2026-04-16 00:22:26.516723: 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) I0416 00:22:26.633186 123305545759872 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0416 00:23:22.942243 123305545759872 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 00:23:22.942360 123305545759872 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 00:23:22.942394 123305545759872 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 I0416 00:23:22.942418 123305545759872 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 00:23:22.942441 123305545759872 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 00:23:22.942563 123305545759872 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 00:23:23.033831 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:23.439144 123305545759872 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 00:23:23.439404 123305545759872 checkpointing.py:677] Setting up checkpoint logger... I0416 00:23:23.439447 123305545759872 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0416 00:23:23.439483 123305545759872 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 00:23:23.439921 123305545759872 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 0x7024b85009b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 00:23:25.780308 123305545759872 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 00:23:25.780476 123305545759872 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 0x701e3c73dfd0>}, handler_registry=None I0416 00:23:25.780715 123305545759872 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x701e3c73dfd0>` 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`. I0416 00:23:25.780756 123305545759872 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x701e3c73cf50>` 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`. I0416 00:23:25.780784 123305545759872 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 0x701e3c73dfd0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x701e3c73dfd0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x701e3c73cf50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x701e3c73cf50>}). I0416 00:23:25.781078 123305545759872 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 00:23:25.781131 123305545759872 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x701e3c7394e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 00:23:25.920559 123305545759872 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints I0416 00:23:25.920779 123305545759872 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_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x701e3dba2db0> I0416 00:23:25.920855 123305545759872 checkpointing.py:301] Checkpoint manager created! W0416 00:23:25.936919 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:26.114634 123305545759872 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 00:23:26.114737 123305545759872 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:23:26.203128 123305545759872 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:23:26.203208 123305545759872 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:23:26.217569 123305545759872 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:23:26.217631 123305545759872 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:23:26.238224 123305545759872 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 00:23:26.238281 123305545759872 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:23:26.252545 123305545759872 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 00:23:26.252596 123305545759872 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:23:26.266839 123305545759872 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 00:23:26.266888 123305545759872 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:23:26.288444 123305545759872 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 00:23:26.288499 123305545759872 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:23:26.306699 123305545759872 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 00:23:26.306751 123305545759872 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 00:23:26.488301 123305545759872 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 00:23:26.488396 123305545759872 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0416 00:23:26.544356 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. [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 I0416 00:23:26.931484 123305545759872 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 00:23:26.931679 123305545759872 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0416 00:23:26.931724 123305545759872 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0416 00:23:26.931769 123305545759872 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0416 00:23:26.931813 123305545759872 maxtext_utils.py:1623] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0416 00:23:26.931841 123305545759872 maxtext_utils.py:1623] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0416 00:23:26.931879 123305545759872 maxtext_utils.py:1623] 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) I0416 00:23:26.931918 123305545759872 maxtext_utils.py:1623] 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') I0416 00:23:26.931946 123305545759872 maxtext_utils.py:1623] 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) I0416 00:23:26.931971 123305545759872 maxtext_utils.py:1623] 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) I0416 00:23:26.932005 123305545759872 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0416 00:23:26.932038 123305545759872 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0416 00:23:27.366091 123305545759872 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:23:27.366240 123305545759872 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 00:23:27.379719 123305545759872 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:23:27.379767 123305545759872 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 00:23:28.037622 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:28.157900 123305545759872 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. I0416 00:23:28.158382 123305545759872 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 00:23:28.158519 123305545759872 metric_logger.py:289] number parameters: 1.104 billion W0416 00:23:29.007482 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:29.103715 123305545759872 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0416 00:23:29.220950 123305545759872 checkpointing.py:776] Waited 0.11721920967102051 seconds for step 0 to finish before starting checkpointing. I0416 00:23:29.221399 123305545759872 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:23:29.221548 123305545759872 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0416 00:23:29.221904 123305545759872 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0. I0416 00:23:29.306279 123305545759872 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 00:23:29.395738 123305545759872 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:23:29.395880 123198070851136 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0 I0416 00:23:29.396282 123305545759872 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 W0416 00:23:29.610826 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.621492 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.631402 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.643310 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.655739 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.665740 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.671627 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:29.676905 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:30.065603 123198060365376 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/items I0416 00:23:30.212982 123198024713792 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776299009978176115, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776299009978176115, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0416 00:23:30.510120 123305545759872 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.115263s I0416 00:23:30.510519 123305545759872 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 10.258 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0416 00:23:30.510567 123305545759872 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.203083s (batch_requests_ready=0.083684s, total_serialization_initiated=1.119096s, others=0.000303s) I0416 00:23:30.510628 123305545759872 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.203886s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=1.203867) I0416 00:23:30.511377 123198003742272 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:23:30.511472 123305545759872 async_checkpointer.py:561] Finished blocking save. Time taken: 1.289886s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0. I0416 00:23:30.511669 123305545759872 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:23:30.511904 123198081336896 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:23:30.512009 123305545759872 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776299009.2213867, 'wait_for_prev_duration_secs': 4.076957702636719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776299009.2215693, 'checkpointer_blocking_duration_secs': 1.2900216579437256, 'get_old_steps_start_time': 1776299010.5116115, 'get_old_steps_duration_secs': 2.3365020751953125e-05, 'checkpoint_manager_blocking_start_time': 1776299009.2213054, 'checkpoint_manager_blocking_duration_secs': 1.290677547454834} I0416 00:23:30.512193 123305545759872 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0416 00:23:30.512232 123305545759872 max_utils.py:750] Memstats: After params initialized: I0416 00:23:30.512309 123305545759872 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0)) I0416 00:23:30.512334 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 00:23:30.512355 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 00:23:30.512374 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 00:23:30.512392 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 00:23:30.512408 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 00:23:30.512425 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 00:23:30.512442 123305545759872 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 00:23:30.518414 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:30.521741 3428518 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 00:23:30.521796 3428518 google_auth_provider.cc:156] Using OAuth2 AuthProvider W0416 00:23:30.530335 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:23:30.536410 3426304 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:23:30.728567 123305545759872 metric_logger.py:185] completed step: 0, seconds: 0.945, TFLOP/s/device: 14.382, Tokens/s/device: 2167.806, total_weights: 16384, loss: 10.874 I0416 00:23:30.729473 123305545759872 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/tensorboard/' I0416 00:23:30.844960 123305545759872 metric_logger.py:185] completed step: 1, seconds: 1.623, TFLOP/s/device: 8.371, Tokens/s/device: 1261.777, total_weights: 16384, loss: 10.874 I0416 00:23:30.963370 123305545759872 metric_logger.py:185] completed step: 2, seconds: 0.006, TFLOP/s/device: 2443.728, Tokens/s/device: 368345.324, total_weights: 16384, loss: 10.276 I0416 00:23:31.082221 123305545759872 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 115.885, Tokens/s/device: 17467.398, total_weights: 16384, loss: 9.282 I0416 00:23:31.086894 123305545759872 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0416 00:23:31.280725 123198049879616 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/items/array_metadatas/process_0 I0416 00:23:31.318647 123305545759872 checkpointing.py:776] Waited 0.2317211627960205 seconds for step 5 to finish before starting checkpointing. I0416 00:23:31.319036 123305545759872 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:23:59.193427 123198070851136 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 28.669627s (commit=28.230696s, array_metadata_write=0.438931s) I0416 00:23:59.194476 123198003742272 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 422.820 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 29 seconds) (per-host) I0416 00:23:59.194581 123198003742272 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.683064s. I0416 00:23:59.431711 123198003742272 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776299009978176115, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0416 00:23:59.625353 123198003742272 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 00:23:59.806957 123198024713792 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776299009978176115, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0416 00:24:00.168621 123198003742272 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.692905s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/items I0416 00:24:00.169315 123198003742272 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0/items I0416 00:24:00.415560 123198003742272 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0 I0416 00:24:01.083948 123198003742272 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0`. I0416 00:24:01.084584 123198003742272 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 31.863003s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/0 I0416 00:24:01.084660 123198003742272 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.573145s. I0416 00:24:01.084829 123198081336896 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:24:01.084938 123198081336896 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:24:01.084988 123198081336896 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:24:01.085030 123198081336896 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 00:24:01.085164 123305545759872 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. W0416 00:24:01.085276 123305545759872 checkpoint_manager.py:1441] Waiting for previous save to complete took 29.766245 seconds. If this number is high, consider checkpointing less frequently. I0416 00:24:01.085842 123305545759872 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0416 00:24:01.086128 123305545759872 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5. I0416 00:24:01.252518 123198081336896 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5 I0416 00:24:01.265714 123305545759872 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:24:01.265806 123305545759872 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 I0416 00:24:01.973149 123197940827712 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5/items I0416 00:24:05.667741 123305545759872 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 4.402490s I0416 00:24:05.671572 123305545759872 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.741 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host) I0416 00:24:05.671636 123305545759872 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.501809s (batch_requests_ready=0.092713s, total_serialization_initiated=4.405366s, others=0.003730s) I0416 00:24:05.671709 123305545759872 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.502314s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=4.502298) I0416 00:24:05.673023 123198081336896 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:24:05.673181 123305545759872 async_checkpointer.py:561] Finished blocking save. Time taken: 4.587295s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5. I0416 00:24:05.673403 123305545759872 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:24:05.673560 123198003742272 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:24:05.673646 123305545759872 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776299011.3190112, 'wait_for_prev_duration_secs': 29.766244649887085, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776299041.0858681, 'checkpointer_blocking_duration_secs': 4.587438344955444, 'get_old_steps_start_time': 1776299045.6733289, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776299011.318948, 'checkpoint_manager_blocking_duration_secs': 34.35467505455017} I0416 00:24:05.673741 123305545759872 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0416 00:24:05.674159 123305545759872 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.723, Tokens/s/device: 17292.332, total_weights: 16384, loss: 8.327 I0416 00:24:05.680203 123305545759872 metric_logger.py:185] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.465, Tokens/s/device: 17253.435, total_weights: 16384, loss: 7.367 I0416 00:24:05.796891 123305545759872 metric_logger.py:185] completed step: 6, seconds: 34.592, TFLOP/s/device: 0.393, Tokens/s/device: 59.205, total_weights: 16384, loss: 6.459 I0416 00:24:05.915161 123305545759872 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2757.688, Tokens/s/device: 415668.764, total_weights: 16384, loss: 5.622 I0416 00:24:06.034027 123305545759872 metric_logger.py:185] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.346, Tokens/s/device: 17386.137, total_weights: 16384, loss: 4.864 I0416 00:24:06.152137 123305545759872 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0416 00:24:06.152963 123305545759872 checkpointing.py:776] Waited 0.0008466243743896484 seconds for step 9 to finish before starting checkpointing. I0416 00:24:06.153216 123305545759872 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:24:06.451768 123198049879616 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5/items/array_metadatas/process_0 I0416 00:24:38.966593 123197940827712 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 33.294277s (commit=32.860242s, array_metadata_write=0.434035s) I0416 00:24:38.967599 123198081336896 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 334.327 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 37 seconds) (per-host) I0416 00:24:38.967703 123198081336896 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 33.294479s. I0416 00:24:39.412938 123198081336896 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 00:24:39.878914 123198081336896 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.635323s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5/items I0416 00:24:39.879617 123198081336896 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5/items I0416 00:24:40.127930 123198081336896 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5 I0416 00:24:40.795920 123198081336896 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5`. I0416 00:24:40.796547 123198081336896 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 39.710667s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/5 I0416 00:24:40.796622 123198081336896 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 35.123400s. I0416 00:24:40.796753 123198003742272 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:24:40.796801 123198003742272 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:24:40.796852 123198003742272 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:24:40.796891 123198003742272 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0416 00:24:40.797033 123305545759872 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. W0416 00:24:40.797164 123305545759872 checkpoint_manager.py:1441] Waiting for previous save to complete took 34.643950 seconds. If this number is high, consider checkpointing less frequently. I0416 00:24:40.797377 123305545759872 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0416 00:24:40.797664 123305545759872 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9. I0416 00:24:40.975029 123305545759872 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:24:40.975175 123198081336896 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9 I0416 00:24:40.975248 123305545759872 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 I0416 00:24:41.663637 123198070851136 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9/items I0416 00:24:52.367956 123305545759872 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.393581s I0416 00:24:52.371746 123305545759872 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.074 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host) I0416 00:24:52.371804 123305545759872 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 11.493421s (batch_requests_ready=0.093058s, total_serialization_initiated=11.396697s, others=0.003665s) I0416 00:24:52.372038 123305545759872 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 11.494304s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=11.494290) I0416 00:24:52.373302 123198049879616 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:24:52.373401 123305545759872 async_checkpointer.py:561] Finished blocking save. Time taken: 11.575981s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9. I0416 00:24:52.373605 123305545759872 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:24:52.373765 123198003742272 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:24:52.373873 123305545759872 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776299046.153189, 'wait_for_prev_duration_secs': 34.64395046234131, 'time_between_consecutive_saves_sec': 5.0681304931640625, 'checkpointer_blocking_start_time': 1776299080.7974, 'checkpointer_blocking_duration_secs': 11.576108932495117, 'get_old_steps_start_time': 1776299092.37353, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776299046.1531584, 'checkpoint_manager_blocking_duration_secs': 46.22067737579346} I0416 00:24:52.373991 123305545759872 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0416 00:24:52.374027 123305545759872 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:24:53.134377 123197368305216 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9/items/array_metadatas/process_0 I0416 00:25:28.077578 123197972284992 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.705088s (commit=35.227217s, array_metadata_write=0.477871s) I0416 00:25:28.078596 123198049879616 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 267.728 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 47 seconds) (per-host) I0416 00:25:28.078705 123198049879616 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.705260s. I0416 00:25:28.521992 123198049879616 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 00:25:29.042169 123198049879616 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.668997s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9/items I0416 00:25:29.042859 123198049879616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9/items I0416 00:25:29.292196 123198049879616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9 I0416 00:25:29.988437 123198049879616 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9`. I0416 00:25:29.989071 123198049879616 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 49.191659s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_false/checkpoints/9 I0416 00:25:29.989138 123198049879616 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 37.615695s. I0416 00:25:29.989320 123198003742272 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:25:29.989433 123198003742272 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:25:29.989486 123198003742272 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:25:29.989532 123198003742272 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 00:25:29.989668 123305545759872 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. I0416 00:25:29.989806 123305545759872 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:25:29.990437 123305545759872 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 115.034, Tokens/s/device: 17339.181, total_weights: 16384, loss: 4.189 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops