XPK Start: Tue Apr 21 07:42:37 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. `rope_parameters`'s factor field must be a float >= 1, got 40 `rope_parameters`'s beta_fast field must be a float, got 32 `rope_parameters`'s beta_slow field must be a float, got 1 DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 2026-04-21 07:43:02.032903: 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) I0421 07:43:02.241881 136206255867712 max_utils.py:273] Attempting to initialize the jax distributed system... I0421 07:43:11.283191 136206255867712 distributed.py:149] Starting JAX distributed service on [::]:8482 I0421 07:43:11.285728 136206255867712 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-gowi6-slice-job-0-0.mt-16-zarr3-false-sv-gowi6:8482 I0421 07:43:12.502489 136206255867712 max_utils.py:284] Jax distributed system initialized! I0421 07:43:18.735981 136206255867712 max_utils.py:800] System Information: Jax Version: 0.9.2 I0421 07:43:18.736086 136206255867712 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0421 07:43:18.736125 136206255867712 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 I0421 07:43:18.736161 136206255867712 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 07:43:19.432278 136206255867712 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 07:43:19.432562 136206255867712 checkpointing.py:677] Setting up checkpoint logger... I0421 07:43:19.432616 136206255867712 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0421 07:43:19.432676 136206255867712 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0421 07:43:19.433017 136206255867712 base_pytree_checkpoint_handler.py:441] 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 0x7be045f38fe0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 07:43:22.758280 136206255867712 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 07:43:22.758624 136206255867712 checkpoint_manager.py:708] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bdf39262fc0>}, handler_registry=None I0421 07:43:22.758877 136206255867712 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bdf39262fc0>` 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`. I0421 07:43:22.758926 136206255867712 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bcc50481c10>` 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`. I0421 07:43:22.758962 136206255867712 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 0x7bdf39262fc0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bdf39262fc0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bcc50481c10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bcc50481c10>}). I0421 07:43:22.759279 136206255867712 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0421 07:43:22.759351 136206255867712 async_checkpointer.py:192] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bcb48795580> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0421 07:43:23.918870 136206255867712 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints I0421 07:43:23.941732 136206255867712 checkpoint_manager.py:929] [process=4][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=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, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bcc5047ab10> I0421 07:43:23.941858 136206255867712 checkpointing.py:301] Checkpoint manager created! I0421 07:43:24.865872 136206255867712 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 07:43:24.865991 136206255867712 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 07:43:25.249296 136206255867712 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 07:43:25.249391 136206255867712 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 07:43:25.265856 136206255867712 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 07:43:25.265916 136206255867712 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 07:43:25.289575 136206255867712 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 07:43:25.289656 136206255867712 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 07:43:25.306106 136206255867712 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 07:43:25.306167 136206255867712 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 07:43:25.322696 136206255867712 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 07:43:25.322755 136206255867712 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 07:43:25.347565 136206255867712 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 07:43:25.347635 136206255867712 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 07:43:25.368874 136206255867712 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 07:43:25.368945 136206255867712 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 07:43:25.584012 136206255867712 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 07:43:25.584126 136206255867712 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 07:43:27.004399 136206255867712 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0421 07:43:27.004528 136206255867712 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 07:43:27.004588 136206255867712 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 07:43:27.004659 136206255867712 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0421 07:43:27.004716 136206255867712 maxtext_utils.py:1668] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0421 07:43:27.004756 136206255867712 maxtext_utils.py:1668] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0421 07:43:27.004809 136206255867712 maxtext_utils.py:1668] 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) I0421 07:43:27.004860 136206255867712 maxtext_utils.py:1668] 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') I0421 07:43:27.004899 136206255867712 maxtext_utils.py:1668] 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) I0421 07:43:27.004934 136206255867712 maxtext_utils.py:1668] 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) I0421 07:43:27.004981 136206255867712 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 07:43:27.005027 136206255867712 maxtext_utils.py:1668] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 07:43:27.501902 136206255867712 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 07:43:27.501998 136206255867712 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 07:43:27.517616 136206255867712 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 07:43:27.517683 136206255867712 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 07:43:38.366546 136206255867712 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. I0421 07:43:38.367291 136206255867712 metric_logger.py:301] number parameters: 1.104 billion I0421 07:43:49.383898 136206255867712 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 07:43:50.075062 136206255867712 checkpointing.py:776] Waited 0.6911430358886719 seconds for step 0 to finish before starting checkpointing. I0421 07:43:50.077504 136206255867712 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 07:43:50.079295 136206255867712 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 0 I0421 07:43:50.080679 136206255867712 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/0. I0421 07:43:50.827591 136206255867712 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 07:43:50.828500 136206255867712 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 07:43:50.828556 136206255867712 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 I0421 07:43:51.103585 136206255867712 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276070s I0421 07:43:51.103772 136206255867712 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 5.472 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28190040588378906 s) (per-host) I0421 07:43:51.103825 136206255867712 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281964s (batch_requests_ready=0.002618s, total_serialization_initiated=0.279273s, others=0.000072s) I0421 07:43:51.103914 136206255867712 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286775s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.286757) I0421 07:43:51.104758 136206255867712 event_tracking.py:125] [process=4] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/0. I0421 07:43:51.105065 136077247510272 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 08:03:51.105031 I0421 07:43:51.115226 136206255867712 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 07:43:51.115536 136076125652736 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 07:43:51.115726 136206255867712 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776757430.077486, 'wait_for_prev_duration_secs': 6.365776062011719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776757430.0793347, 'checkpointer_blocking_duration_secs': 1.0258803367614746, 'get_old_steps_start_time': 1776757431.1052403, 'get_old_steps_duration_secs': 3.4332275390625e-05, 'checkpoint_manager_blocking_start_time': 1776757430.0756283, 'checkpoint_manager_blocking_duration_secs': 1.0400588512420654} I0421 07:43:51.115835 136206255867712 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 07:43:51.115890 136206255867712 max_utils.py:750] Memstats: After params initialized: I0421 07:43:51.115935 136206255867712 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0421 07:43:51.115967 136206255867712 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0421 07:43:51.115993 136206255867712 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0421 07:43:51.116017 136206255867712 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0421 07:43:51.433220 136206255867712 metric_logger.py:196] completed step: 0, seconds: 11.016, TFLOP/s/device: 1.233, Tokens/s/device: 185.903, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 07:43:51.612695 136206255867712 metric_logger.py:196] completed step: 1, seconds: 2.048, TFLOP/s/device: 6.635, Tokens/s/device: 1000.164, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 07:43:52.052893 136206255867712 metric_logger.py:196] completed step: 2, seconds: 0.032, TFLOP/s/device: 421.137, Tokens/s/device: 63478.288, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0421 07:43:52.200116 136206255867712 metric_logger.py:196] completed step: 3, seconds: 0.440, TFLOP/s/device: 30.855, Tokens/s/device: 4650.793, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0421 07:43:52.205173 136206255867712 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0421 07:43:52.494352 136206255867712 checkpointing.py:776] Waited 0.28914451599121094 seconds for step 5 to finish before starting checkpointing. I0421 07:43:52.497264 136206255867712 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 07:43:54.489792 2799 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 07:43:57.460722 136076134045440 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/0/items/array_metadatas/process_4 I0421 07:44:30.422548 136077247510272 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 39.888 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 39.60064101219177 s) (per-host) I0421 07:44:30.422677 136077247510272 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.317506s. I0421 07:44:40.483259 136077247510272 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 49.378072s. I0421 07:44:40.483552 136076125652736 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 07:44:40.483700 136076125652736 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 07:44:40.483761 136076125652736 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 07:44:40.486033 136076125652736 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 07:44:40.486259 136206255867712 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 07:44:40.486402 136206255867712 checkpoint_manager.py:1452] Waiting for previous save to complete took 47.989145 seconds. If this number is high, consider checkpointing less frequently. I0421 07:44:40.488424 136206255867712 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 5 I0421 07:44:40.490930 136206255867712 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/5. I0421 07:44:41.196359 136206255867712 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 07:44:41.196457 136206255867712 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 I0421 07:44:41.234437 136206255867712 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039176s I0421 07:44:41.234614 136206255867712 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 36.135 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.042688846588134766 s) (per-host) I0421 07:44:41.234678 136206255867712 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042763s (batch_requests_ready=0.001882s, total_serialization_initiated=0.040801s, others=0.000080s) I0421 07:44:41.234769 136206255867712 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048004s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.047988) I0421 07:44:41.235436 136206255867712 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/5. I0421 07:44:41.235770 136077230724864 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 08:04:41.235731 I0421 07:44:41.240278 136206255867712 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 07:44:41.240530 136076125652736 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 07:44:41.240670 136206255867712 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776757432.4972277, 'wait_for_prev_duration_secs': 47.989144802093506, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776757480.4884624, 'checkpointer_blocking_duration_secs': 0.7474527359008789, 'get_old_steps_start_time': 1776757481.2359369, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776757432.4949975, 'checkpoint_manager_blocking_duration_secs': 48.74562454223633} I0421 07:44:41.240782 136206255867712 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0421 07:44:41.241623 136206255867712 metric_logger.py:196] completed step: 4, seconds: 0.154, TFLOP/s/device: 88.408, Tokens/s/device: 13325.872, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0421 07:44:41.247460 136206255867712 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.306, Tokens/s/device: 13913.327, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0421 07:44:48.928462 136076134045440 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/5/items/array_metadatas/process_4 I0421 07:45:03.927679 136206255867712 metric_logger.py:196] completed step: 6, seconds: 49.041, TFLOP/s/device: 0.277, Tokens/s/device: 41.761, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0421 07:45:04.074917 136206255867712 metric_logger.py:196] completed step: 7, seconds: 22.534, TFLOP/s/device: 0.603, Tokens/s/device: 90.885, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0421 07:45:04.222367 136206255867712 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.300, Tokens/s/device: 13460.312, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0421 07:45:04.369174 136206255867712 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 07:45:04.369795 136206255867712 checkpointing.py:776] Waited 0.0006406307220458984 seconds for step 9 to finish before starting checkpointing. I0421 07:45:04.371914 136206255867712 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 07:45:25.271780 136077230724864 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 35.835 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.07983469963074 s) (per-host) I0421 07:45:25.271888 136077230724864 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.036004s. I0421 07:45:34.893063 136077230724864 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 53.657162s. I0421 07:45:34.893319 136076125652736 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 07:45:34.893453 136076125652736 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 07:45:34.893511 136076125652736 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 07:45:34.895386 136076125652736 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 07:45:34.895572 136206255867712 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0421 07:45:34.895722 136206255867712 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.523812 seconds. If this number is high, consider checkpointing less frequently. I0421 07:45:34.897273 136206255867712 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 9 I0421 07:45:34.899355 136206255867712 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/9. I0421 07:45:35.185899 136206255867712 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 07:45:35.185992 136206255867712 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 I0421 07:45:35.221105 136206255867712 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036050s I0421 07:45:35.221241 136206255867712 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 39.063 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0394892692565918 s) (per-host) I0421 07:45:35.221288 136206255867712 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039544s (batch_requests_ready=0.001724s, total_serialization_initiated=0.037761s, others=0.000060s) I0421 07:45:35.221362 136206255867712 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043753s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.043742) I0421 07:45:35.221971 136206255867712 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.32 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/9. I0421 07:45:35.222259 136083102734080 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 08:05:35.222212 I0421 07:45:35.223930 136206255867712 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 07:45:35.224192 136076125652736 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 07:45:35.224319 136206255867712 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776757504.3718827, 'wait_for_prev_duration_secs': 30.52381157875061, 'time_between_consecutive_saves_sec': 23.885807991027832, 'checkpointer_blocking_start_time': 1776757534.8973145, 'checkpointer_blocking_duration_secs': 0.3250892162322998, 'get_old_steps_start_time': 1776757535.222419, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1776757504.3700213, 'checkpoint_manager_blocking_duration_secs': 30.85426902770996} I0421 07:45:35.224472 136206255867712 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 07:45:35.224514 136206255867712 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 07:45:42.141307 136075594934016 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_false/checkpoints/9/items/array_metadatas/process_4 I0421 07:46:18.224081 136083102734080 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 36.699 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.042290925979614 s) (per-host) I0421 07:46:18.224202 136083102734080 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.001828s. I0421 07:46:25.933858 136083102734080 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 50.711468s. I0421 07:46:25.934137 136076125652736 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 07:46:25.934252 136076125652736 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 07:46:25.934295 136076125652736 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 07:46:25.935847 136076125652736 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 07:46:25.936034 136206255867712 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 07:46:25.936197 136206255867712 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 07:46:25.937180 136206255867712 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.252, Tokens/s/device: 13905.203, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 07:46:37 UTC 2026 EXIT_CODE=0