MaxView

‹ 09_pdb_lt_1Case: 10_shardy_false10_shardy_true ›

Metrics: Linen vs NNX  ·  main

MetricLinen  147168c3aNNX  1b5e4eb5aDiff (NNX − Linen)
Parameters1.104 billion
Final loss8.1790
TFLOP/s91.843
Tok/s13843.5
Avg s/step2.747
Memory %1.38
JAX0.9.20.9.2

Diff = NNX value − Linen value. Green = NNX improved. Red = NNX regressed.

Linen  ·  147168c3a  ·  main_20260417_235819  ·  full log
XPK Start: Sat Apr 18 00:50:13 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 00:50:38.095671: 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 00:50:38.275058 140674705372992 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 00:50:47.315420 140674705372992 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 00:50:47.317778 140674705372992 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-llruj-slice-job-0-0.mt-10-shardy-false-llruj:8482
I0418 00:50:48.520721 140674705372992 max_utils.py:284] Jax distributed system initialized!
I0418 00:50:54.590502 140674705372992 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 00:50:54.590604 140674705372992 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 00:50:54.590643 140674705372992 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 00:50:54.590677 140674705372992 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 00:50:55.271253 140674705372992 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 00:50:55.271529 140674705372992 checkpointing.py:677] Setting up checkpoint logger...
I0418 00:50:55.271583 140674705372992 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 00:50:55.271627 140674705372992 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 00:50:55.271973 140674705372992 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 0x7ff0aa350e90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 00:50:59.196719 140674705372992 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 00:50:59.197071 140674705372992 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 0x7fdc64290bf0>}, handler_registry=None
I0418 00:50:59.197314 140674705372992 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fdc64290bf0>` 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 00:50:59.197363 140674705372992 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fdc64331400>` 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 00:50:59.197397 140674705372992 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 0x7fdc64290bf0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fdc64290bf0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fdc64331400>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fdc64331400>}).
I0418 00:50:59.197715 140674705372992 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 00:50:59.197786 140674705372992 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7fdc4c3f0180> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 00:51:00.443703 140674705372992 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints
I0418 00:51:01.290555 140674705372992 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_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7fdc6432f2f0>
I0418 00:51:01.290728 140674705372992 checkpointing.py:301] Checkpoint manager created!
I0418 00:51:02.201372 140674705372992 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 00:51:02.201486 140674705372992 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 00:51:02.585350 140674705372992 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 00:51:02.585439 140674705372992 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 00:51:02.601544 140674705372992 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 00:51:02.601622 140674705372992 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 00:51:02.625161 140674705372992 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 00:51:02.625226 140674705372992 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:51:02.641246 140674705372992 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 00:51:02.641300 140674705372992 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:51:02.657377 140674705372992 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 00:51:02.657435 140674705372992 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:51:02.681753 140674705372992 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 00:51:02.681829 140674705372992 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:51:02.703243 140674705372992 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 00:51:02.703308 140674705372992 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 00:51:02.919278 140674705372992 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 00:51:02.919382 140674705372992 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 00:51:04.299088 140674705372992 maxtext_utils.py:1654]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0418 00:51:04.299210 140674705372992 maxtext_utils.py:1654]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0418 00:51:04.299263 140674705372992 maxtext_utils.py:1654]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0418 00:51:04.299319 140674705372992 maxtext_utils.py:1654]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0418 00:51:04.299371 140674705372992 maxtext_utils.py:1654]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0418 00:51:04.299409 140674705372992 maxtext_utils.py:1654]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0418 00:51:04.299461 140674705372992 maxtext_utils.py:1654]  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 00:51:04.299512 140674705372992 maxtext_utils.py:1654]  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 00:51:04.299551 140674705372992 maxtext_utils.py:1654]  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 00:51:04.299587 140674705372992 maxtext_utils.py:1654]  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 00:51:04.299633 140674705372992 maxtext_utils.py:1654]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0418 00:51:04.299680 140674705372992 maxtext_utils.py:1654]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0418 00:51:04.782456 140674705372992 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 00:51:04.782548 140674705372992 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 00:51:04.797656 140674705372992 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 00:51:04.797715 140674705372992 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 00:51:15.286245 140674705372992 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 00:51:15.293832 140674705372992 metric_logger.py:301] number parameters: 1.104 billion
I0418 00:51:26.900477 140674705372992 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 00:51:27.054573 140674705372992 checkpointing.py:776] Waited 0.15407323837280273 seconds for step 0 to finish before starting checkpointing.
I0418 00:51:27.058028 140674705372992 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:51:27.059562 140674705372992 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 00:51:27.061139 140674705372992 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0.
I0418 00:51:27.790476 140674705372992 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 00:51:27.806447 140674705372992 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 00:51:27.806512 140674705372992 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 00:51:28.081172 140674705372992 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275456s
I0418 00:51:28.081983 140674705372992 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.472 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28192591667175293 s) (per-host)
I0418 00:51:28.082061 140674705372992 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282021s (batch_requests_ready=0.002296s, total_serialization_initiated=0.278991s, others=0.000734s)
I0418 00:51:28.082143 140674705372992 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286900s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.286883)
I0418 00:51:28.082899 140674705372992 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0.
I0418 00:51:28.083202 140545707894528 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:11:28.083169
I0418 00:51:28.088963 140674705372992 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 00:51:28.089275 140544637245184 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 00:51:28.089431 140674705372992 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776473487.0580084, 'wait_for_prev_duration_secs': 6.365776062011719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776473487.0595999, 'checkpointer_blocking_duration_secs': 1.0237486362457275, 'get_old_steps_start_time': 1776473488.0833757, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776473487.0551724, 'checkpoint_manager_blocking_duration_secs': 1.034219741821289}
I0418 00:51:28.089538 140674705372992 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 00:51:28.089585 140674705372992 max_utils.py:750] 
Memstats: After params initialized:
I0418 00:51:28.089632 140674705372992 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 00:51:28.089664 140674705372992 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 00:51:28.089689 140674705372992 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 00:51:28.089712 140674705372992 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 00:51:28.426896 140674705372992 metric_logger.py:196] completed step: 0, seconds: 11.313, TFLOP/s/device: 1.201, Tokens/s/device: 181.026, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 00:51:28.431059 140674705372992 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/tensorboard/'
I0418 00:51:28.851443 140674705372992 metric_logger.py:196] completed step: 1, seconds: 1.525, TFLOP/s/device: 8.912, Tokens/s/device: 1343.272, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 00:51:28.912341 140545221457664 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0
I0418 00:51:28.997728 140674705372992 metric_logger.py:196] completed step: 2, seconds: 0.425, TFLOP/s/device: 31.949, Tokens/s/device: 4815.775, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129
I0418 00:51:29.145098 140674705372992 metric_logger.py:196] completed step: 3, seconds: 0.011, TFLOP/s/device: 1263.919, Tokens/s/device: 190511.628, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203
I0418 00:51:31.647563 140542495516416 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776473491242150369, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776473491242150369, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:51:31.974076 140545204672256 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/items
I0418 00:51:33.119795    2845 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 00:51:35.329474 140544654030592 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/items/array_metadatas/process_0
I0418 00:51:51.315048 140674705372992 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.467, Tokens/s/device: 13937.567, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614
I0418 00:51:51.323774 140674705372992 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.178, Tokens/s/device: 13894.071, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717
I0418 00:51:51.469904 140674705372992 metric_logger.py:196] completed step: 6, seconds: 22.167, TFLOP/s/device: 0.613, Tokens/s/device: 92.389, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759
I0418 00:51:51.617346 140674705372992 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2065.541, Tokens/s/device: 311340.833, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905
I0418 00:51:51.764462 140674705372992 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.228, Tokens/s/device: 13901.616, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406
I0418 00:51:51.911239 140674705372992 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 00:51:51.911893 140674705372992 checkpointing.py:776] Waited 0.0006759166717529297 seconds for step 9 to finish before starting checkpointing.
I0418 00:51:51.914305 140674705372992 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 00:52:09.564036 140544645637888 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.449216s (commit=35.268636s, array_metadata_write=1.180580s)
I0418 00:52:09.565213 140545707894528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.821 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.76512622833252 s) (per-host)
I0418 00:52:09.565337 140545707894528 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.482026s.
I0418 00:52:10.685520 140545707894528 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776473491242150369, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:52:11.451347 140542495516416 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776473491242150369, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:52:11.830504 140545707894528 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000238s.
I0418 00:52:12.062762 140545707894528 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 00:52:12.986526 140545707894528 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.174404s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/items
I0418 00:52:12.989295 140545707894528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0/items
I0418 00:52:13.544588 140545707894528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0
I0418 00:52:15.028885 140545707894528 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0`.
I0418 00:52:15.029767 140545707894528 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 47.97 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/0
I0418 00:52:15.031357 140545707894528 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.948043s.
I0418 00:52:15.031490 140544637245184 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 00:52:15.031573 140544637245184 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 00:52:15.031637 140544637245184 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 00:52:15.033639 140544637245184 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 00:52:15.033813 140674705372992 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 00:52:15.033970 140674705372992 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.119663 seconds. If this number is high, consider checkpointing less frequently.
I0418 00:52:15.035792 140674705372992 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 00:52:15.037905 140674705372992 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9.
I0418 00:52:16.170517 140674705372992 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 00:52:16.170610 140674705372992 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 00:52:16.196264 140674705372992 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026469s
I0418 00:52:16.197043 140674705372992 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.742 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031011343002319336 s) (per-host)
I0418 00:52:16.197116 140674705372992 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031101s (batch_requests_ready=0.001936s, total_serialization_initiated=0.028428s, others=0.000737s)
I0418 00:52:16.197188 140674705372992 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035204s (all_items=0.000021s, per_item={'items': '0.00002122'}, temp_paths=0.035183)
I0418 00:52:16.197937 140674705372992 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.16 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9.
I0418 00:52:16.198203 140544637245184 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:12:16.198181
I0418 00:52:16.213145 140674705372992 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 00:52:16.213420 140543610844928 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 00:52:16.213566 140674705372992 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776473511.9142728, 'wait_for_prev_duration_secs': 23.11966347694397, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776473535.035851, 'checkpointer_blocking_duration_secs': 1.1624481678009033, 'get_old_steps_start_time': 1776473536.1983247, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776473511.912126, 'checkpoint_manager_blocking_duration_secs': 24.301403522491455}
I0418 00:52:16.213677 140674705372992 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 00:52:16.213722 140674705372992 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 00:52:16.843137 140545707894528 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9
I0418 00:52:18.602826 140544147715840 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9/items
I0418 00:52:22.655417 140544109434624 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9/items/array_metadatas/process_0
I0418 00:53:02.555215 140544101041920 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.400224s (commit=40.752859s, array_metadata_write=1.647366s)
I0418 00:53:02.557859 140544637245184 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.049 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.39177846908569 s) (per-host)
I0418 00:53:02.557984 140544637245184 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.359704s.
I0418 00:53:04.805007 140544637245184 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 00:53:04.865093 140544637245184 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000212s.
I0418 00:53:05.460937 140544637245184 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.605604s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9/items
I0418 00:53:05.463496 140544637245184 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9/items
I0418 00:53:06.005228 140544637245184 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9
I0418 00:53:08.535603 140544637245184 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9`.
I0418 00:53:08.536361 140544637245184 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.50 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_10_shardy_false/checkpoints/9
I0418 00:53:08.537631 140544637245184 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.339350s.
I0418 00:53:08.537774 140543610844928 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 00:53:08.537829 140543610844928 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 00:53:08.537919 140543610844928 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 00:53:08.542547 140543610844928 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 00:53:08.542705 140674705372992 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 00:53:08.542862 140674705372992 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:53:08.543827 140674705372992 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.843, Tokens/s/device: 13843.544, 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 00:53:19 UTC 2026
EXIT_CODE=0
NNX  ·  1b5e4eb5a  ·  main_20260417_235819  ·  full log
XPK Start: Sat Apr 18 02:24:43 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 02:25:07.626691: 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 02:25:08.164444 135287218988864 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 02:25:17.206106 135287218988864 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 02:25:17.210659 135287218988864 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-173ca-slice-job-0-0.mt-10-shardy-false-173ca:8482
I0418 02:25:18.620008 135287218988864 max_utils.py:284] Jax distributed system initialized!
I0418 02:25:23.613399 135287218988864 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 02:25:23.613501 135287218988864 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 02:25:23.613539 135287218988864 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 02:25:23.613571 135287218988864 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 744, in <module>
    app.run(main)
  File "/usr/local/lib/python3.12/site-packages/absl/app.py", line 367, in run
    _run_main(main, args)
  File "/usr/local/lib/python3.12/site-packages/absl/app.py", line 312, in _run_main
    sys.exit(main(argv))
             ^^^^^^^^^^
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 740, in main
    train_func()
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 730, in train_func
    run(config, recorder, diagnostic_config)
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 709, in run
    train_loop(config, recorder)
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 536, in train_loop
    ) = train_utils.setup_train_loop(config, recorder)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/deps/src/maxtext/utils/train_utils.py", line 217, in setup_train_loop
    raise NotImplementedError("Pure NNX support has not been implemented yet.")
NotImplementedError: Pure NNX support has not been implemented yet.
XPK End: Sat Apr 18 02:25:29 UTC 2026
EXIT_CODE=1