2026-04-22 18:52:32.150845: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1776883952.163842 2085216 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776883952.167699 2085216 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776883952.179265 2085216 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883952.179282 2085216 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883952.179284 2085216 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776883952.179285 2085216 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} `rope_scaling`'s factor field must be a float >= 1, got 40 `rope_scaling`'s beta_fast field must be a float, got 32 `rope_scaling`'s beta_slow field must be a float, got 1 Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} 2026-04-22 18:52:36.209195: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0422 18:52:36.346647 131560432293696 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0422 18:53:32.115718 131560432293696 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:53:32.115826 131560432293696 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:53:32.115868 131560432293696 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0422 18:53:32.115894 131560432293696 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system I0422 18:53:32.115917 131560432293696 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:53:32.116008 131560432293696 train.py:705] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0422 18:53:32.776993 131560432293696 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:53:32.777259 131560432293696 checkpointing.py:677] Setting up checkpoint logger... I0422 18:53:32.777303 131560432293696 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True I0422 18:53:32.777338 131560432293696 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0422 18:53:32.777808 131560432293696 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x77a705dadc70>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:53:35.209795 131560432293696 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0422 18:53:35.210213 131560432293696 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 0x77a52d369640>}, handler_registry=None I0422 18:53:35.210440 131560432293696 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77a52d369640>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0422 18:53:35.210482 131560432293696 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77a52d369d60>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0422 18:53:35.210511 131560432293696 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 0x77a52d369640>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77a52d369640>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77a52d369d60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77a52d369d60>}). I0422 18:53:35.210798 131560432293696 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0422 18:53:35.210849 131560432293696 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x77a52d378900> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:53:35.352283 131560432293696 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints I0422 18:53:35.352516 131560432293696 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_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x77a52d369b80> I0422 18:53:35.352790 131560432293696 checkpointing.py:301] Checkpoint manager created! I0422 18:53:35.739989 131560432293696 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:53:35.740094 131560432293696 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:53:36.019975 131560432293696 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:53:36.020071 131560432293696 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:53:36.032397 131560432293696 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:53:36.032455 131560432293696 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:53:36.051580 131560432293696 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:53:36.051641 131560432293696 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:53:36.064092 131560432293696 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:53:36.064139 131560432293696 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:53:36.096280 131560432293696 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:53:36.096363 131560432293696 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:53:36.117543 131560432293696 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:53:36.117619 131560432293696 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:53:36.133455 131560432293696 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:53:36.133507 131560432293696 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0422 18:53:36.284951 131560432293696 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:53:36.285059 131560432293696 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. [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 I0422 18:53:36.920105 131560432293696 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0422 18:53:36.920217 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0422 18:53:36.920258 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0422 18:53:36.920302 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0422 18:53:36.920341 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0422 18:53:36.920373 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0422 18:53:36.920413 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0422 18:53:36.920453 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0422 18:53:36.920482 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0422 18:53:36.920508 131560432293696 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0422 18:53:36.920543 131560432293696 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0422 18:53:36.920578 131560432293696 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0422 18:53:37.326622 131560432293696 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0422 18:53:37.326709 131560432293696 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0422 18:53:37.338311 131560432293696 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0422 18:53:37.338358 131560432293696 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0422 18:53:38.094714 131560432293696 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. I0422 18:53:38.095278 131560432293696 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0422 18:53:38.095427 131560432293696 metric_logger.py:301] number parameters: 1.104 billion I0422 18:53:40.920980 131560432293696 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0422 18:53:41.039500 131560432293696 checkpointing.py:776] Waited 0.11851286888122559 seconds for step 0 to finish before starting checkpointing. I0422 18:53:41.040077 131560432293696 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:53:41.040231 131560432293696 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0422 18:53:41.040609 131560432293696 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0. I0422 18:53:41.117707 131560432293696 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0422 18:53:41.211241 131560432293696 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:53:41.211331 131560432293696 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 I0422 18:53:41.217248 131529302869568 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0 I0422 18:53:42.024000 131529292383808 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/items I0422 18:53:42.167370 131528906507840 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776884021923799415, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776884021923799415, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0422 18:53:42.516154 2087413 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0422 18:53:42.516207 2087413 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0422 18:53:42.773006 131560432293696 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.562171s I0422 18:53:42.777154 131560432293696 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 7.443 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0422 18:53:42.777216 131560432293696 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.658120s (batch_requests_ready=0.087930s, total_serialization_initiated=1.566158s, others=0.004032s) I0422 18:53:42.777285 131560432293696 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.659083s (all_items=0.000021s, per_item={'items': '0.00002074'}, temp_paths=1.659062) I0422 18:53:42.778499 131529313355328 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:53:42.778586 131560432293696 async_checkpointer.py:561] Finished blocking save. Time taken: 1.738317s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0. I0422 18:53:42.778791 131560432293696 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:53:42.779086 131528101201472 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:53:42.779193 131560432293696 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776884021.0400639, 'wait_for_prev_duration_secs': 4.363059997558594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776884021.040252, 'checkpointer_blocking_duration_secs': 1.738452434539795, 'get_old_steps_start_time': 1776884022.7787216, 'get_old_steps_duration_secs': 3.457069396972656e-05, 'checkpoint_manager_blocking_start_time': 1776884021.0399601, 'checkpoint_manager_blocking_duration_secs': 1.7392075061798096} I0422 18:53:42.779305 131560432293696 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0422 18:53:42.779346 131560432293696 max_utils.py:750] Memstats: After params initialized: I0422 18:53:42.779395 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0422 18:53:42.779418 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0422 18:53:42.779437 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0422 18:53:42.779454 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0422 18:53:42.779471 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0422 18:53:42.779487 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0422 18:53:42.779504 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0422 18:53:42.779520 131560432293696 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0422 18:53:43.034381 131560432293696 metric_logger.py:196] completed step: 0, seconds: 2.825, TFLOP/s/device: 4.810, Tokens/s/device: 724.949, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0422 18:53:43.035176 131560432293696 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/tensorboard/' I0422 18:53:43.150962 131560432293696 metric_logger.py:196] completed step: 1, seconds: 2.111, TFLOP/s/device: 6.435, Tokens/s/device: 969.948, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0422 18:53:43.270178 131560432293696 metric_logger.py:196] completed step: 2, seconds: 0.006, TFLOP/s/device: 2222.661, Tokens/s/device: 335023.720, total_weights: 16384, loss: 10.276, lm_loss: 10.276, perplexity: 29019.973 I0422 18:53:43.389682 131560432293696 metric_logger.py:196] completed step: 3, seconds: 0.118, TFLOP/s/device: 115.351, Tokens/s/device: 17387.023, total_weights: 16384, loss: 9.282, lm_loss: 9.282, perplexity: 10739.224 I0422 18:53:43.398011 131560432293696 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0422 18:53:43.536322 131528493368896 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_0 I0422 18:53:43.628409 131560432293696 checkpointing.py:776] Waited 0.23033857345581055 seconds for step 5 to finish before starting checkpointing. I0422 18:53:43.629834 131560432293696 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:53:50.136871 131528302528064 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 7.358913s (commit=6.890356s, array_metadata_write=0.468556s) I0422 18:53:50.137873 131529313355328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.368 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 9 seconds) (per-host) I0422 18:53:50.137929 131529313355328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 7.359299s. I0422 18:53:50.384508 131529313355328 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776884021923799415, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA W0422 18:53:50.469684 131529313355328 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/items, I0422 18:53:50.588871 131529313355328 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0422 18:53:50.588986 131529313355328 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.169455s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/items I0422 18:53:50.589646 131529313355328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/items I0422 18:53:50.795478 131528906507840 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776884021923799415, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0422 18:53:50.817867 131529313355328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0 I0422 18:53:51.518857 131529313355328 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0`. I0422 18:53:51.519515 131529313355328 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 10.479250s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/0 I0422 18:53:51.519591 131529313355328 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.740962s. I0422 18:53:51.519789 131528101201472 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:53:51.519902 131528101201472 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:53:51.519973 131528101201472 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:53:51.520023 131528101201472 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 18:53:51.520189 131560432293696 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0422 18:53:51.520245 131560432293696 checkpoint_manager.py:1441] Waiting for previous save to complete took 7.890441 seconds. If this number is high, consider checkpointing less frequently. I0422 18:53:51.521485 131560432293696 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0422 18:53:51.521764 131560432293696 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5. I0422 18:53:51.685294 131529313355328 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5 I0422 18:53:51.707333 131560432293696 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:53:51.707418 131560432293696 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 I0422 18:53:52.389766 131528493368896 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5/items I0422 18:53:53.725784 131560432293696 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.018862s I0422 18:53:53.729712 131560432293696 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.809 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0422 18:53:53.729775 131560432293696 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.124347s (batch_requests_ready=0.098201s, total_serialization_initiated=2.022331s, others=0.003815s) I0422 18:53:53.729861 131560432293696 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.124896s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=2.124881) I0422 18:53:53.731435 131528101201472 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:53:53.731518 131560432293696 async_checkpointer.py:561] Finished blocking save. Time taken: 2.209992s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5. I0422 18:53:53.731706 131560432293696 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:53:53.731857 131529313355328 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:53:53.731959 131560432293696 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776884023.6297922, 'wait_for_prev_duration_secs': 7.890441179275513, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776884031.5215087, 'checkpointer_blocking_duration_secs': 2.2101118564605713, 'get_old_steps_start_time': 1776884033.7316391, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776884023.6296837, 'checkpoint_manager_blocking_duration_secs': 10.102251291275024} I0422 18:53:53.732101 131560432293696 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0422 18:53:53.732788 131560432293696 metric_logger.py:196] completed step: 4, seconds: 0.122, TFLOP/s/device: 111.650, Tokens/s/device: 16829.096, total_weights: 16384, loss: 8.327, lm_loss: 8.327, perplexity: 4134.812 I0422 18:53:53.742000 131560432293696 metric_logger.py:196] completed step: 5, seconds: 0.120, TFLOP/s/device: 113.542, Tokens/s/device: 17114.301, total_weights: 16384, loss: 7.367, lm_loss: 7.367, perplexity: 1583.522 I0422 18:53:53.858335 131560432293696 metric_logger.py:196] completed step: 6, seconds: 10.342, TFLOP/s/device: 1.314, Tokens/s/device: 198.023, total_weights: 16384, loss: 6.459, lm_loss: 6.459, perplexity: 638.363 I0422 18:53:53.979745 131560432293696 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2372.883, Tokens/s/device: 357666.783, total_weights: 16384, loss: 5.622, lm_loss: 5.622, perplexity: 276.577 I0422 18:53:54.098677 131560432293696 metric_logger.py:196] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.305, Tokens/s/device: 17380.088, total_weights: 16384, loss: 4.864, lm_loss: 4.864, perplexity: 129.523 I0422 18:53:54.221671 131560432293696 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0422 18:53:54.248324 131560432293696 checkpointing.py:776] Waited 0.02677464485168457 seconds for step 9 to finish before starting checkpointing. I0422 18:53:54.254937 131560432293696 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:53:54.500002 131529122514496 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_0 I0422 18:54:02.965654 131528493368896 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 9.235446s (commit=8.717240s, array_metadata_write=0.518205s) I0422 18:54:02.966732 131528101201472 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.086 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host) I0422 18:54:02.966852 131528101201472 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 9.235289s. W0422 18:54:03.380594 131528101201472 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5/items, I0422 18:54:03.514916 131528101201472 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0422 18:54:03.515031 131528101201472 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.211083s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5/items I0422 18:54:03.515681 131528101201472 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5/items I0422 18:54:03.741599 131528101201472 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5 I0422 18:54:04.423517 131528101201472 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5`. I0422 18:54:04.424215 131528101201472 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 12.902694s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/5 I0422 18:54:04.424289 131528101201472 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 10.692730s. I0422 18:54:04.424520 131529313355328 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:54:04.424635 131529313355328 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:54:04.424694 131529313355328 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:54:04.424736 131529313355328 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 18:54:04.424906 131560432293696 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0422 18:54:04.425022 131560432293696 checkpoint_manager.py:1441] Waiting for previous save to complete took 10.170096 seconds. If this number is high, consider checkpointing less frequently. I0422 18:54:04.425336 131560432293696 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0422 18:54:04.425624 131560432293696 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9. I0422 18:54:04.596102 131528101201472 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9 I0422 18:54:04.603684 131560432293696 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:54:04.603778 131560432293696 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 I0422 18:54:05.268438 131529122514496 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9/items I0422 18:54:09.815446 131560432293696 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 5.212163s I0422 18:54:09.819505 131560432293696 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.325 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 5 seconds) (per-host) I0422 18:54:09.819566 131560432293696 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 5.307420s (batch_requests_ready=0.088111s, total_serialization_initiated=5.215360s, others=0.003949s) I0422 18:54:09.819637 131560432293696 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 5.308175s (all_items=0.000014s, per_item={'items': '0.00001431'}, temp_paths=5.308161) I0422 18:54:09.821303 131529313355328 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:54:09.821385 131560432293696 async_checkpointer.py:561] Finished blocking save. Time taken: 5.396008s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9. I0422 18:54:09.821547 131560432293696 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:54:09.821687 131528101201472 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:54:09.821793 131560432293696 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776884034.2549071, 'wait_for_prev_duration_secs': 10.170096158981323, 'time_between_consecutive_saves_sec': 2.7348690032958984, 'checkpointer_blocking_start_time': 1776884044.42536, 'checkpointer_blocking_duration_secs': 5.396129608154297, 'get_old_steps_start_time': 1776884049.8214993, 'get_old_steps_duration_secs': 1.8358230590820312e-05, 'checkpoint_manager_blocking_start_time': 1776884034.2548723, 'checkpoint_manager_blocking_duration_secs': 15.566896200180054} I0422 18:54:09.821932 131560432293696 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0422 18:54:09.821964 131560432293696 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:54:10.663408 131529112028736 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_0 I0422 18:54:16.683272 131529122514496 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 6.863075s (commit=6.425231s, array_metadata_write=0.437844s) I0422 18:54:16.684156 131529313355328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.014 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0422 18:54:16.684207 131529313355328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 6.862778s. W0422 18:54:17.027134 131529313355328 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9/items, I0422 18:54:17.148493 131529313355328 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0422 18:54:17.148607 131529313355328 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.171031s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9/items I0422 18:54:17.149301 131529313355328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9/items I0422 18:54:17.396035 131529313355328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9 I0422 18:54:18.147023 131529313355328 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9`. I0422 18:54:18.147758 131529313355328 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 13.722387s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_15_ocdbt_false/checkpoints/9 I0422 18:54:18.147819 131529313355328 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.326392s. I0422 18:54:18.148000 131528101201472 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:54:18.148116 131528101201472 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:54:18.148161 131528101201472 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:54:18.148200 131528101201472 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 18:54:18.148337 131560432293696 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0422 18:54:18.148485 131560432293696 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:54:18.149262 131560432293696 metric_logger.py:196] completed step: 9, seconds: 0.134, TFLOP/s/device: 101.106, Tokens/s/device: 15239.796, total_weights: 16384, loss: 4.189, lm_loss: 4.189, perplexity: 65.979 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops