm_loss: 3.101, perplexity: 22.211 I0417 14:10:02.017564 140035665976448 metric_logger.py:196] completed step: 12, seconds: 0.005, TFLOP/s/device: 2588.518, Tokens/s/device: 390169.556, total_weights: 16384, loss: 2.684, lm_loss: 2.684, perplexity: 14.644 I0417 14:10:02.135534 140035665976448 metric_logger.py:196] completed step: 13, seconds: 0.117, TFLOP/s/device: 116.086, Tokens/s/device: 17497.693, total_weights: 16384, loss: 2.348, lm_loss: 2.348, perplexity: 10.461 I0417 14:10:02.140186 140035665976448 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0417 14:10:02.277900 139927977199168 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10/items/array_metadatas/process_0 I0417 14:10:02.371534 140035665976448 checkpointing.py:776] Waited 0.23131847381591797 seconds for step 15 to finish before starting checkpointing. I0417 14:10:02.371926 140035665976448 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 14:10:42.076209 139927998170688 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.528139s (commit=40.078115s, array_metadata_write=0.450024s) I0417 14:10:42.077307 139928019142208 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 294.302 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 42 seconds) (per-host) I0417 14:10:42.077363 139928019142208 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.528437s. I0417 14:10:42.502941 139928019142208 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0417 14:10:42.693447 139927966713408 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776434999832995220, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA I0417 14:10:42.992065 139928019142208 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.643839s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10/items I0417 14:10:42.992795 139928019142208 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10/items I0417 14:10:43.232058 139928019142208 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10 I0417 14:10:43.930895 139928019142208 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10`. I0417 14:10:43.931518 139928019142208 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 44.864165s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/10 I0417 14:10:43.931596 139928019142208 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 42.382671s. I0417 14:10:43.931751 139927987684928 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 14:10:43.931847 139927987684928 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 14:10:43.931921 139927987684928 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0417 14:10:43.931970 139927987684928 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0417 14:10:43.932329 140035665976448 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0417 14:10:43.933285 140035665976448 checkpoint_manager.py:1441] Waiting for previous save to complete took 41.561356 seconds. If this number is high, consider checkpointing less frequently. I0417 14:10:43.934421 140035665976448 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 15 I0417 14:10:43.934718 140035665976448 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15. I0417 14:10:44.096815 140035665976448 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 14:10:44.096905 140035665976448 replica_slices.py:410] 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 I0417 14:10:44.098479 139927987684928 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15 I0417 14:10:44.780224 139927956227648 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15/items I0417 14:10:48.068285 140035665976448 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 3.971867s I0417 14:10:48.071989 140035665976448 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.039 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host) I0417 14:10:48.072060 140035665976448 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.061208s (batch_requests_ready=0.082318s, total_serialization_initiated=3.975293s, others=0.003596s) I0417 14:10:48.072132 140035665976448 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.061760s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=4.061745) I0417 14:10:48.073458 139929491342912 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 14:10:48.073541 140035665976448 async_checkpointer.py:561] Finished blocking save. Time taken: 4.139073s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15. I0417 14:10:48.073759 140035665976448 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0417 14:10:48.073917 139928019142208 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 14:10:48.074016 140035665976448 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776435002.3719022, 'wait_for_prev_duration_secs': 41.56135582923889, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776435043.9344466, 'checkpointer_blocking_duration_secs': 4.139220952987671, 'get_old_steps_start_time': 1776435048.073688, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776435002.3718414, 'checkpoint_manager_blocking_duration_secs': 45.70214891433716} I0417 14:10:48.074148 140035665976448 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0417 14:10:48.074801 140035665976448 metric_logger.py:196] completed step: 14, seconds: 0.119, TFLOP/s/device: 114.329, Tokens/s/device: 17232.965, total_weights: 16384, loss: 2.083, lm_loss: 2.083, perplexity: 8.030 I0417 14:10:48.082616 140035665976448 metric_logger.py:196] completed step: 15, seconds: 0.118, TFLOP/s/device: 115.005, Tokens/s/device: 17334.778, total_weights: 16384, loss: 1.880, lm_loss: 1.880, perplexity: 6.551 I0417 14:10:48.198484 140035665976448 metric_logger.py:196] completed step: 16, seconds: 45.940, TFLOP/s/device: 0.296, Tokens/s/device: 44.580, total_weights: 16384, loss: 1.726, lm_loss: 1.726, perplexity: 5.616 I0417 14:10:48.316959 140035665976448 metric_logger.py:196] completed step: 17, seconds: 0.006, TFLOP/s/device: 2233.256, Tokens/s/device: 336620.644, total_weights: 16384, loss: 1.610, lm_loss: 1.610, perplexity: 5.002 I0417 14:10:48.435551 140035665976448 metric_logger.py:196] completed step: 18, seconds: 0.116, TFLOP/s/device: 116.698, Tokens/s/device: 17589.968, total_weights: 16384, loss: 1.521, lm_loss: 1.521, perplexity: 4.577 I0417 14:10:48.553217 140035665976448 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0417 14:10:48.554078 140035665976448 checkpointing.py:776] Waited 0.0008802413940429688 seconds for step 19 to finish before starting checkpointing. I0417 14:10:48.554315 140035665976448 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 14:10:48.850539 139929501828672 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15/items/array_metadatas/process_0 I0417 14:11:35.777175 139927987684928 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 47.704370s (commit=47.239201s, array_metadata_write=0.465169s) I0417 14:11:35.778340 139929491342912 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 244.107 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 51 seconds) (per-host) I0417 14:11:35.778396 139929491342912 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.704798s. I0417 14:11:36.239619 139929491342912 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0417 14:11:36.734643 139929491342912 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.648517s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15/items I0417 14:11:36.735359 139929491342912 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15/items I0417 14:11:36.993087 139929491342912 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15 I0417 14:11:37.688457 139929491342912 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15`. I0417 14:11:37.689082 139929491342912 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 53.754620s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/15 I0417 14:11:37.689157 139929491342912 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 49.615561s. I0417 14:11:37.689326 139928019142208 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 14:11:37.689428 139928019142208 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 14:11:37.689494 139928019142208 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0417 14:11:37.689536 139928019142208 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0417 14:11:37.689940 140035665976448 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0417 14:11:37.690077 140035665976448 checkpoint_manager.py:1441] Waiting for previous save to complete took 49.135761 seconds. If this number is high, consider checkpointing less frequently. I0417 14:11:37.690298 140035665976448 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 19 I0417 14:11:37.690596 140035665976448 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19. I0417 14:11:37.869317 139928019142208 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19 I0417 14:11:37.877395 140035665976448 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 14:11:37.877486 140035665976448 replica_slices.py:410] 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 I0417 14:11:38.584583 139929459885632 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19/items I0417 14:11:43.443583 140035665976448 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 5.566610s I0417 14:11:43.444134 140035665976448 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.179 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 5 seconds) (per-host) I0417 14:11:43.444185 140035665976448 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 5.663065s (batch_requests_ready=0.092377s, total_serialization_initiated=5.570246s, others=0.000443s) I0417 14:11:43.444252 140035665976448 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 5.663846s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=5.663831) I0417 14:11:43.448714 139929438914112 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 14:11:43.448877 140035665976448 async_checkpointer.py:561] Finished blocking save. Time taken: 5.758538s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19. I0417 14:11:43.449095 140035665976448 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0417 14:11:43.449245 139929491342912 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 14:11:43.449327 140035665976448 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776435048.5542934, 'wait_for_prev_duration_secs': 49.13576054573059, 'time_between_consecutive_saves_sec': 4.622308254241943, 'checkpointer_blocking_start_time': 1776435097.6903198, 'checkpointer_blocking_duration_secs': 5.758690118789673, 'get_old_steps_start_time': 1776435103.4490256, 'get_old_steps_duration_secs': 3.3855438232421875e-05, 'checkpoint_manager_blocking_start_time': 1776435048.5542626, 'checkpoint_manager_blocking_duration_secs': 54.89504098892212} I0417 14:11:43.449421 140035665976448 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0417 14:11:43.449449 140035665976448 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 14:11:44.219765 139929501828672 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19/items/array_metadatas/process_0 I0417 14:12:19.860933 139927987684928 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.413098s (commit=35.931914s, array_metadata_write=0.481184s) I0417 14:12:19.861985 139929438914112 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 300.298 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 42 seconds) (per-host) I0417 14:12:19.862108 139929438914112 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.413177s. I0417 14:12:20.302959 139929438914112 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0417 14:12:20.865989 139929438914112 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.710347s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19/items I0417 14:12:20.866712 139929438914112 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19/items I0417 14:12:21.126339 139929438914112 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19 I0417 14:12:21.810928 139929438914112 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19`. I0417 14:12:21.811559 139929438914112 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 44.121223s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_16_zarr3_false/checkpoints/19 I0417 14:12:21.811630 139929438914112 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 38.362705s. I0417 14:12:21.811744 139929491342912 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 14:12:21.811799 139929491342912 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 14:12:21.811864 139929491342912 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0417 14:12:21.811910 139929491342912 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0417 14:12:21.812067 140035665976448 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0417 14:12:21.812212 140035665976448 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 14:12:21.812968 140035665976448 metric_logger.py:196] completed step: 19, seconds: 0.118, TFLOP/s/device: 114.725, Tokens/s/device: 17292.624, total_weights: 16384, loss: 1.449, lm_loss: 1.449, perplexity: 4.260 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops