2026-04-15 22:07:34.081546: 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) I0415 22:07:34.199872 123413702483072 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 22:08:25.875583 123413702483072 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 22:08:25.875696 123413702483072 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 22:08:25.875730 123413702483072 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 I0415 22:08:25.875758 123413702483072 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 22:08:25.875779 123413702483072 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0415 22:08:25.875857 123413702483072 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 22:08:25.967309 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:08:26.371257 123413702483072 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 22:08:26.371576 123413702483072 checkpointing.py:677] Setting up checkpoint logger... I0415 22:08:26.371625 123413702483072 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True I0415 22:08:26.371665 123413702483072 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 22:08:26.372226 123413702483072 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 0x703de6ffc740>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 22:08:28.795741 123413702483072 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 22:08:28.795916 123413702483072 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 0x7038e22a8440>}, handler_registry=None I0415 22:08:28.796184 123413702483072 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7038e22a8440>` 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`. I0415 22:08:28.796226 123413702483072 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x70376ad3a330>` 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`. I0415 22:08:28.796257 123413702483072 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 0x7038e22a8440>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7038e22a8440>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x70376ad3a330>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x70376ad3a330>}). I0415 22:08:28.796549 123413702483072 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 22:08:28.796627 123413702483072 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x70376ad2d3a0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 22:08:28.921285 123413702483072 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints I0415 22:08:28.921500 123413702483072 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_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x70376ad38e90> I0415 22:08:28.921578 123413702483072 checkpointing.py:301] Checkpoint manager created! W0415 22:08:28.938185 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:08:29.119318 123413702483072 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 22:08:29.119426 123413702483072 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:08:29.214271 123413702483072 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 22:08:29.214349 123413702483072 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:08:29.228796 123413702483072 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 22:08:29.228851 123413702483072 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:08:29.249694 123413702483072 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 22:08:29.249753 123413702483072 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:08:29.264259 123413702483072 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 22:08:29.264307 123413702483072 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:08:29.278869 123413702483072 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 22:08:29.278917 123413702483072 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:08:29.301129 123413702483072 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 22:08:29.301180 123413702483072 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:08:29.319689 123413702483072 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0415 22:08:29.319742 123413702483072 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0415 22:08:29.510341 123413702483072 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 22:08:29.510512 123413702483072 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 22:08:29.563059 3283945 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 I0415 22:08:29.950165 123413702483072 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 22:08:29.950270 123413702483072 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0415 22:08:29.950308 123413702483072 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0415 22:08:29.950351 123413702483072 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0415 22:08:29.950390 123413702483072 maxtext_utils.py:1620] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0415 22:08:29.950422 123413702483072 maxtext_utils.py:1620] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0415 22:08:29.950460 123413702483072 maxtext_utils.py:1620] 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) I0415 22:08:29.950499 123413702483072 maxtext_utils.py:1620] 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') I0415 22:08:29.950528 123413702483072 maxtext_utils.py:1620] 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) I0415 22:08:29.950552 123413702483072 maxtext_utils.py:1620] 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) I0415 22:08:29.950609 123413702483072 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0415 22:08:29.950644 123413702483072 maxtext_utils.py:1620] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0415 22:08:30.379705 123413702483072 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 22:08:30.379793 123413702483072 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0415 22:08:30.393679 123413702483072 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 22:08:30.393727 123413702483072 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0415 22:08:31.044516 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:08:31.165397 123413702483072 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. I0415 22:08:31.165874 123413702483072 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 22:08:31.166012 123413702483072 metric_logger.py:289] number parameters: 1.104 billion W0415 22:08:32.004780 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:08:32.097236 123413702483072 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 22:08:32.214401 123413702483072 checkpointing.py:776] Waited 0.11715102195739746 seconds for step 0 to finish before starting checkpointing. I0415 22:08:32.214838 123413702483072 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 22:08:32.214992 123413702483072 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 22:08:32.215358 123413702483072 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0. I0415 22:08:32.304500 123413702483072 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 22:08:32.386061 123306229368384 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0 I0415 22:08:32.391561 123413702483072 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:08:32.391649 123413702483072 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 W0415 22:08:32.604461 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.616978 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.626013 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.639424 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.654697 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.659759 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.663964 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:32.668083 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:08:33.047467 123306218882624 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/items I0415 22:08:33.197036 123306187425344 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776290912950238189, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776290912950238189, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0415 22:08:33.537335 3286073 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 22:08:33.537390 3286073 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 22:08:33.953273 123413702483072 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.562107s I0415 22:08:33.957309 123413702483072 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 7.471 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0415 22:08:33.957372 123413702483072 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.651932s (batch_requests_ready=0.081960s, total_serialization_initiated=1.566031s, others=0.003940s) I0415 22:08:33.957446 123413702483072 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.652489s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=1.652470) I0415 22:08:33.958964 123306114025024 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:08:33.959102 123413702483072 async_checkpointer.py:561] Finished blocking save. Time taken: 1.744070s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0. I0415 22:08:33.959233 123413702483072 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:08:33.959474 123306239854144 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:08:33.959581 123413702483072 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776290912.2148182, 'wait_for_prev_duration_secs': 4.9114227294921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776290912.2150133, 'checkpointer_blocking_duration_secs': 1.7441487312316895, 'get_old_steps_start_time': 1776290913.9591787, 'get_old_steps_duration_secs': 2.1457672119140625e-05, 'checkpoint_manager_blocking_start_time': 1776290912.2147348, 'checkpoint_manager_blocking_duration_secs': 1.7448210716247559} I0415 22:08:33.959686 123413702483072 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 22:08:33.959724 123413702483072 max_utils.py:750] Memstats: After params initialized: I0415 22:08:33.959767 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0415 22:08:33.959791 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0415 22:08:33.959810 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0415 22:08:33.959828 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0415 22:08:33.959845 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0415 22:08:33.959861 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0415 22:08:33.959877 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0415 22:08:33.959893 123413702483072 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0415 22:08:33.965400 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:33.975256 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:08:33.980211 3283945 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:08:34.163411 123413702483072 metric_logger.py:185] completed step: 0, seconds: 0.931, TFLOP/s/device: 14.598, Tokens/s/device: 2200.371, total_weights: 16384, loss: 10.874 I0415 22:08:34.164309 123413702483072 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/tensorboard/' I0415 22:08:34.279563 123413702483072 metric_logger.py:185] completed step: 1, seconds: 2.065, TFLOP/s/device: 6.581, Tokens/s/device: 991.990, total_weights: 16384, loss: 10.874 I0415 22:08:34.398161 123413702483072 metric_logger.py:185] completed step: 2, seconds: 0.005, TFLOP/s/device: 2549.658, Tokens/s/device: 384312.254, total_weights: 16384, loss: 10.276 I0415 22:08:34.517752 123413702483072 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.076, Tokens/s/device: 17496.198, total_weights: 16384, loss: 9.282 I0415 22:08:34.524844 123413702483072 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0415 22:08:34.721747 123306197911104 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_0 I0415 22:08:34.755146 123413702483072 checkpointing.py:776] Waited 0.23036766052246094 seconds for step 5 to finish before starting checkpointing. I0415 22:08:34.755658 123413702483072 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:08:39.322740 123306134996544 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 5.364399s (commit=5.003080s, array_metadata_write=0.361319s) I0415 22:08:39.323782 123306114025024 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.758 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host) I0415 22:08:39.323889 123306114025024 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 5.364840s. I0415 22:08:39.513761 123306114025024 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776290912950238189, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA W0415 22:08:39.606433 123306114025024 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_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/items, I0415 22:08:39.689268 123306114025024 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0415 22:08:39.689373 123306114025024 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.139457s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/items I0415 22:08:39.689998 123306114025024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/items I0415 22:08:39.887621 123306187425344 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776290912950238189, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0415 22:08:39.924005 123306114025024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0 I0415 22:08:40.585540 123306114025024 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0`. I0415 22:08:40.586186 123306114025024 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 8.371161s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/0 I0415 22:08:40.586262 123306114025024 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 6.627215s. I0415 22:08:40.586447 123306239854144 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:08:40.586559 123306239854144 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:08:40.586628 123306239854144 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:08:40.586672 123306239854144 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 22:08:40.586802 123413702483072 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. W0415 22:08:40.586914 123413702483072 checkpoint_manager.py:1441] Waiting for previous save to complete took 5.831270 seconds. If this number is high, consider checkpointing less frequently. I0415 22:08:40.587674 123413702483072 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0415 22:08:40.587954 123413702483072 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5. I0415 22:08:40.758980 123413702483072 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:08:40.759086 123413702483072 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 I0415 22:08:40.762627 123306239854144 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5 I0415 22:08:41.438599 123306134996544 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5/items I0415 22:08:51.795221 123413702483072 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.036642s I0415 22:08:51.799420 123413702483072 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.109 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host) I0415 22:08:51.799487 123413702483072 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 11.130072s (batch_requests_ready=0.085969s, total_serialization_initiated=11.039997s, others=0.004105s) I0415 22:08:51.799563 123413702483072 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 11.130602s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=11.130585) I0415 22:08:51.800846 123306103539264 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:08:51.800925 123413702483072 async_checkpointer.py:561] Finished blocking save. Time taken: 11.213206s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5. I0415 22:08:51.801123 123413702483072 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:08:51.801291 123306114025024 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:08:51.801392 123413702483072 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776290914.7556272, 'wait_for_prev_duration_secs': 5.83126974105835, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776290920.5876994, 'checkpointer_blocking_duration_secs': 11.213332891464233, 'get_old_steps_start_time': 1776290931.8010597, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776290914.7555537, 'checkpoint_manager_blocking_duration_secs': 17.04581308364868} I0415 22:08:51.801511 123413702483072 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0415 22:08:51.801992 123413702483072 metric_logger.py:185] completed step: 4, seconds: 0.119, TFLOP/s/device: 113.975, Tokens/s/device: 17179.623, total_weights: 16384, loss: 8.327 I0415 22:08:51.808716 123413702483072 metric_logger.py:185] completed step: 5, seconds: 0.121, TFLOP/s/device: 112.146, Tokens/s/device: 16903.826, total_weights: 16384, loss: 7.367 I0415 22:08:51.924805 123413702483072 metric_logger.py:185] completed step: 6, seconds: 17.282, TFLOP/s/device: 0.786, Tokens/s/device: 118.502, total_weights: 16384, loss: 6.459 I0415 22:08:52.044131 123413702483072 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2473.085, Tokens/s/device: 372770.295, total_weights: 16384, loss: 5.622 I0415 22:08:52.162725 123413702483072 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 116.233, Tokens/s/device: 17519.847, total_weights: 16384, loss: 4.864 I0415 22:08:52.281301 123413702483072 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 22:08:52.282350 123413702483072 checkpointing.py:776] Waited 0.0010716915130615234 seconds for step 9 to finish before starting checkpointing. I0415 22:08:52.282686 123413702483072 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:08:52.518326 123306302768704 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_0 I0415 22:08:57.555455 123306134996544 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 5.755302s (commit=5.348505s, array_metadata_write=0.406797s) I0415 22:08:57.556344 123306103539264 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 748.319 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 16 seconds) (per-host) I0415 22:08:57.556396 123306103539264 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 5.755429s. W0415 22:08:57.858085 123306103539264 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_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5/items, I0415 22:08:57.941256 123306103539264 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0415 22:08:57.941365 123306103539264 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.138721s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5/items I0415 22:08:57.942037 123306103539264 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5/items I0415 22:08:58.182224 123306103539264 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5 I0415 22:08:58.806610 123306103539264 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5`. I0415 22:08:58.807234 123306103539264 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 18.219520s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/5 I0415 22:08:58.807307 123306103539264 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 7.006341s. I0415 22:08:58.808005 123306114025024 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:08:58.808112 123306114025024 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:08:58.808167 123306114025024 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:08:58.808208 123306114025024 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0415 22:08:58.808328 123413702483072 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. W0415 22:08:58.808441 123413702483072 checkpoint_manager.py:1441] Waiting for previous save to complete took 6.525763 seconds. If this number is high, consider checkpointing less frequently. I0415 22:08:58.808660 123413702483072 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 22:08:58.808944 123413702483072 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9. I0415 22:08:58.977108 123306114025024 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9 I0415 22:08:58.988688 123413702483072 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:08:58.988783 123413702483072 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 I0415 22:08:59.634975 123306134996544 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9/items I0415 22:09:10.577162 123413702483072 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.588873s I0415 22:09:10.577732 123413702483072 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.056 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host) I0415 22:09:10.577793 123413702483072 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 11.681520s (batch_requests_ready=0.089054s, total_serialization_initiated=11.591998s, others=0.000467s) I0415 22:09:10.577859 123413702483072 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 11.682216s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=11.682201) I0415 22:09:10.582819 123306292282944 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:09:10.582920 123413702483072 async_checkpointer.py:561] Finished blocking save. Time taken: 11.774217s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9. I0415 22:09:10.583121 123413702483072 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:09:10.583255 123306103539264 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:09:10.583327 123413702483072 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776290932.2826562, 'wait_for_prev_duration_secs': 6.5257627964019775, 'time_between_consecutive_saves_sec': 11.69596815109253, 'checkpointer_blocking_start_time': 1776290938.808684, 'checkpointer_blocking_duration_secs': 11.774346351623535, 'get_old_steps_start_time': 1776290950.5830557, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776290932.2826173, 'checkpoint_manager_blocking_duration_secs': 18.300686597824097} I0415 22:09:10.583432 123413702483072 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 22:09:10.583466 123413702483072 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:09:11.381808 123306302768704 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_0 I0415 22:09:16.587101 123306134996544 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 6.004937s (commit=5.576908s, array_metadata_write=0.428028s) I0415 22:09:16.587995 123306292282944 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 714.280 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 17 seconds) (per-host) I0415 22:09:16.588108 123306292282944 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 6.005145s. W0415 22:09:16.898175 123306292282944 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_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9/items, I0415 22:09:16.984993 123306292282944 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0415 22:09:16.985116 123306292282944 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.146108s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9/items I0415 22:09:16.985749 123306292282944 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9/items I0415 22:09:17.230844 123306292282944 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9 I0415 22:09:17.915068 123306292282944 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9`. I0415 22:09:17.915705 123306292282944 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 19.107010s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_15_ocdbt_false/checkpoints/9 I0415 22:09:17.915777 123306292282944 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 7.332814s. I0415 22:09:17.915875 123306103539264 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:09:17.915926 123306103539264 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:09:17.915971 123306103539264 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:09:17.916011 123306103539264 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 22:09:17.916140 123413702483072 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. I0415 22:09:17.916271 123413702483072 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 22:09:17.916845 123413702483072 metric_logger.py:185] completed step: 9, seconds: 0.120, TFLOP/s/device: 112.816, Tokens/s/device: 17004.882, 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