XPK Start: Sat Apr 18 20:46:20 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 20:46:44.739531: 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) I0418 20:46:44.919739 133855713437504 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 20:46:53.962220 133855713437504 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 20:46:53.964596 133855713437504 distributed.py:172] Connecting to JAX distributed service on mt-02-synthetic-mluc1-slice-job-0-0.mt-02-synthetic-mluc1:8482 I0418 20:46:55.526878 133855713437504 max_utils.py:284] Jax distributed system initialized! I0418 20:47:01.726409 133855713437504 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 20:47:01.726511 133855713437504 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 20:47:01.726554 133855713437504 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0418 20:47:01.726590 133855713437504 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 20:47:02.412431 133855713437504 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 20:47:02.413033 133855713437504 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 20:47:02.413217 133855713437504 checkpointing.py:688] Setting up checkpoint logger... I0418 20:47:02.413269 133855713437504 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 20:47:02.413311 133855713437504 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 20:47:02.413645 133855713437504 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x79bd201ea6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 20:47:05.709137 133855713437504 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0418 20:47:05.709472 133855713437504 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a8e8521a00>}, handler_registry=None I0418 20:47:05.709709 133855713437504 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a8e8521a00>` 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`. I0418 20:47:05.709759 133855713437504 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79a8e8524620>` 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`. I0418 20:47:05.709795 133855713437504 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 0x79a8e8521a00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a8e8521a00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79a8e8524620>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79a8e8524620>}). I0418 20:47:05.710132 133855713437504 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 20:47:05.710206 133855713437504 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79a8e8497880> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 20:47:08.229368 133855713437504 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints I0418 20:47:09.595794 133855713437504 checkpoint_manager.py:929] [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_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=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79a8e8523080> I0418 20:47:09.596004 133855713437504 checkpointing.py:302] Checkpoint manager created! I0418 20:47:10.531794 133855713437504 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 20:47:10.531910 133855713437504 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 20:47:10.916664 133855713437504 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 20:47:10.916757 133855713437504 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 20:47:10.933067 133855713437504 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 20:47:10.933127 133855713437504 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 20:47:10.956652 133855713437504 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 20:47:10.956717 133855713437504 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 20:47:10.973124 133855713437504 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 20:47:10.973184 133855713437504 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 20:47:10.989504 133855713437504 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 20:47:10.989565 133855713437504 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 20:47:11.014067 133855713437504 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 20:47:11.014137 133855713437504 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 20:47:11.034935 133855713437504 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 20:47:11.034995 133855713437504 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 20:47:11.253196 133855713437504 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0418 20:47:11.253302 133855713437504 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 20:47:12.671516 133855713437504 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 20:47:12.671641 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 20:47:12.671693 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 20:47:12.671758 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0418 20:47:12.671834 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 20:47:12.671881 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 20:47:12.671936 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0418 20:47:12.671989 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0418 20:47:12.672030 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0418 20:47:12.672068 133855713437504 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0418 20:47:12.672115 133855713437504 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 20:47:12.672164 133855713437504 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 20:47:13.166585 133855713437504 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 20:47:13.166676 133855713437504 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 20:47:13.182316 133855713437504 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 20:47:13.182379 133855713437504 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 20:47:23.752968 133855713437504 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0418 20:47:23.761725 133855713437504 metric_logger.py:301] number parameters: 1.104 billion I0418 20:47:35.388723 133855713437504 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0418 20:47:35.542691 133855713437504 checkpointing.py:798] Waited 0.15395188331604004 seconds for step 0 to finish before starting checkpointing. I0418 20:47:35.545223 133855713437504 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 20:47:35.546969 133855713437504 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0418 20:47:35.548504 133855713437504 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0. I0418 20:47:36.266288 133855713437504 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 20:47:36.280822 133855713437504 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 20:47:36.280892 133855713437504 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0418 20:47:36.525835 133727489353472 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0 I0418 20:47:36.550973 133855713437504 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.270962s I0418 20:47:36.551946 133855713437504 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.560 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.277449369430542 s) (per-host) I0418 20:47:36.552052 133855713437504 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.277583s (batch_requests_ready=0.002143s, total_serialization_initiated=0.274523s, others=0.000916s) I0418 20:47:36.552164 133855713437504 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.281518s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.281501) I0418 20:47:36.553039 133855713437504 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0. I0418 20:47:36.553326 133727497746176 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 21:07:36.553292 I0418 20:47:36.567464 133855713437504 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 20:47:36.567778 133726914725632 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 20:47:36.567947 133855713437504 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776545255.5451996, 'wait_for_prev_duration_secs': 0.00010657310485839844, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776545255.5470076, 'checkpointer_blocking_duration_secs': 1.0064587593078613, 'get_old_steps_start_time': 1776545256.5534918, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776545255.5434384, 'checkpoint_manager_blocking_duration_secs': 1.024468183517456} I0418 20:47:36.568056 133855713437504 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0418 20:47:36.568116 133855713437504 max_utils.py:750] Memstats: After params initialized: I0418 20:47:36.568167 133855713437504 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0418 20:47:36.568198 133855713437504 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0418 20:47:36.568224 133855713437504 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0418 20:47:36.568249 133855713437504 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0418 20:47:36.923985 133855713437504 metric_logger.py:196] completed step: 0, seconds: 11.347, TFLOP/s/device: 1.197, Tokens/s/device: 180.495, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 20:47:36.927410 133855713437504 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/tensorboard/' I0418 20:47:37.348143 133855713437504 metric_logger.py:196] completed step: 1, seconds: 1.533, TFLOP/s/device: 8.860, Tokens/s/device: 1335.530, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 20:47:37.494105 133855713437504 metric_logger.py:196] completed step: 2, seconds: 0.425, TFLOP/s/device: 31.980, Tokens/s/device: 4820.434, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0418 20:47:37.641667 133855713437504 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 1421.099, Tokens/s/device: 214203.535, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0418 20:47:38.916765 133727472568064 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/items I0418 20:47:39.059566 133725837154048 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776545258657407999, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776545258657407999, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA I0418 20:47:40.464917 2799 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 20:47:43.510089 133726931511040 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/items/array_metadatas/process_0 I0418 20:48:00.019921 133855713437504 metric_logger.py:196] completed step: 4, seconds: 0.146, TFLOP/s/device: 92.832, Tokens/s/device: 13992.607, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0418 20:48:00.028319 133855713437504 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 91.957, Tokens/s/device: 13860.783, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0418 20:48:00.174512 133855713437504 metric_logger.py:196] completed step: 6, seconds: 22.377, TFLOP/s/device: 0.607, Tokens/s/device: 91.524, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0418 20:48:00.321848 133855713437504 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2113.741, Tokens/s/device: 318606.098, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0418 20:48:00.469065 133855713437504 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.395, Tokens/s/device: 13926.762, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0418 20:48:00.615900 133855713437504 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0418 20:48:00.616544 133855713437504 checkpointing.py:798] Waited 0.0006682872772216797 seconds for step 9 to finish before starting checkpointing. I0418 20:48:00.618680 133855713437504 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 20:48:18.026098 133726923118336 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.567367s (commit=36.294044s, array_metadata_write=1.273323s) I0418 20:48:18.027343 133727497746176 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.832 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.75282955169678 s) (per-host) I0418 20:48:18.027477 133727497746176 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.474042s. I0418 20:48:19.230330 133727497746176 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776545258657407999, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA I0418 20:48:19.973370 133725837154048 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776545258657407999, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA I0418 20:48:20.705085 133727497746176 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 20:48:20.818272 133727497746176 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000237s. I0418 20:48:21.628984 133727497746176 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.271666s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/items I0418 20:48:21.630639 133727497746176 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/items I0418 20:48:22.581079 133727497746176 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0 I0418 20:48:25.343152 133727497746176 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0`. I0418 20:48:25.344004 133727497746176 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.80 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0 I0418 20:48:25.345483 133727497746176 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 48.792046s. I0418 20:48:25.345681 133726914725632 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 20:48:25.345799 133726914725632 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 20:48:25.345874 133726914725632 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 20:48:25.347702 133726914725632 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 20:48:25.347836 133855713437504 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0418 20:48:25.347910 133855713437504 checkpoint_manager.py:1452] Waiting for previous save to complete took 24.729232 seconds. If this number is high, consider checkpointing less frequently. I0418 20:48:25.349948 133855713437504 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0418 20:48:25.351980 133855713437504 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9. I0418 20:48:26.071678 133855713437504 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 20:48:26.071774 133855713437504 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0418 20:48:26.098523 133855713437504 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027595s I0418 20:48:26.099316 133855713437504 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 47.947 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.032172203063964844 s) (per-host) I0418 20:48:26.099383 133855713437504 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032255s (batch_requests_ready=0.001925s, total_serialization_initiated=0.029591s, others=0.000740s) I0418 20:48:26.099454 133855713437504 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036288s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.036271) I0418 20:48:26.100133 133855713437504 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9. I0418 20:48:26.100465 133726914725632 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 21:08:26.100428 I0418 20:48:26.107924 133855713437504 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 20:48:26.108233 133726374381312 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 20:48:26.108395 133855713437504 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776545280.618648, 'wait_for_prev_duration_secs': 24.7292320728302, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776545305.3499908, 'checkpointer_blocking_duration_secs': 0.7506110668182373, 'get_old_steps_start_time': 1776545306.1006255, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776545280.6167657, 'checkpoint_manager_blocking_duration_secs': 25.49159598350525} I0418 20:48:26.108513 133855713437504 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0418 20:48:26.108559 133855713437504 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 20:48:26.306549 133727497746176 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9 I0418 20:48:28.901654 133723710285568 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9/items I0418 20:48:33.894074 133726399559424 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9/items/array_metadatas/process_0 I0418 20:49:13.459520 133726382774016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.143734s (commit=40.574350s, array_metadata_write=1.569384s) I0418 20:49:13.460711 133726914725632 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.329 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.393515825271606 s) (per-host) I0418 20:49:13.460849 133726914725632 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.360273s. I0418 20:49:15.723875 133726914725632 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000234s. I0418 20:49:16.134492 133726914725632 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 20:49:17.054988 133726914725632 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.725111s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9/items I0418 20:49:17.056549 133726914725632 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9/items I0418 20:49:17.566606 133726914725632 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9 I0418 20:49:19.019752 133726914725632 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9`. I0418 20:49:19.020596 133726914725632 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.67 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9 I0418 20:49:19.021918 133726914725632 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.921341s. I0418 20:49:19.022140 133726374381312 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 20:49:19.022269 133726374381312 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 20:49:19.022332 133726374381312 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 20:49:19.024305 133726374381312 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 20:49:19.024505 133855713437504 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0418 20:49:19.024648 133855713437504 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 20:49:19.025585 133855713437504 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.246, Tokens/s/device: 13904.259, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 20:49:30 UTC 2026 EXIT_CODE=0