2026-07-27 12:03:34.947 2012bc4f83ae:0 olmo.util:275 INFO Set up torchrun environment 2026-07-27 12:03:35.197 2012bc4f83ae:0 train:61 INFO Configuration: 2026-07-27 12:03:35.198 2012bc4f83ae:0 train:62 INFO TrainConfig(run_name='firewarning-molmopoint-strong-curated-ground-v1', model=MolmoPointConfig(model_name='token_indexing_video_molmo2', data_formatter=MolmoPointDataFormatter(prompt_templates='none', message_format='qwen3', system_prompt='style_and_length_v2', always_start_with_space=False, default_inference_len=65, select_answer='best', debug=False, eval_system_prompt_mapping=None, p_choice_content_in_mc=1.0, points_decimal_places=1, use_seperate_non_pointing_qa_style=False, timestamp_mode='50-percent-seconds', output_timestamp_mode='seconds', seconds_decimal_places=1, p_multi_point_all_image=0.0, use_seperate_count_without_pointing_style=False, sample_random_initial_point=True, _point_start_token=False, include_point_number='no_space_id_last', _end_with_patch=True, _location_token=True), llm=LlmConfig(d_model=4096, n_heads=32, n_kv_heads=8, head_dim=128, qkv_bias=False, clip_qkv=None, n_layers=36, mlp_ratio=4, mlp_hidden_size=24576, activation_type='swiglu', block_type='sequential', rope=True, rope_full_precision=True, rope_theta=1000000.0, rope_type='default', rope_factor=None, rope_high_freq_factor=None, rope_low_freq_factor=None, rope_original_max_position_embeddings=None, rope_attention_factor=None, rope_beta_fast=None, rope_beta_slow=None, rope_mscale=None, rope_mscale_all_dim=None, rope_truncate=None, attention_type='sdpa', full_attention_layers=None, sliding_attention_rope_scaling=False, float32_attention=True, attention_dropout=0.0, attention_layer_norm=True, attention_layer_norm_type='qwen3', residual_dropout=0.1, response_residual_dropout=0.0, layer_norm_type='rms', layer_norm_with_affine=True, layer_norm_eps=1e-06, attention_layer_norm_with_affine=True, max_sequence_length=4096, max_position_embeddings=None, include_bias=False, bias_for_layer_norm=None, norm_after=False, moe_num_experts=8, moe_top_k=2, moe_mlp_impl='sparse', moe_log_expert_assignment=False, moe_shared_expert=False, moe_lbl_in_fp32=False, moe_interleave=False, moe_loss_weight=0.1, moe_zloss_weight=None, moe_dropless=True, moe_capacity_factor=1.25, embedding_dropout=0.0, scale_logits=False, vocab_size=151936, additional_vocab_size=128, weight_tying=False, embedding_size=151936, use_position_ids=True, tokenizer=TokenizerConfig(identifier='Qwen/Qwen3-8B', tokenizer_dir=None), can_predict_extra_tokens=True, init_path='/weka/oe-training-default/mm-olmo/pretrained_llms/qwen3-8b.pt', init_incremental=None, new_embedding_init_range=0.02, initializer_range=0.02, normalize_input_embeds=False, activation_checkpoint='whole_layer', compile='blocks', fix_pad_tokenizer=False, init_std=0.02, init_fn='normal', init_cutoff_factor=None), vit=VitConfig(image_model_type='siglip', image_default_input_size=(378, 378), image_patch_size=14, image_pos_patch_size=14, image_emb_dim=1152, image_num_heads=16, image_num_key_value_heads=16, image_num_layers=27, image_head_dim=72, image_mlp_dim=4304, image_mlp_activations='gelu_pytorch_tanh', image_dropout_rate=0.0, image_num_pos=729, image_norm_eps=1e-06, attention_dropout=0.0, residual_dropout=0.0, initializer_range=0.02, float32_attention=True, attention_type='sdpa', sdpa_backend='all', activation_checkpointing=True, init_path='/weka/oe-training-default/mm-olmo/pretrained_image_encoders/siglip2-so400m-14-384.pt', resize_mode='siglip', pad_value=0.0, normalize='siglip'), connector=ConnectorConfig(pooling_attention_mask=True, image_projector='mlp', pooling_out_layer=False, vit_layers=(-3, -9), skip_unused_layers=True, positional_embeddings=None, connector_activation_checkpointing=True, compile_connector=None, normalize_on_gpu=True), mm_preprocessor=MolmoPointPreprocessorConfig(max_answer_len=None, last_message_loss_only=False, max_text_tokens=None, loss_token_weighting=None, video=VideoPreprocessorConfig(max_frames=16, frame_sample_mode='uniform_last_frame', candidate_sampling_fps=[0.25, 0.5, 1.0, 2.0, 4.0, 6.0, 8.0, 16.0], cache_videos=True, loading_method='torchcodec_exact', max_fps=[2.0], time_sampling=True, time_mode='per-frame-compact', subtitle_mode='frame_1', pooling_w=3, pooling_h=3, use_frame_special_tokens=True, per_frame_special_token=True, max_subtitle_tokens=None), image=MultiCropConfig(crop_mode='overlap-and-resize-c2', use_col_tokens=True, max_crops=8, high_res_max_crops=8, p_high_res=0.0, pooling_w=2, pooling_h=2, overlap_margins=[4, 4], max_images=1, max_multi_image_crops=1, multi_image_pooling_w=2, multi_image_pooling_h=2, use_single_crop_col_tokens=False, use_single_crop_start_token=True), remove_repeats=None), bi_directional_attn='image_tokens', shared_low_high_embedding=True, patch_location='3x3', no_more_points_class=True, patch_embed_dim=512, patch_embedding_kind='image_feature0', embed_selected_vit_patch='linear', embed_location=False, layer_norm_x=True, norm_x=True, norm_logits=True, mask_patches='always', sort_points=True, mask_subpatches='inference', mask_repeats='inference', token_prediction_rotary='one_d', token_prediction_rotary_theta=50000.0, token_prediction_rotary_dims=None, debug=None), parallelism=ParallelismConfig(data_parallel_replicate_degree=1, enable_compiled_autograd=False, data_parallel_shard_degree=-1, fsdp_reshard_after_forward='default', context_parallel_config=TransformerContextParallelConfig(degree=1, attention_type='ulysses', load_balancer='ulysses', head_stride=1), tensor_parallel_config=TransformerTensorParallelConfig(degree=1, enable_async=False), data_parallel_config=TransformerDataParallelConfig(name='fsdp', param_dtype=None, reduce_dtype='float32', num_replicas=None, shard_degree=None, wrapping_strategy='full', prefetch_factor=0), context_parallel_rotate_method='allgather'), seed=6198, epoch=None, dry_run=False, ft_llm=False, ft_vit=False, ft_connector=False, ft_embedding='lm_head', optimizer=OptimizerConfig(name='adamw', learning_rate=0.0001, weight_decay=0.0, betas=(0.9, 0.95), eps=1e-05, connector_learning_rate=5e-06, vit_learning_rate=5e-06, llm_learning_rate=1e-05, frame_selector_learning_rate=5e-05, connector_weight_decay=0.0, vit_weight_decay=0.0, llm_weight_decay=0.0, frame_selector_weight_decay=0.0, connector_betas=(0.9, 0.95), vit_betas=(0.9, 0.95), llm_betas=(0.9, 0.95), frame_selector_betas=[0.9, 0.95], connector_eps=1e-06, vit_eps=1e-06, llm_eps=1e-06, frame_selector_eps=1e-06, metrics_log_interval=-1), scheduler=SchedulerConfig(name='multimodal', units='steps', t_warmup=100, t_max=None, alpha_f=0.1, connector_t_warmup=200, vit_t_warmup=200, llm_t_warmup=200, frame_selector_t_warmup=50, temporal_token_scorer_t_warmup=200, grad_clip_warmup_steps=None, grad_clip_warmup_factor=None, warmup_min_lr=0.0), data=DataLoaderConfig(dataset=None, mixture=None, root_size_mixture=None, kwargs_mixture=[KwargsMixture(rate=1.0, datasets=[WeightedDataset(dataset_name='firewarning_pointing', sampling_rate=None, root_size_factor=None, message_weight=MessageWeight(weight=0.35, root_subsegments=False, root_length=False), override_p_high_res=0.3)], name='firewarning_pointing')], split='train', seed=50189, pad='to_max', sequence_length=4096, max_text_seq_len=None, shuffle=True, start_index=0, packing=PackingConfig(buffer_size=48, mode='dynamic_solver', text_weight=1.0, image_weight=30.0, shortcut_max_len_images=False, track_packing_state=False, cp_world_size=1), enable_variable_sized_token_pooling=True, num_workers=8, drop_last=True, pin_memory=True, prefetch_factor=4, persistent_workers=False, timeout=0), restore_dataloader=False, fast_forward_batches=None, evaluators=[], eval_interval=-1, inf_evaluators=[], inf_eval_interval=-1, eval_on_last_step=True, eval_on_load=False, eval_on=[], save_folder='/runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1', checkpointer_config=CheckpointerConfig(save_thread_count=1, load_thread_count=None, pre_download=False, work_dir=None, throttle_uploads=False), canceled_check_interval=50, save_interval=100, save_at=None, save_final_optim=False, save_num_checkpoints_to_keep=4, save_final_unsharded_checkpoint=False, save_interval_ephemeral=None, save_overwrite=True, load_path=None, reset_optimizer_state=False, reset_trainer_state=False, initial_model_checkpoint='/workspace/firewarning-runtime/models/molmopoint-native/pretrain/MolmoPoint-8B-Pretrain/step23000', allow_resume=True, max_duration=600, global_train_batch_size=4, device_train_microbatch_size=1, max_grad_norm=1.0, batch_divisor='global_batch', max_grad_norm_ratio=None, precision='amp_bf16', wandb=None, beaker_log_interval=50, speed_monitor=SpeedMonitorConfig(window_size=20, gpu_flops_available=None), console_log_interval=1, gen1_gc_interval=1, compile=None, activation_checkpointing=True, fsdp=FSDPConfig(fsdp2=True, precision='pure', use_orig_params=True, wrapping_strategy=None, sharding_strategy=, hybrid_sharding_num_model_replicas=None), softmax_auxiliary_loss=True, softmax_auxiliary_loss_scale=0.0001, response_logits_only=True, saliency_score_loss_wt=None, frame_score_loss_wt=None, frame_score_loss_type='mse', frame_score_loss_target=0.7, time_limit=None, extra_steps_after_cancel=0, python_profiling=False, torch_profiling=False, stop_at=600, stop_after=None, fused_loss=False, compile_loss=False, runtime_data=None, inf_eval_config=InfEvalConfig(distributed=True, offload_model=False, offload_optim=False), save_inloop_predictions=True) 2026-07-27 12:03:35.209 2012bc4f83ae:0 train:89 INFO Not resuming since no latest checkpoint found 2026-07-27 12:03:37.402 2012bc4f83ae:0 root:139 INFO Padding tokenizer with 267 tokens 2026-07-27 12:03:38.438 2012bc4f83ae:0 train:116 INFO Freezing connector 2026-07-27 12:03:38.439 2012bc4f83ae:0 train:120 INFO Freezing vision backbone 2026-07-27 12:03:38.440 2012bc4f83ae:0 train:124 INFO Freezing LLM 2026-07-27 12:03:38.475 2012bc4f83ae:0 train:191 INFO Wrapping model with FSDP2... 2026-07-27 12:03:40.535 2012bc4f83ae:0 train:222 INFO Total number of parameters: 8,677,855,065 2026-07-27 12:03:40.538 2012bc4f83ae:0 train:223 INFO Number of non-embedding parameters: 8,055,000,921 2026-07-27 12:03:40.540 2012bc4f83ae:0 train:226 INFO Peak GPU Memory (MB) after FSDP: 34732 2026-07-27 12:03:40.540 2012bc4f83ae:0 train:227 INFO Model: 2026-07-27 12:03:40.540 2012bc4f83ae:0 train:228 INFO FSDPMolmoPoint( (transformer): Llm( (blocks): ModuleList( (0-35): 36 x FSDPCheckpointWrapper( (_checkpoint_wrapped_module): OLMoSequentialBlock( (dropout): Dropout(p=0.1, inplace=False) (k_norm): RMSLayerNorm() (q_norm): RMSLayerNorm() (act): SwiGLU() (attn_out): Linear(in_features=4096, out_features=4096, bias=False) (ff_out): Linear(in_features=12288, out_features=4096, bias=False) (rotary_emb): RotaryEmbedding() (attn_norm): RMSLayerNorm() (ff_norm): RMSLayerNorm() (att_proj): Linear(in_features=4096, out_features=6144, bias=False) (ff_proj): Linear(in_features=4096, out_features=24576, bias=False) ) ) ) (ln_f): FSDPRMSLayerNorm() (emb_drop): Dropout(p=0.0, inplace=False) (wte): FSDPEmbedding() (ff_out): FSDPProjectWithExtra() ) (connector): FSDPMolmoPointConnector( (image_projector): FSDPCheckpointWrapper( (_checkpoint_wrapped_module): ImageProjectorMLP( (w1): Linear(in_features=1152, out_features=12288, bias=False) (w2): Linear(in_features=12288, out_features=4096, bias=False) (w3): Linear(in_features=1152, out_features=12288, bias=False) (act): LlamaSwiGLU() (dropout): Dropout(p=0.0, inplace=False) ) ) (image_pooling_2d): FSDPCheckpointWrapper( (_checkpoint_wrapped_module): ViTMultiHeadDotProductAttention( (wq): Linear(in_features=2304, out_features=1152, bias=True) (wk): Linear(in_features=2304, out_features=1152, bias=True) (wv): Linear(in_features=2304, out_features=1152, bias=True) (residual_dropout): Dropout(p=0.0, inplace=False) ) ) ) (vit): FSDPSiglipVisionTransformer( (patch_embedding): Linear(in_features=588, out_features=1152, bias=True) (transformer): BlockCollection( (resblocks): ModuleList( (0-24): 25 x FSDPCheckpointWrapper( (_checkpoint_wrapped_module): ResidualAttentionBlock( (attention): ViTMultiHeadDotProductAttention( (wq): Linear(in_features=1152, out_features=1152, bias=True) (wk): Linear(in_features=1152, out_features=1152, bias=True) (wv): Linear(in_features=1152, out_features=1152, bias=True) (wo): Linear(in_features=1152, out_features=1152, bias=True) (residual_dropout): Dropout(p=0.0, inplace=False) ) (feed_forward): ViTMLP( (w1): Linear(in_features=1152, out_features=4304, bias=True) (act): GELUTanh() (w2): Linear(in_features=4304, out_features=1152, bias=True) ) (attention_norm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True) (ffn_norm): LayerNorm((1152,), eps=1e-06, elementwise_affine=True) ) ) ) ) ) (build_vit_embedding): FSDPLinear(in_features=2304, out_features=4096, bias=True) (x_norm): FSDPRMSLayerNorm() (patch_q): FSDPLinear(in_features=4096, out_features=512, bias=True) (patch_k): FSDPLinear(in_features=4096, out_features=512, bias=True) (subpatch_q): FSDPLinear(in_features=4096, out_features=512, bias=True) (subpatch_k): FSDPLinear(in_features=2304, out_features=512, bias=True) (add_no_point_class_embed): FSDPPadWithLearnedVector() (subpatch_loc_k): FSDPLinear(in_features=4096, out_features=9, bias=True) (patch_rotary): FlatRotaryEmbedding() ) 2026-07-27 12:03:40.559 2012bc4f83ae:0 olmo.train.optim:179 INFO Constructing optimizer with 1 param groups 2026-07-27 12:03:40.561 2012bc4f83ae:0 olmo.data.data_loader:246 INFO Loading train dataset firewarning_pointing/train 2026-07-27 12:03:41.030 2012bc4f83ae:0 olmo.data.data_loader:302 INFO Sampling rates: 2026-07-27 12:03:41.030 2012bc4f83ae:0 olmo.data.data_loader:305 INFO firewarning_pointing: 100.00 (size=336) 2026-07-27 12:03:41.765 2012bc4f83ae:0 olmo.data.data_loader:340 INFO Packing with text=4096, images=16 2026-07-27 12:03:41.765 2012bc4f83ae:0 olmo.data.data_loader:346 INFO Building train dataset with output shapes: tokens=[4096], token_pooling=[1, 1], images=[16, 729, 588], point_target_ids=[1, 3], image_pos_ids=[1, 1] 2026-07-27 12:03:41.855 2012bc4f83ae:0 train:375 INFO Loading model from /workspace/firewarning-runtime/models/molmopoint-native/pretrain/MolmoPoint-8B-Pretrain/step23000 2026-07-27 12:03:41.964 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_projector.w1.weight' to 'connectors.0.image_projector.w1.weight' in checkpoint 2026-07-27 12:03:41.965 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_projector.w2.weight' to 'connectors.0.image_projector.w2.weight' in checkpoint 2026-07-27 12:03:41.965 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_projector.w3.weight' to 'connectors.0.image_projector.w3.weight' in checkpoint 2026-07-27 12:03:41.965 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_pooling_2d.wq.weight' to 'connectors.0.image_pooling_2d.wq.weight' in checkpoint 2026-07-27 12:03:41.966 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_pooling_2d.wq.bias' to 'connectors.0.image_pooling_2d.wq.bias' in checkpoint 2026-07-27 12:03:41.966 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_pooling_2d.wk.weight' to 'connectors.0.image_pooling_2d.wk.weight' in checkpoint 2026-07-27 12:03:41.967 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_pooling_2d.wk.bias' to 'connectors.0.image_pooling_2d.wk.bias' in checkpoint 2026-07-27 12:03:41.967 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_pooling_2d.wv.weight' to 'connectors.0.image_pooling_2d.wv.weight' in checkpoint 2026-07-27 12:03:41.968 2012bc4f83ae:0 olmo.train.distributed_checkpointing:262 INFO Mapping current param 'connector.image_pooling_2d.wv.bias' to 'connectors.0.image_pooling_2d.wv.bias' in checkpoint 2026-07-27 12:04:13.104 2012bc4f83ae:0 train:386 INFO Checkpoint successfully loaded in 31.2 seconds 2026-07-27 12:04:13.107 2012bc4f83ae:0 train:396 INFO Initializing FireWarning weights from /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-ground-v1/step100 2026-07-27 12:04:14.668 2012bc4f83ae:0 train:410 INFO Starting training... 2026-07-27 12:04:14.671 2012bc4f83ae:0 olmo.train.trainer:1050 INFO Pre-train system metrics System/Peak GPU Memory (MB)=39,088 2026-07-27 12:05:47.006 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=1/600, eta=4 hours, 41 minutes] train/CrossEntropyLoss=0.0215 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=2.248 train/patch_acc=0.5000 train/subpatch_loss=1.480 train/subpatch_acc=0.2500 train/location_loss=2.847 train/location_acc=0.0625 train/token_losses=35.29 throughput/total_tokens=16,384 System/Peak GPU Memory (MB)=41,640 2026-07-27 12:06:11.750 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=2/600, eta=4 hours, 24 minutes] train/CrossEntropyLoss=0.0225 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.343 train/patch_acc=0.5312 train/subpatch_loss=1.383 train/subpatch_acc=0.3750 train/location_loss=2.303 train/location_acc=0.1875 train/token_losses=33.49 throughput/total_tokens=32,768 throughput/device/tokens_per_second=662.4 throughput/device/batches_per_second=0.0404 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:06:36.172 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=3/600, eta=4 hours, 17 minutes] train/CrossEntropyLoss=0.0157 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=2.298 train/patch_acc=0.5938 train/subpatch_loss=1.616 train/subpatch_acc=0.3750 train/location_loss=2.326 train/location_acc=0.1250 train/token_losses=34.16 throughput/total_tokens=49,152 throughput/device/tokens_per_second=666.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:07:00.659 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=4/600, eta=4 hours, 13 minutes] train/CrossEntropyLoss=0.0474 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.908 train/patch_acc=0.5625 train/subpatch_loss=1.657 train/subpatch_acc=0.3125 train/location_loss=2.570 train/location_acc=0.1250 train/token_losses=32.17 throughput/total_tokens=65,536 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:07:25.167 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=5/600, eta=4 hours, 11 minutes] train/CrossEntropyLoss=0.0127 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=2.027 train/patch_acc=0.5938 train/subpatch_loss=1.522 train/subpatch_acc=0.3125 train/location_loss=2.494 train/location_acc=0.0 train/token_losses=32.28 throughput/total_tokens=81,920 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:07:49.674 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=6/600, eta=4 hours, 9 minutes] train/CrossEntropyLoss=0.0313 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=2.247 train/patch_acc=0.5000 train/subpatch_loss=1.226 train/subpatch_acc=0.5000 train/location_loss=2.328 train/location_acc=0.1250 train/token_losses=32.19 throughput/total_tokens=98,304 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 12:08:14.211 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=7/600, eta=4 hours, 8 minutes] train/CrossEntropyLoss=0.0498 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=2.319 train/patch_acc=0.5625 train/subpatch_loss=1.475 train/subpatch_acc=0.3125 train/location_loss=2.141 train/location_acc=0.1250 train/token_losses=33.02 throughput/total_tokens=114,688 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:08:38.747 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=8/600, eta=4 hours, 7 minutes] train/CrossEntropyLoss=0.0011 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=2.094 train/patch_acc=0.5938 train/subpatch_loss=1.615 train/subpatch_acc=0.1250 train/location_loss=2.347 train/location_acc=0.1875 train/token_losses=32.60 throughput/total_tokens=131,072 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:09:03.306 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=9/600, eta=4 hours, 6 minutes] train/CrossEntropyLoss=0.1649 train/ZLoss=0.0006 train/Accuracy=0.9604 train/patch_loss=2.344 train/patch_acc=0.4688 train/subpatch_loss=1.110 train/subpatch_acc=0.5625 train/location_loss=2.286 train/location_acc=0.0625 train/token_losses=32.34 throughput/total_tokens=147,456 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:09:27.837 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=10/600, eta=4 hours, 5 minutes] train/CrossEntropyLoss=0.0480 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.415 train/patch_acc=0.5000 train/subpatch_loss=1.188 train/subpatch_acc=0.5000 train/location_loss=2.247 train/location_acc=0.1875 train/token_losses=33.06 throughput/total_tokens=163,840 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:09:52.403 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=11/600, eta=4 hours, 5 minutes] train/CrossEntropyLoss=0.0257 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=2.021 train/patch_acc=0.5000 train/subpatch_loss=1.619 train/subpatch_acc=0.3125 train/location_loss=2.466 train/location_acc=0.0625 train/token_losses=32.51 throughput/total_tokens=180,224 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:10:16.962 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=12/600, eta=4 hours, 4 minutes] train/CrossEntropyLoss=0.0357 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.327 train/patch_acc=0.5000 train/subpatch_loss=1.518 train/subpatch_acc=0.3750 train/location_loss=2.164 train/location_acc=0.1250 train/token_losses=33.34 throughput/total_tokens=196,608 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:10:41.509 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=13/600, eta=4 hours, 3 minutes] train/CrossEntropyLoss=0.0322 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=2.135 train/patch_acc=0.5625 train/subpatch_loss=1.289 train/subpatch_acc=0.5625 train/location_loss=2.452 train/location_acc=0.0 train/token_losses=32.04 throughput/total_tokens=212,992 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:11:06.059 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=14/600, eta=4 hours, 3 minutes] train/CrossEntropyLoss=0.2041 train/ZLoss=0.0006 train/Accuracy=0.9543 train/patch_loss=1.802 train/patch_acc=0.6250 train/subpatch_loss=1.516 train/subpatch_acc=0.2500 train/location_loss=2.369 train/location_acc=0.0625 train/token_losses=29.95 throughput/total_tokens=229,376 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:11:30.699 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=15/600, eta=4 hours, 2 minutes] train/CrossEntropyLoss=0.0342 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=2.023 train/patch_acc=0.6250 train/subpatch_loss=1.285 train/subpatch_acc=0.3125 train/location_loss=1.983 train/location_acc=0.1875 train/token_losses=29.25 throughput/total_tokens=245,760 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:11:55.233 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=16/600, eta=4 hours, 2 minutes] train/CrossEntropyLoss=0.3306 train/ZLoss=0.0009 train/Accuracy=0.9177 train/patch_loss=1.971 train/patch_acc=0.5625 train/subpatch_loss=1.035 train/subpatch_acc=0.5000 train/location_loss=2.280 train/location_acc=0.0625 train/token_losses=29.03 throughput/total_tokens=262,144 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:12:19.755 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=17/600, eta=4 hours, 1 minute] train/CrossEntropyLoss=0.0355 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.851 train/patch_acc=0.5938 train/subpatch_loss=1.005 train/subpatch_acc=0.6875 train/location_loss=2.202 train/location_acc=0.0625 train/token_losses=27.64 throughput/total_tokens=278,528 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:12:44.296 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=18/600, eta=4 hours, 1 minute] train/CrossEntropyLoss=0.0216 train/ZLoss=0.0003 train/Accuracy=0.9878 train/patch_loss=2.104 train/patch_acc=0.5938 train/subpatch_loss=1.494 train/subpatch_acc=0.3750 train/location_loss=2.377 train/location_acc=0.0625 train/token_losses=32.32 throughput/total_tokens=294,912 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:13:08.853 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=19/600, eta=4 hours] train/CrossEntropyLoss=0.0149 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.966 train/patch_acc=0.6250 train/subpatch_loss=1.059 train/subpatch_acc=0.6875 train/location_loss=2.093 train/location_acc=0.2500 train/token_losses=28.33 throughput/total_tokens=311,296 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:13:33.392 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=20/600, eta=4 hours] train/CrossEntropyLoss=0.0328 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.561 train/patch_acc=0.6250 train/subpatch_loss=1.103 train/subpatch_acc=0.5000 train/location_loss=2.276 train/location_acc=0.1250 train/token_losses=26.01 throughput/total_tokens=327,680 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:13:57.922 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=21/600, eta=3 hours, 59 minutes] train/CrossEntropyLoss=0.0466 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=1.688 train/patch_acc=0.5938 train/subpatch_loss=0.8705 train/subpatch_acc=0.7500 train/location_loss=2.314 train/location_acc=0.0625 train/token_losses=26.24 throughput/total_tokens=344,064 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:14:22.459 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=22/600, eta=3 hours, 59 minutes] train/CrossEntropyLoss=0.0072 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=2.300 train/patch_acc=0.5312 train/subpatch_loss=0.8311 train/subpatch_acc=0.6250 train/location_loss=2.164 train/location_acc=0.0625 train/token_losses=30.38 throughput/total_tokens=360,448 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:14:46.990 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=23/600, eta=3 hours, 58 minutes] train/CrossEntropyLoss=0.0063 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=2.310 train/patch_acc=0.5938 train/subpatch_loss=0.9790 train/subpatch_acc=0.6875 train/location_loss=2.114 train/location_acc=0.1875 train/token_losses=30.85 throughput/total_tokens=376,832 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:15:11.543 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=24/600, eta=3 hours, 58 minutes] train/CrossEntropyLoss=0.0539 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=2.955 train/patch_acc=0.5625 train/subpatch_loss=1.114 train/subpatch_acc=0.4375 train/location_loss=2.180 train/location_acc=0.1250 train/token_losses=36.81 throughput/total_tokens=393,216 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:15:36.097 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=25/600, eta=3 hours, 57 minutes] train/CrossEntropyLoss=0.0563 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.987 train/patch_acc=0.6250 train/subpatch_loss=0.9503 train/subpatch_acc=0.4375 train/location_loss=2.263 train/location_acc=0.0 train/token_losses=28.75 throughput/total_tokens=409,600 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:16:00.643 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=26/600, eta=3 hours, 57 minutes] train/CrossEntropyLoss=0.0113 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=2.543 train/patch_acc=0.5312 train/subpatch_loss=1.054 train/subpatch_acc=0.5625 train/location_loss=2.065 train/location_acc=0.2500 train/token_losses=32.82 throughput/total_tokens=425,984 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:16:25.180 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=27/600, eta=3 hours, 56 minutes] train/CrossEntropyLoss=0.0364 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.027 train/patch_acc=0.6250 train/subpatch_loss=0.8206 train/subpatch_acc=0.5625 train/location_loss=2.169 train/location_acc=0.1875 train/token_losses=28.18 throughput/total_tokens=442,368 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:16:49.751 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=28/600, eta=3 hours, 56 minutes] train/CrossEntropyLoss=0.0257 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.544 train/patch_acc=0.6250 train/subpatch_loss=0.8610 train/subpatch_acc=0.6250 train/location_loss=2.321 train/location_acc=0.0625 train/token_losses=25.08 throughput/total_tokens=458,752 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:17:14.290 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=29/600, eta=3 hours, 55 minutes] train/CrossEntropyLoss=0.0042 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.736 train/patch_acc=0.5938 train/subpatch_loss=1.001 train/subpatch_acc=0.5625 train/location_loss=2.181 train/location_acc=0.1250 train/token_losses=26.62 throughput/total_tokens=475,136 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:17:38.813 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=30/600, eta=3 hours, 55 minutes] train/CrossEntropyLoss=0.0162 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.014 train/patch_acc=0.5625 train/subpatch_loss=1.006 train/subpatch_acc=0.5000 train/location_loss=2.043 train/location_acc=0.2500 train/token_losses=28.31 throughput/total_tokens=491,520 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:18:03.341 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=31/600, eta=3 hours, 54 minutes] train/CrossEntropyLoss=0.0447 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=2.232 train/patch_acc=0.5625 train/subpatch_loss=0.9663 train/subpatch_acc=0.5000 train/location_loss=2.093 train/location_acc=0.1875 train/token_losses=30.10 throughput/total_tokens=507,904 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:18:27.857 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=32/600, eta=3 hours, 54 minutes] train/CrossEntropyLoss=0.0182 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.706 train/patch_acc=0.5938 train/subpatch_loss=0.9503 train/subpatch_acc=0.5000 train/location_loss=1.993 train/location_acc=0.3125 train/token_losses=33.42 throughput/total_tokens=524,288 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:18:52.368 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=33/600, eta=3 hours, 53 minutes] train/CrossEntropyLoss=0.0048 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=2.283 train/patch_acc=0.5625 train/subpatch_loss=0.9671 train/subpatch_acc=0.6250 train/location_loss=2.240 train/location_acc=0.0625 train/token_losses=31.09 throughput/total_tokens=540,672 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:19:16.889 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=34/600, eta=3 hours, 53 minutes] train/CrossEntropyLoss=0.0064 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=1.899 train/patch_acc=0.5938 train/subpatch_loss=0.8040 train/subpatch_acc=0.6250 train/location_loss=2.054 train/location_acc=0.1250 train/token_losses=26.62 throughput/total_tokens=557,056 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:19:41.418 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=35/600, eta=3 hours, 53 minutes] train/CrossEntropyLoss=0.0267 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.239 train/patch_acc=0.5625 train/subpatch_loss=0.8463 train/subpatch_acc=0.6875 train/location_loss=2.120 train/location_acc=0.0625 train/token_losses=29.78 throughput/total_tokens=573,440 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:20:05.949 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=36/600, eta=3 hours, 52 minutes] train/CrossEntropyLoss=0.0033 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=1.534 train/patch_acc=0.7500 train/subpatch_loss=0.7947 train/subpatch_acc=0.6875 train/location_loss=2.476 train/location_acc=0.0 train/token_losses=25.36 throughput/total_tokens=589,824 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:20:30.466 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=37/600, eta=3 hours, 52 minutes] train/CrossEntropyLoss=0.0602 train/ZLoss=0.0005 train/Accuracy=0.9878 train/patch_loss=1.690 train/patch_acc=0.5625 train/subpatch_loss=0.7795 train/subpatch_acc=0.7500 train/location_loss=2.191 train/location_acc=0.0625 train/token_losses=25.40 throughput/total_tokens=606,208 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:20:54.990 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=38/600, eta=3 hours, 51 minutes] train/CrossEntropyLoss=0.0543 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.901 train/patch_acc=0.6250 train/subpatch_loss=0.8586 train/subpatch_acc=0.6875 train/location_loss=2.238 train/location_acc=0.0 train/token_losses=27.59 throughput/total_tokens=622,592 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:21:19.513 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=39/600, eta=3 hours, 51 minutes] train/CrossEntropyLoss=0.0249 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.760 train/patch_acc=0.6562 train/subpatch_loss=0.7173 train/subpatch_acc=0.6875 train/location_loss=2.132 train/location_acc=0.1875 train/token_losses=25.47 throughput/total_tokens=638,976 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:21:44.045 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=40/600, eta=3 hours, 50 minutes] train/CrossEntropyLoss=0.0472 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.951 train/patch_acc=0.5625 train/subpatch_loss=1.063 train/subpatch_acc=0.5000 train/location_loss=2.169 train/location_acc=0.1875 train/token_losses=28.53 throughput/total_tokens=655,360 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:22:08.575 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=41/600, eta=3 hours, 50 minutes] train/CrossEntropyLoss=0.0645 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=1.961 train/patch_acc=0.5938 train/subpatch_loss=0.7808 train/subpatch_acc=0.8125 train/location_loss=2.074 train/location_acc=0.2500 train/token_losses=27.11 throughput/total_tokens=671,744 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 12:22:33.095 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=42/600, eta=3 hours, 50 minutes] train/CrossEntropyLoss=0.0311 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=2.266 train/patch_acc=0.5625 train/subpatch_loss=0.6734 train/subpatch_acc=0.7500 train/location_loss=2.092 train/location_acc=0.1875 train/token_losses=29.19 throughput/total_tokens=688,128 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 12:22:57.606 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=43/600, eta=3 hours, 49 minutes] train/CrossEntropyLoss=0.0060 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.523 train/patch_acc=0.6250 train/subpatch_loss=0.6371 train/subpatch_acc=0.7500 train/location_loss=2.347 train/location_acc=0.1250 train/token_losses=24.12 throughput/total_tokens=704,512 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 12:23:22.116 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=44/600, eta=3 hours, 49 minutes] train/CrossEntropyLoss=0.0195 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.728 train/patch_acc=0.6250 train/subpatch_loss=1.124 train/subpatch_acc=0.5000 train/location_loss=2.075 train/location_acc=0.0625 train/token_losses=26.62 throughput/total_tokens=720,896 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 12:23:46.629 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=45/600, eta=3 hours, 48 minutes] train/CrossEntropyLoss=0.0820 train/ZLoss=0.0007 train/Accuracy=0.9787 train/patch_loss=1.827 train/patch_acc=0.5625 train/subpatch_loss=0.9899 train/subpatch_acc=0.6875 train/location_loss=2.213 train/location_acc=0.1250 train/token_losses=27.43 throughput/total_tokens=737,280 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 12:24:11.125 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=46/600, eta=3 hours, 48 minutes] train/CrossEntropyLoss=0.0628 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.365 train/patch_acc=0.5312 train/subpatch_loss=1.168 train/subpatch_acc=0.6250 train/location_loss=2.064 train/location_acc=0.0625 train/token_losses=31.85 throughput/total_tokens=753,664 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:24:35.648 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=47/600, eta=3 hours, 47 minutes] train/CrossEntropyLoss=0.0279 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.025 train/patch_acc=0.5312 train/subpatch_loss=1.327 train/subpatch_acc=0.3750 train/location_loss=2.337 train/location_acc=0.1875 train/token_losses=30.85 throughput/total_tokens=770,048 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:25:00.166 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=48/600, eta=3 hours, 47 minutes] train/CrossEntropyLoss=0.0114 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.513 train/patch_acc=0.6562 train/subpatch_loss=0.9656 train/subpatch_acc=0.5625 train/location_loss=2.020 train/location_acc=0.2500 train/token_losses=24.05 throughput/total_tokens=786,432 throughput/device/tokens_per_second=668.2 throughput/device/batches_per_second=0.0408 2026-07-27 12:25:24.690 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=49/600, eta=3 hours, 46 minutes] train/CrossEntropyLoss=0.0334 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.635 train/patch_acc=0.6562 train/subpatch_loss=0.5610 train/subpatch_acc=0.8125 train/location_loss=2.137 train/location_acc=0.0 train/token_losses=23.87 throughput/total_tokens=802,816 throughput/device/tokens_per_second=668.2 throughput/device/batches_per_second=0.0408 2026-07-27 12:25:49.229 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=50/600, eta=3 hours, 46 minutes] train/CrossEntropyLoss=0.0215 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.639 train/patch_acc=0.6250 train/subpatch_loss=0.9289 train/subpatch_acc=0.5625 train/location_loss=2.021 train/location_acc=0.0625 train/token_losses=24.91 throughput/total_tokens=819,200 throughput/device/tokens_per_second=668.2 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:26:13.749 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=51/600, eta=3 hours, 46 minutes] train/CrossEntropyLoss=0.0071 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.657 train/patch_acc=0.6562 train/subpatch_loss=0.6881 train/subpatch_acc=0.7500 train/location_loss=2.049 train/location_acc=0.0625 train/token_losses=24.20 throughput/total_tokens=835,584 throughput/device/tokens_per_second=668.2 throughput/device/batches_per_second=0.0408 2026-07-27 12:26:38.283 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=52/600, eta=3 hours, 45 minutes] train/CrossEntropyLoss=0.0142 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=2.030 train/patch_acc=0.5938 train/subpatch_loss=0.6443 train/subpatch_acc=0.6875 train/location_loss=1.972 train/location_acc=0.1875 train/token_losses=26.71 throughput/total_tokens=851,968 throughput/device/tokens_per_second=668.2 throughput/device/batches_per_second=0.0408 2026-07-27 12:27:02.830 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=53/600, eta=3 hours, 45 minutes] train/CrossEntropyLoss=0.0112 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.548 train/patch_acc=0.6562 train/subpatch_loss=0.8384 train/subpatch_acc=0.5625 train/location_loss=2.154 train/location_acc=0.0625 train/token_losses=24.36 throughput/total_tokens=868,352 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:27:27.353 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=54/600, eta=3 hours, 44 minutes] train/CrossEntropyLoss=0.0262 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.824 train/patch_acc=0.5625 train/subpatch_loss=0.8051 train/subpatch_acc=0.6875 train/location_loss=2.104 train/location_acc=0.1875 train/token_losses=26.23 throughput/total_tokens=884,736 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:27:51.872 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=55/600, eta=3 hours, 44 minutes] train/CrossEntropyLoss=0.0256 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.971 train/patch_acc=0.6250 train/subpatch_loss=0.9569 train/subpatch_acc=0.6250 train/location_loss=2.293 train/location_acc=0.0625 train/token_losses=28.77 throughput/total_tokens=901,120 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:28:16.389 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=56/600, eta=3 hours, 44 minutes] train/CrossEntropyLoss=0.0121 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.754 train/patch_acc=0.5938 train/subpatch_loss=1.037 train/subpatch_acc=0.6875 train/location_loss=1.989 train/location_acc=0.1875 train/token_losses=26.14 throughput/total_tokens=917,504 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:28:40.946 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=57/600, eta=3 hours, 43 minutes] train/CrossEntropyLoss=0.0282 train/ZLoss=0.0004 train/Accuracy=0.9939 train/patch_loss=1.549 train/patch_acc=0.5625 train/subpatch_loss=0.7415 train/subpatch_acc=0.6250 train/location_loss=2.097 train/location_acc=0.3125 train/token_losses=23.74 throughput/total_tokens=933,888 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:29:05.482 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=58/600, eta=3 hours, 43 minutes] train/CrossEntropyLoss=0.0097 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.648 train/patch_acc=0.6875 train/subpatch_loss=0.8569 train/subpatch_acc=0.6875 train/location_loss=2.152 train/location_acc=0.0 train/token_losses=25.22 throughput/total_tokens=950,272 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:29:30.006 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=59/600, eta=3 hours, 42 minutes] train/CrossEntropyLoss=0.1114 train/ZLoss=0.0007 train/Accuracy=0.9726 train/patch_loss=1.768 train/patch_acc=0.5625 train/subpatch_loss=0.6982 train/subpatch_acc=0.6875 train/location_loss=2.263 train/location_acc=0.0625 train/token_losses=25.99 throughput/total_tokens=966,656 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:29:54.538 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=60/600, eta=3 hours, 42 minutes] train/CrossEntropyLoss=0.0627 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.609 train/patch_acc=0.5625 train/subpatch_loss=0.6068 train/subpatch_acc=0.7500 train/location_loss=1.872 train/location_acc=0.3125 train/token_losses=22.79 throughput/total_tokens=983,040 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:30:19.077 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=61/600, eta=3 hours, 41 minutes] train/CrossEntropyLoss=0.1999 train/ZLoss=0.0006 train/Accuracy=0.9543 train/patch_loss=1.820 train/patch_acc=0.6250 train/subpatch_loss=0.6233 train/subpatch_acc=0.6875 train/location_loss=2.193 train/location_acc=0.0625 train/token_losses=25.83 throughput/total_tokens=999,424 throughput/device/tokens_per_second=668.1 throughput/device/batches_per_second=0.0408 2026-07-27 12:30:43.637 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=62/600, eta=3 hours, 41 minutes] train/CrossEntropyLoss=0.0414 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.615 train/patch_acc=0.5938 train/subpatch_loss=0.6822 train/subpatch_acc=0.6875 train/location_loss=2.172 train/location_acc=0.1250 train/token_losses=24.34 throughput/total_tokens=1,015,808 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 12:31:08.171 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=63/600, eta=3 hours, 41 minutes] train/CrossEntropyLoss=0.0276 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.926 train/patch_acc=0.5938 train/subpatch_loss=0.6812 train/subpatch_acc=0.8750 train/location_loss=2.034 train/location_acc=0.1875 train/token_losses=26.27 throughput/total_tokens=1,032,192 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 12:31:32.690 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=64/600, eta=3 hours, 40 minutes] train/CrossEntropyLoss=0.0521 train/ZLoss=0.0002 train/Accuracy=0.9787 train/patch_loss=2.078 train/patch_acc=0.5938 train/subpatch_loss=0.6017 train/subpatch_acc=0.7500 train/location_loss=2.190 train/location_acc=0.2500 train/token_losses=27.79 throughput/total_tokens=1,048,576 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 12:31:57.245 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=65/600, eta=3 hours, 40 minutes] train/CrossEntropyLoss=0.0282 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.650 train/patch_acc=0.6562 train/subpatch_loss=0.8913 train/subpatch_acc=0.7500 train/location_loss=2.158 train/location_acc=0.1250 train/token_losses=25.40 throughput/total_tokens=1,064,960 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 12:32:21.771 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=66/600, eta=3 hours, 39 minutes] train/CrossEntropyLoss=0.0329 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.561 train/patch_acc=0.6250 train/subpatch_loss=0.5292 train/subpatch_acc=0.8750 train/location_loss=2.055 train/location_acc=0.0625 train/token_losses=22.82 throughput/total_tokens=1,081,344 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 12:32:46.308 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=67/600, eta=3 hours, 39 minutes] train/CrossEntropyLoss=0.0409 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.837 train/patch_acc=0.5938 train/subpatch_loss=0.8088 train/subpatch_acc=0.7500 train/location_loss=2.059 train/location_acc=0.1875 train/token_losses=26.16 throughput/total_tokens=1,097,728 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:33:10.827 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=68/600, eta=3 hours, 39 minutes] train/CrossEntropyLoss=0.1565 train/ZLoss=0.0006 train/Accuracy=0.9573 train/patch_loss=2.061 train/patch_acc=0.5938 train/subpatch_loss=0.9147 train/subpatch_acc=0.5625 train/location_loss=2.236 train/location_acc=0.1250 train/token_losses=29.09 throughput/total_tokens=1,114,112 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:33:35.360 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=69/600, eta=3 hours, 38 minutes] train/CrossEntropyLoss=0.0007 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=1.928 train/patch_acc=0.6250 train/subpatch_loss=0.5580 train/subpatch_acc=0.8125 train/location_loss=2.060 train/location_acc=0.3125 train/token_losses=25.90 throughput/total_tokens=1,130,496 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:33:59.910 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=70/600, eta=3 hours, 38 minutes] train/CrossEntropyLoss=0.0874 train/ZLoss=0.0005 train/Accuracy=0.9726 train/patch_loss=1.382 train/patch_acc=0.6250 train/subpatch_loss=0.3766 train/subpatch_acc=0.8750 train/location_loss=1.952 train/location_acc=0.3125 train/token_losses=20.37 throughput/total_tokens=1,146,880 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:34:24.428 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=71/600, eta=3 hours, 37 minutes] train/CrossEntropyLoss=0.0695 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.619 train/patch_acc=0.6250 train/subpatch_loss=0.7147 train/subpatch_acc=0.6875 train/location_loss=1.833 train/location_acc=0.1875 train/token_losses=23.14 throughput/total_tokens=1,163,264 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:34:48.970 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=72/600, eta=3 hours, 37 minutes] train/CrossEntropyLoss=0.0145 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.701 train/patch_acc=0.5625 train/subpatch_loss=0.5460 train/subpatch_acc=0.8125 train/location_loss=1.833 train/location_acc=0.2500 train/token_losses=23.12 throughput/total_tokens=1,179,648 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:35:13.523 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=73/600, eta=3 hours, 36 minutes] train/CrossEntropyLoss=0.0031 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=1.572 train/patch_acc=0.5938 train/subpatch_loss=0.5784 train/subpatch_acc=0.7500 train/location_loss=2.106 train/location_acc=0.1250 train/token_losses=23.31 throughput/total_tokens=1,196,032 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:35:38.034 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=74/600, eta=3 hours, 36 minutes] train/CrossEntropyLoss=0.0123 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=1.521 train/patch_acc=0.6250 train/subpatch_loss=0.4288 train/subpatch_acc=0.8125 train/location_loss=2.005 train/location_acc=0.2500 train/token_losses=21.91 throughput/total_tokens=1,212,416 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:36:02.576 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=75/600, eta=3 hours, 36 minutes] train/CrossEntropyLoss=0.0408 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.365 train/patch_acc=0.6562 train/subpatch_loss=0.5951 train/subpatch_acc=0.6875 train/location_loss=2.282 train/location_acc=0.0 train/token_losses=22.43 throughput/total_tokens=1,228,800 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:36:27.093 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=76/600, eta=3 hours, 35 minutes] train/CrossEntropyLoss=0.0704 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.414 train/patch_acc=0.6875 train/subpatch_loss=0.7394 train/subpatch_acc=0.6875 train/location_loss=1.962 train/location_acc=0.1250 train/token_losses=22.12 throughput/total_tokens=1,245,184 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:36:51.651 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=77/600, eta=3 hours, 35 minutes] train/CrossEntropyLoss=0.0054 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.185 train/patch_acc=0.6562 train/subpatch_loss=0.5053 train/subpatch_acc=0.8750 train/location_loss=1.873 train/location_acc=0.3750 train/token_losses=18.99 throughput/total_tokens=1,261,568 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:37:16.193 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=78/600, eta=3 hours, 34 minutes] train/CrossEntropyLoss=0.0450 train/ZLoss=0.0004 train/Accuracy=0.9848 train/patch_loss=1.284 train/patch_acc=0.6562 train/subpatch_loss=0.5920 train/subpatch_acc=0.7500 train/location_loss=2.061 train/location_acc=0.1250 train/token_losses=20.89 throughput/total_tokens=1,277,952 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:37:40.736 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=79/600, eta=3 hours, 34 minutes] train/CrossEntropyLoss=0.0126 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.665 train/patch_acc=0.6250 train/subpatch_loss=0.7387 train/subpatch_acc=0.7500 train/location_loss=1.927 train/location_acc=0.3750 train/token_losses=23.98 throughput/total_tokens=1,294,336 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:38:05.264 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=80/600, eta=3 hours, 34 minutes] train/CrossEntropyLoss=0.0672 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=1.334 train/patch_acc=0.6562 train/subpatch_loss=0.4021 train/subpatch_acc=0.8750 train/location_loss=2.202 train/location_acc=0.2500 train/token_losses=21.08 throughput/total_tokens=1,310,720 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:38:29.800 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=81/600, eta=3 hours, 33 minutes] train/CrossEntropyLoss=0.0382 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.330 train/patch_acc=0.6562 train/subpatch_loss=0.4915 train/subpatch_acc=0.7500 train/location_loss=1.935 train/location_acc=0.3750 train/token_losses=20.35 throughput/total_tokens=1,327,104 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:38:54.347 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=82/600, eta=3 hours, 33 minutes] train/CrossEntropyLoss=0.0272 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.371 train/patch_acc=0.6875 train/subpatch_loss=0.5538 train/subpatch_acc=0.8125 train/location_loss=2.021 train/location_acc=0.0625 train/token_losses=21.26 throughput/total_tokens=1,343,488 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:39:18.902 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=83/600, eta=3 hours, 32 minutes] train/CrossEntropyLoss=0.1663 train/ZLoss=0.0006 train/Accuracy=0.9634 train/patch_loss=1.777 train/patch_acc=0.5625 train/subpatch_loss=0.4451 train/subpatch_acc=0.8750 train/location_loss=1.935 train/location_acc=0.3125 train/token_losses=23.74 throughput/total_tokens=1,359,872 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:39:43.416 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=84/600, eta=3 hours, 32 minutes] train/CrossEntropyLoss=0.0263 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.054 train/patch_acc=0.6875 train/subpatch_loss=0.3436 train/subpatch_acc=0.8125 train/location_loss=1.935 train/location_acc=0.4375 train/token_losses=17.54 throughput/total_tokens=1,376,256 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:40:07.931 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=85/600, eta=3 hours, 31 minutes] train/CrossEntropyLoss=0.0121 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.605 train/patch_acc=0.5938 train/subpatch_loss=0.7571 train/subpatch_acc=0.6875 train/location_loss=2.133 train/location_acc=0.1250 train/token_losses=24.40 throughput/total_tokens=1,392,640 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:40:32.451 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=86/600, eta=3 hours, 31 minutes] train/CrossEntropyLoss=0.0221 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.327 train/patch_acc=0.6250 train/subpatch_loss=0.5604 train/subpatch_acc=0.7500 train/location_loss=1.730 train/location_acc=0.4375 train/token_losses=19.78 throughput/total_tokens=1,409,024 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:40:56.982 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=87/600, eta=3 hours, 31 minutes] train/CrossEntropyLoss=0.0425 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.434 train/patch_acc=0.6875 train/subpatch_loss=0.3956 train/subpatch_acc=0.9375 train/location_loss=1.915 train/location_acc=0.2500 train/token_losses=20.71 throughput/total_tokens=1,425,408 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:41:21.498 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=88/600, eta=3 hours, 30 minutes] train/CrossEntropyLoss=0.0336 train/ZLoss=0.0003 train/Accuracy=0.9939 train/patch_loss=1.891 train/patch_acc=0.5938 train/subpatch_loss=0.7188 train/subpatch_acc=0.6875 train/location_loss=1.962 train/location_acc=0.1875 train/token_losses=25.85 throughput/total_tokens=1,441,792 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:41:46.026 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=89/600, eta=3 hours, 30 minutes] train/CrossEntropyLoss=0.0586 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.790 train/patch_acc=0.5625 train/subpatch_loss=0.8238 train/subpatch_acc=0.6250 train/location_loss=2.065 train/location_acc=0.2500 train/token_losses=25.87 throughput/total_tokens=1,458,176 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:42:10.557 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=90/600, eta=3 hours, 29 minutes] train/CrossEntropyLoss=0.0620 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.413 train/patch_acc=0.6875 train/subpatch_loss=0.3235 train/subpatch_acc=0.9375 train/location_loss=1.841 train/location_acc=0.3750 train/token_losses=19.96 throughput/total_tokens=1,474,560 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:42:35.084 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=91/600, eta=3 hours, 29 minutes] train/CrossEntropyLoss=0.0473 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.697 train/patch_acc=0.6250 train/subpatch_loss=0.4441 train/subpatch_acc=0.8750 train/location_loss=1.935 train/location_acc=0.1875 train/token_losses=23.10 throughput/total_tokens=1,490,944 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:42:59.633 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=92/600, eta=3 hours, 29 minutes] train/CrossEntropyLoss=0.0487 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.659 train/patch_acc=0.6250 train/subpatch_loss=0.7066 train/subpatch_acc=0.8125 train/location_loss=1.826 train/location_acc=0.3750 train/token_losses=23.40 throughput/total_tokens=1,507,328 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:43:24.196 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=93/600, eta=3 hours, 28 minutes] train/CrossEntropyLoss=0.0575 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.521 train/patch_acc=0.6562 train/subpatch_loss=0.6655 train/subpatch_acc=0.6875 train/location_loss=2.134 train/location_acc=0.0 train/token_losses=23.36 throughput/total_tokens=1,523,712 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:43:48.736 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=94/600, eta=3 hours, 28 minutes] train/CrossEntropyLoss=0.0360 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.240 train/patch_acc=0.7188 train/subpatch_loss=0.5959 train/subpatch_acc=0.7500 train/location_loss=2.001 train/location_acc=0.2500 train/token_losses=20.31 throughput/total_tokens=1,540,096 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:44:13.268 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=95/600, eta=3 hours, 27 minutes] train/CrossEntropyLoss=0.0269 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.484 train/patch_acc=0.5938 train/subpatch_loss=0.7732 train/subpatch_acc=0.5625 train/location_loss=1.748 train/location_acc=0.3125 train/token_losses=21.96 throughput/total_tokens=1,556,480 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:44:37.818 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=96/600, eta=3 hours, 27 minutes] train/CrossEntropyLoss=0.0221 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.193 train/patch_acc=0.6875 train/subpatch_loss=0.3718 train/subpatch_acc=0.8750 train/location_loss=2.161 train/location_acc=0.1250 train/token_losses=19.67 throughput/total_tokens=1,572,864 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:45:02.356 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=97/600, eta=3 hours, 26 minutes] train/CrossEntropyLoss=0.0637 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.708 train/patch_acc=0.5938 train/subpatch_loss=0.3657 train/subpatch_acc=0.8750 train/location_loss=1.853 train/location_acc=0.3750 train/token_losses=22.54 throughput/total_tokens=1,589,248 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:45:26.896 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=98/600, eta=3 hours, 26 minutes] train/CrossEntropyLoss=0.1550 train/ZLoss=0.0006 train/Accuracy=0.9665 train/patch_loss=1.640 train/patch_acc=0.5938 train/subpatch_loss=0.3962 train/subpatch_acc=0.9375 train/location_loss=1.976 train/location_acc=0.1875 train/token_losses=22.61 throughput/total_tokens=1,605,632 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:45:51.457 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=99/600, eta=3 hours, 26 minutes] train/CrossEntropyLoss=0.0322 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.305 train/patch_acc=0.6875 train/subpatch_loss=0.7145 train/subpatch_acc=0.7500 train/location_loss=2.007 train/location_acc=0.2500 train/token_losses=21.32 throughput/total_tokens=1,622,016 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:46:15.992 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=100/600, eta=3 hours, 25 minutes] train/CrossEntropyLoss=0.0467 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=2.040 train/patch_acc=0.5938 train/subpatch_loss=0.6327 train/subpatch_acc=0.8125 train/location_loss=2.074 train/location_acc=0.1875 train/token_losses=27.15 throughput/total_tokens=1,638,400 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:46:15.993 2012bc4f83ae:0 olmo.train.trainer:1415 INFO Saving checkpoint... 2026-07-27 12:46:17.871 2012bc4f83ae:0 olmo.train.trainer:1420 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step100 2026-07-27 12:46:42.857 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=101/600, eta=3 hours, 25 minutes] train/CrossEntropyLoss=0.0013 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=1.330 train/patch_acc=0.6250 train/subpatch_loss=0.2301 train/subpatch_acc=0.9375 train/location_loss=1.832 train/location_acc=0.3750 train/token_losses=18.89 throughput/total_tokens=1,654,784 throughput/device/tokens_per_second=655.8 throughput/device/batches_per_second=0.0400 2026-07-27 12:47:07.413 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=102/600, eta=3 hours, 25 minutes] train/CrossEntropyLoss=0.0579 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.655 train/patch_acc=0.5625 train/subpatch_loss=0.5927 train/subpatch_acc=0.8125 train/location_loss=2.163 train/location_acc=0.1250 train/token_losses=24.26 throughput/total_tokens=1,671,168 throughput/device/tokens_per_second=661.5 throughput/device/batches_per_second=0.0404 2026-07-27 12:47:31.938 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=103/600, eta=3 hours, 24 minutes] train/CrossEntropyLoss=0.0151 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.400 train/patch_acc=0.6875 train/subpatch_loss=0.5329 train/subpatch_acc=0.8750 train/location_loss=1.983 train/location_acc=0.2500 train/token_losses=21.26 throughput/total_tokens=1,687,552 throughput/device/tokens_per_second=663.6 throughput/device/batches_per_second=0.0405 2026-07-27 12:47:56.480 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=104/600, eta=3 hours, 24 minutes] train/CrossEntropyLoss=0.0365 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.092 train/patch_acc=0.7188 train/subpatch_loss=0.8266 train/subpatch_acc=0.7500 train/location_loss=1.855 train/location_acc=0.2500 train/token_losses=19.46 throughput/total_tokens=1,703,936 throughput/device/tokens_per_second=664.6 throughput/device/batches_per_second=0.0406 2026-07-27 12:48:21.042 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=105/600, eta=3 hours, 23 minutes] train/CrossEntropyLoss=0.2087 train/ZLoss=0.0007 train/Accuracy=0.9482 train/patch_loss=1.456 train/patch_acc=0.6250 train/subpatch_loss=0.6834 train/subpatch_acc=0.6875 train/location_loss=1.888 train/location_acc=0.2500 train/token_losses=21.94 throughput/total_tokens=1,720,320 throughput/device/tokens_per_second=665.1 throughput/device/batches_per_second=0.0406 2026-07-27 12:48:45.600 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=106/600, eta=3 hours, 23 minutes] train/CrossEntropyLoss=0.0760 train/ZLoss=0.0001 train/Accuracy=0.9787 train/patch_loss=1.217 train/patch_acc=0.6875 train/subpatch_loss=0.5997 train/subpatch_acc=0.6875 train/location_loss=1.830 train/location_acc=0.2500 train/token_losses=19.45 throughput/total_tokens=1,736,704 throughput/device/tokens_per_second=665.5 throughput/device/batches_per_second=0.0406 2026-07-27 12:49:10.168 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=107/600, eta=3 hours, 23 minutes] train/CrossEntropyLoss=0.1519 train/ZLoss=0.0011 train/Accuracy=0.9573 train/patch_loss=1.339 train/patch_acc=0.6562 train/subpatch_loss=0.5618 train/subpatch_acc=0.7500 train/location_loss=1.938 train/location_acc=0.3750 train/token_losses=20.71 throughput/total_tokens=1,753,088 throughput/device/tokens_per_second=665.7 throughput/device/batches_per_second=0.0406 2026-07-27 12:49:34.760 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=108/600, eta=3 hours, 22 minutes] train/CrossEntropyLoss=0.0192 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.366 train/patch_acc=0.6250 train/subpatch_loss=0.4100 train/subpatch_acc=1.0000 train/location_loss=1.883 train/location_acc=0.2500 train/token_losses=20.09 throughput/total_tokens=1,769,472 throughput/device/tokens_per_second=665.7 throughput/device/batches_per_second=0.0406 2026-07-27 12:49:59.342 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=109/600, eta=3 hours, 22 minutes] train/CrossEntropyLoss=0.0426 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.607 train/patch_acc=0.5625 train/subpatch_loss=0.6100 train/subpatch_acc=0.7500 train/location_loss=1.431 train/location_acc=0.7500 train/token_losses=21.02 throughput/total_tokens=1,785,856 throughput/device/tokens_per_second=665.8 throughput/device/batches_per_second=0.0406 2026-07-27 12:50:23.888 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=110/600, eta=3 hours, 21 minutes] train/CrossEntropyLoss=0.1849 train/ZLoss=0.0006 train/Accuracy=0.9573 train/patch_loss=1.392 train/patch_acc=0.6250 train/subpatch_loss=0.6008 train/subpatch_acc=0.8125 train/location_loss=2.211 train/location_acc=0.1250 train/token_losses=22.39 throughput/total_tokens=1,802,240 throughput/device/tokens_per_second=666.0 throughput/device/batches_per_second=0.0406 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:50:48.413 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=111/600, eta=3 hours, 21 minutes] train/CrossEntropyLoss=0.0166 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.419 train/patch_acc=0.5938 train/subpatch_loss=0.3755 train/subpatch_acc=0.9375 train/location_loss=1.551 train/location_acc=0.3125 train/token_losses=19.05 throughput/total_tokens=1,818,624 throughput/device/tokens_per_second=666.2 throughput/device/batches_per_second=0.0407 2026-07-27 12:51:12.958 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=112/600, eta=3 hours, 20 minutes] train/CrossEntropyLoss=0.0223 train/ZLoss=0.0005 train/Accuracy=0.9909 train/patch_loss=1.411 train/patch_acc=0.6562 train/subpatch_loss=0.4618 train/subpatch_acc=0.8125 train/location_loss=2.008 train/location_acc=0.3750 train/token_losses=21.17 throughput/total_tokens=1,835,008 throughput/device/tokens_per_second=666.3 throughput/device/batches_per_second=0.0407 2026-07-27 12:51:37.516 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=113/600, eta=3 hours, 20 minutes] train/CrossEntropyLoss=0.0436 train/ZLoss=0.0005 train/Accuracy=0.9787 train/patch_loss=1.517 train/patch_acc=0.5938 train/subpatch_loss=0.2870 train/subpatch_acc=0.9375 train/location_loss=1.583 train/location_acc=0.3125 train/token_losses=19.61 throughput/total_tokens=1,851,392 throughput/device/tokens_per_second=666.3 throughput/device/batches_per_second=0.0407 2026-07-27 12:52:02.061 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=114/600, eta=3 hours, 20 minutes] train/CrossEntropyLoss=0.0157 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.549 train/patch_acc=0.6250 train/subpatch_loss=0.5199 train/subpatch_acc=0.8125 train/location_loss=2.054 train/location_acc=0.3125 train/token_losses=22.69 throughput/total_tokens=1,867,776 throughput/device/tokens_per_second=666.4 throughput/device/batches_per_second=0.0407 2026-07-27 12:52:26.587 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=115/600, eta=3 hours, 19 minutes] train/CrossEntropyLoss=0.0375 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.185 train/patch_acc=0.6562 train/subpatch_loss=0.8220 train/subpatch_acc=0.6250 train/location_loss=1.838 train/location_acc=0.1875 train/token_losses=20.12 throughput/total_tokens=1,884,160 throughput/device/tokens_per_second=666.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:52:51.104 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=116/600, eta=3 hours, 19 minutes] train/CrossEntropyLoss=0.0118 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.379 train/patch_acc=0.6250 train/subpatch_loss=0.3588 train/subpatch_acc=0.8125 train/location_loss=1.447 train/location_acc=0.3750 train/token_losses=18.25 throughput/total_tokens=1,900,544 throughput/device/tokens_per_second=666.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:53:15.636 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=117/600, eta=3 hours, 18 minutes] train/CrossEntropyLoss=0.0422 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.358 train/patch_acc=0.5938 train/subpatch_loss=0.2071 train/subpatch_acc=1.0000 train/location_loss=1.697 train/location_acc=0.4375 train/token_losses=18.48 throughput/total_tokens=1,916,928 throughput/device/tokens_per_second=666.7 throughput/device/batches_per_second=0.0407 2026-07-27 12:53:40.174 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=118/600, eta=3 hours, 18 minutes] train/CrossEntropyLoss=0.0236 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.336 train/patch_acc=0.6562 train/subpatch_loss=0.6415 train/subpatch_acc=0.7500 train/location_loss=1.893 train/location_acc=0.1875 train/token_losses=20.82 throughput/total_tokens=1,933,312 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 12:54:04.703 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=119/600, eta=3 hours, 18 minutes] train/CrossEntropyLoss=0.0440 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=1.170 train/patch_acc=0.7188 train/subpatch_loss=0.4573 train/subpatch_acc=0.8125 train/location_loss=1.937 train/location_acc=0.1250 train/token_losses=18.93 throughput/total_tokens=1,949,696 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 12:54:29.245 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=120/600, eta=3 hours, 17 minutes] train/CrossEntropyLoss=0.0082 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.337 train/patch_acc=0.5938 train/subpatch_loss=0.3712 train/subpatch_acc=0.9375 train/location_loss=1.791 train/location_acc=0.2500 train/token_losses=19.34 throughput/total_tokens=1,966,080 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:54:53.775 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=121/600, eta=3 hours, 17 minutes] train/CrossEntropyLoss=0.0336 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.396 train/patch_acc=0.6875 train/subpatch_loss=0.3610 train/subpatch_acc=0.8125 train/location_loss=1.605 train/location_acc=0.5000 train/token_losses=19.03 throughput/total_tokens=1,982,464 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:55:18.316 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=122/600, eta=3 hours, 16 minutes] train/CrossEntropyLoss=0.1988 train/ZLoss=0.0008 train/Accuracy=0.9512 train/patch_loss=1.206 train/patch_acc=0.6875 train/subpatch_loss=0.3625 train/subpatch_acc=0.8750 train/location_loss=1.861 train/location_acc=0.3750 train/token_losses=18.54 throughput/total_tokens=1,998,848 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:55:42.865 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=123/600, eta=3 hours, 16 minutes] train/CrossEntropyLoss=0.0182 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9094 train/patch_acc=0.7500 train/subpatch_loss=0.4175 train/subpatch_acc=0.8125 train/location_loss=1.828 train/location_acc=0.4375 train/token_losses=16.26 throughput/total_tokens=2,015,232 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:56:07.401 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=124/600, eta=3 hours, 16 minutes] train/CrossEntropyLoss=0.0139 train/ZLoss=0.0005 train/Accuracy=0.9939 train/patch_loss=1.441 train/patch_acc=0.5938 train/subpatch_loss=0.2543 train/subpatch_acc=1.0000 train/location_loss=1.819 train/location_acc=0.1875 train/token_losses=19.82 throughput/total_tokens=2,031,616 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:56:31.925 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=125/600, eta=3 hours, 15 minutes] train/CrossEntropyLoss=0.0238 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.553 train/patch_acc=0.5312 train/subpatch_loss=0.4405 train/subpatch_acc=0.9375 train/location_loss=1.780 train/location_acc=0.2500 train/token_losses=21.31 throughput/total_tokens=2,048,000 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 12:56:56.444 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=126/600, eta=3 hours, 15 minutes] train/CrossEntropyLoss=0.1086 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=1.395 train/patch_acc=0.7188 train/subpatch_loss=0.5356 train/subpatch_acc=0.8750 train/location_loss=1.939 train/location_acc=0.2500 train/token_losses=21.05 throughput/total_tokens=2,064,384 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:57:21.001 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=127/600, eta=3 hours, 14 minutes] train/CrossEntropyLoss=0.0365 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.157 train/patch_acc=0.7188 train/subpatch_loss=0.5170 train/subpatch_acc=0.8125 train/location_loss=1.781 train/location_acc=0.3125 train/token_losses=18.44 throughput/total_tokens=2,080,768 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 12:57:45.562 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=128/600, eta=3 hours, 14 minutes] train/CrossEntropyLoss=0.0266 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8668 train/patch_acc=0.7500 train/subpatch_loss=0.2891 train/subpatch_acc=0.9375 train/location_loss=1.856 train/location_acc=0.1875 train/token_losses=15.51 throughput/total_tokens=2,097,152 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:58:10.086 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=129/600, eta=3 hours, 13 minutes] train/CrossEntropyLoss=0.0334 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.429 train/patch_acc=0.5938 train/subpatch_loss=0.2467 train/subpatch_acc=0.9375 train/location_loss=1.959 train/location_acc=0.1250 train/token_losses=20.26 throughput/total_tokens=2,113,536 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:58:34.624 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=130/600, eta=3 hours, 13 minutes] train/CrossEntropyLoss=0.1707 train/ZLoss=0.0006 train/Accuracy=0.9573 train/patch_loss=1.732 train/patch_acc=0.5938 train/subpatch_loss=0.6073 train/subpatch_acc=0.8125 train/location_loss=2.004 train/location_acc=0.1250 train/token_losses=24.30 throughput/total_tokens=2,129,920 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 12:58:59.150 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=131/600, eta=3 hours, 13 minutes] train/CrossEntropyLoss=0.0336 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.114 train/patch_acc=0.8125 train/subpatch_loss=0.3193 train/subpatch_acc=0.8750 train/location_loss=1.713 train/location_acc=0.3750 train/token_losses=17.04 throughput/total_tokens=2,146,304 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 12:59:23.683 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=132/600, eta=3 hours, 12 minutes] train/CrossEntropyLoss=0.1020 train/ZLoss=0.0007 train/Accuracy=0.9695 train/patch_loss=1.392 train/patch_acc=0.5938 train/subpatch_loss=0.4300 train/subpatch_acc=0.8125 train/location_loss=1.842 train/location_acc=0.5000 train/token_losses=20.22 throughput/total_tokens=2,162,688 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 12:59:48.202 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=133/600, eta=3 hours, 12 minutes] train/CrossEntropyLoss=0.0605 train/ZLoss=0.0007 train/Accuracy=0.9848 train/patch_loss=1.475 train/patch_acc=0.6875 train/subpatch_loss=0.3237 train/subpatch_acc=0.8750 train/location_loss=1.845 train/location_acc=0.3125 train/token_losses=20.47 throughput/total_tokens=2,179,072 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:00:12.767 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=134/600, eta=3 hours, 11 minutes] train/CrossEntropyLoss=0.0575 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.634 train/patch_acc=0.5312 train/subpatch_loss=0.6145 train/subpatch_acc=0.6875 train/location_loss=1.942 train/location_acc=0.1875 train/token_losses=23.30 throughput/total_tokens=2,195,456 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:00:37.283 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=135/600, eta=3 hours, 11 minutes] train/CrossEntropyLoss=0.0111 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.194 train/patch_acc=0.6562 train/subpatch_loss=0.2538 train/subpatch_acc=0.9375 train/location_loss=1.749 train/location_acc=0.4375 train/token_losses=17.57 throughput/total_tokens=2,211,840 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:01:01.819 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=136/600, eta=3 hours, 11 minutes] train/CrossEntropyLoss=0.0540 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.416 train/patch_acc=0.6250 train/subpatch_loss=0.4363 train/subpatch_acc=0.8750 train/location_loss=1.720 train/location_acc=0.3750 train/token_losses=19.95 throughput/total_tokens=2,228,224 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:01:26.392 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=137/600, eta=3 hours, 10 minutes] train/CrossEntropyLoss=0.0908 train/ZLoss=0.0006 train/Accuracy=0.9756 train/patch_loss=1.184 train/patch_acc=0.6875 train/subpatch_loss=0.4609 train/subpatch_acc=0.7500 train/location_loss=1.786 train/location_acc=0.3125 train/token_losses=18.46 throughput/total_tokens=2,244,608 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:01:50.929 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=138/600, eta=3 hours, 10 minutes] train/CrossEntropyLoss=0.0051 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.359 train/patch_acc=0.6562 train/subpatch_loss=0.3542 train/subpatch_acc=0.8125 train/location_loss=1.872 train/location_acc=0.3125 train/token_losses=19.78 throughput/total_tokens=2,260,992 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:02:15.465 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=139/600, eta=3 hours, 9 minutes] train/CrossEntropyLoss=0.0216 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.101 train/patch_acc=0.6875 train/subpatch_loss=0.4480 train/subpatch_acc=0.8125 train/location_loss=1.658 train/location_acc=0.3125 train/token_losses=17.24 throughput/total_tokens=2,277,376 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:02:39.993 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=140/600, eta=3 hours, 9 minutes] train/CrossEntropyLoss=0.0368 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9627 train/patch_acc=0.7188 train/subpatch_loss=0.3435 train/subpatch_acc=0.9375 train/location_loss=1.804 train/location_acc=0.3125 train/token_losses=16.29 throughput/total_tokens=2,293,760 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:03:04.520 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=141/600, eta=3 hours, 9 minutes] train/CrossEntropyLoss=0.0525 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.332 train/patch_acc=0.6562 train/subpatch_loss=0.4711 train/subpatch_acc=0.8125 train/location_loss=1.731 train/location_acc=0.2500 train/token_losses=19.46 throughput/total_tokens=2,310,144 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:03:29.061 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=142/600, eta=3 hours, 8 minutes] train/CrossEntropyLoss=0.0642 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.9765 train/patch_acc=0.7188 train/subpatch_loss=0.3014 train/subpatch_acc=0.8750 train/location_loss=1.723 train/location_acc=0.4375 train/token_losses=15.91 throughput/total_tokens=2,326,528 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:03:53.595 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=143/600, eta=3 hours, 8 minutes] train/CrossEntropyLoss=0.0065 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.298 train/patch_acc=0.6562 train/subpatch_loss=0.3005 train/subpatch_acc=0.9375 train/location_loss=1.505 train/location_acc=0.3125 train/token_losses=17.60 throughput/total_tokens=2,342,912 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:04:18.117 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=144/600, eta=3 hours, 7 minutes] train/CrossEntropyLoss=0.0146 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.491 train/patch_acc=0.6250 train/subpatch_loss=0.4538 train/subpatch_acc=0.7500 train/location_loss=1.602 train/location_acc=0.3125 train/token_losses=20.15 throughput/total_tokens=2,359,296 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:04:42.628 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=145/600, eta=3 hours, 7 minutes] train/CrossEntropyLoss=0.0028 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=1.285 train/patch_acc=0.6562 train/subpatch_loss=0.3834 train/subpatch_acc=0.8750 train/location_loss=1.727 train/location_acc=0.1875 train/token_losses=18.72 throughput/total_tokens=2,375,680 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:05:07.162 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=146/600, eta=3 hours, 6 minutes] train/CrossEntropyLoss=0.0282 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.255 train/patch_acc=0.6562 train/subpatch_loss=0.3107 train/subpatch_acc=0.8750 train/location_loss=1.670 train/location_acc=0.3125 train/token_losses=17.97 throughput/total_tokens=2,392,064 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:05:31.677 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=147/600, eta=3 hours, 6 minutes] train/CrossEntropyLoss=0.0533 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.565 train/patch_acc=0.5938 train/subpatch_loss=0.3059 train/subpatch_acc=0.9375 train/location_loss=1.772 train/location_acc=0.3125 train/token_losses=20.83 throughput/total_tokens=2,408,448 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:05:56.203 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=148/600, eta=3 hours, 6 minutes] train/CrossEntropyLoss=0.0157 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.108 train/patch_acc=0.7188 train/subpatch_loss=0.2035 train/subpatch_acc=1.0000 train/location_loss=1.672 train/location_acc=0.5000 train/token_losses=16.37 throughput/total_tokens=2,424,832 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:06:20.727 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=149/600, eta=3 hours, 5 minutes] train/CrossEntropyLoss=0.0101 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.064 train/patch_acc=0.7188 train/subpatch_loss=0.4900 train/subpatch_acc=0.8125 train/location_loss=1.786 train/location_acc=0.2500 train/token_losses=17.62 throughput/total_tokens=2,441,216 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:06:45.254 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=150/600, eta=3 hours, 5 minutes] train/CrossEntropyLoss=0.0163 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.165 train/patch_acc=0.7812 train/subpatch_loss=0.2850 train/subpatch_acc=1.0000 train/location_loss=1.599 train/location_acc=0.4375 train/token_losses=16.86 throughput/total_tokens=2,457,600 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:07:09.800 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=151/600, eta=3 hours, 4 minutes] train/CrossEntropyLoss=0.0474 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=1.170 train/patch_acc=0.7188 train/subpatch_loss=0.4678 train/subpatch_acc=0.8125 train/location_loss=1.663 train/location_acc=0.2500 train/token_losses=17.88 throughput/total_tokens=2,473,984 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:07:34.328 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=152/600, eta=3 hours, 4 minutes] train/CrossEntropyLoss=0.0191 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.241 train/patch_acc=0.6250 train/subpatch_loss=0.3746 train/subpatch_acc=0.9375 train/location_loss=1.537 train/location_acc=0.3750 train/token_losses=17.58 throughput/total_tokens=2,490,368 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:07:58.863 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=153/600, eta=3 hours, 4 minutes] train/CrossEntropyLoss=0.0190 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.465 train/patch_acc=0.5938 train/subpatch_loss=0.3709 train/subpatch_acc=0.8750 train/location_loss=1.708 train/location_acc=0.3750 train/token_losses=20.04 throughput/total_tokens=2,506,752 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:08:23.403 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=154/600, eta=3 hours, 3 minutes] train/CrossEntropyLoss=0.0112 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8831 train/patch_acc=0.7500 train/subpatch_loss=0.3709 train/subpatch_acc=0.8750 train/location_loss=1.668 train/location_acc=0.3125 train/token_losses=15.22 throughput/total_tokens=2,523,136 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:08:47.931 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=155/600, eta=3 hours, 3 minutes] train/CrossEntropyLoss=0.0064 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.453 train/patch_acc=0.6875 train/subpatch_loss=0.4556 train/subpatch_acc=0.8125 train/location_loss=1.675 train/location_acc=0.4375 train/token_losses=20.15 throughput/total_tokens=2,539,520 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:09:12.452 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=156/600, eta=3 hours, 2 minutes] train/CrossEntropyLoss=0.0109 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.463 train/patch_acc=0.6250 train/subpatch_loss=0.3136 train/subpatch_acc=0.8750 train/location_loss=1.628 train/location_acc=0.3125 train/token_losses=19.47 throughput/total_tokens=2,555,904 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:09:36.991 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=157/600, eta=3 hours, 2 minutes] train/CrossEntropyLoss=0.0836 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.431 train/patch_acc=0.6250 train/subpatch_loss=0.4361 train/subpatch_acc=0.8750 train/location_loss=1.859 train/location_acc=0.2500 train/token_losses=20.63 throughput/total_tokens=2,572,288 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:10:01.545 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=158/600, eta=3 hours, 2 minutes] train/CrossEntropyLoss=0.0269 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.083 train/patch_acc=0.6562 train/subpatch_loss=0.3754 train/subpatch_acc=0.8125 train/location_loss=1.622 train/location_acc=0.3750 train/token_losses=16.65 throughput/total_tokens=2,588,672 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:10:26.079 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=159/600, eta=3 hours, 1 minute] train/CrossEntropyLoss=0.0153 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.072 train/patch_acc=0.7188 train/subpatch_loss=0.2790 train/subpatch_acc=0.8750 train/location_loss=1.598 train/location_acc=0.4375 train/token_losses=16.08 throughput/total_tokens=2,605,056 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:10:50.618 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=160/600, eta=3 hours, 1 minute] train/CrossEntropyLoss=0.0260 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.132 train/patch_acc=0.6562 train/subpatch_loss=0.2304 train/subpatch_acc=0.9375 train/location_loss=1.668 train/location_acc=0.5000 train/token_losses=16.65 throughput/total_tokens=2,621,440 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:11:15.129 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=161/600, eta=3 hours] train/CrossEntropyLoss=0.0129 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.368 train/patch_acc=0.6250 train/subpatch_loss=0.5876 train/subpatch_acc=0.8750 train/location_loss=1.330 train/location_acc=0.5000 train/token_losses=18.62 throughput/total_tokens=2,637,824 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:11:39.657 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=162/600, eta=3 hours] train/CrossEntropyLoss=0.0313 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.614 train/patch_acc=0.6562 train/subpatch_loss=0.5735 train/subpatch_acc=0.7500 train/location_loss=1.315 train/location_acc=0.5000 train/token_losses=20.47 throughput/total_tokens=2,654,208 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:12:04.193 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=163/600, eta=2 hours, 59 minutes] train/CrossEntropyLoss=0.0144 train/ZLoss=0.0006 train/Accuracy=0.9878 train/patch_loss=0.9331 train/patch_acc=0.7812 train/subpatch_loss=0.3364 train/subpatch_acc=0.8125 train/location_loss=1.756 train/location_acc=0.3125 train/token_losses=15.84 throughput/total_tokens=2,670,592 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:12:28.734 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=164/600, eta=2 hours, 59 minutes] train/CrossEntropyLoss=0.0112 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.141 train/patch_acc=0.6875 train/subpatch_loss=0.3451 train/subpatch_acc=0.8750 train/location_loss=1.258 train/location_acc=0.6875 train/token_losses=15.54 throughput/total_tokens=2,686,976 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:12:53.281 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=165/600, eta=2 hours, 59 minutes] train/CrossEntropyLoss=0.0094 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.152 train/patch_acc=0.6250 train/subpatch_loss=0.1018 train/subpatch_acc=1.0000 train/location_loss=1.549 train/location_acc=0.4375 train/token_losses=15.82 throughput/total_tokens=2,703,360 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:13:17.823 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=166/600, eta=2 hours, 58 minutes] train/CrossEntropyLoss=0.0410 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.257 train/patch_acc=0.6875 train/subpatch_loss=0.1011 train/subpatch_acc=1.0000 train/location_loss=1.639 train/location_acc=0.3125 train/token_losses=17.01 throughput/total_tokens=2,719,744 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:13:42.371 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=167/600, eta=2 hours, 58 minutes] train/CrossEntropyLoss=0.0672 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.282 train/patch_acc=0.6875 train/subpatch_loss=0.4091 train/subpatch_acc=0.7500 train/location_loss=1.628 train/location_acc=0.3750 train/token_losses=18.40 throughput/total_tokens=2,736,128 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:14:06.899 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=168/600, eta=2 hours, 57 minutes] train/CrossEntropyLoss=0.0142 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.163 train/patch_acc=0.7500 train/subpatch_loss=0.3293 train/subpatch_acc=0.8750 train/location_loss=1.592 train/location_acc=0.2500 train/token_losses=16.99 throughput/total_tokens=2,752,512 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:14:31.429 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=169/600, eta=2 hours, 57 minutes] train/CrossEntropyLoss=0.0060 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.139 train/patch_acc=0.6875 train/subpatch_loss=0.2333 train/subpatch_acc=0.9375 train/location_loss=1.667 train/location_acc=0.3125 train/token_losses=16.72 throughput/total_tokens=2,768,896 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:14:55.983 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=170/600, eta=2 hours, 57 minutes] train/CrossEntropyLoss=0.0703 train/ZLoss=0.0002 train/Accuracy=0.9787 train/patch_loss=1.083 train/patch_acc=0.5938 train/subpatch_loss=0.2817 train/subpatch_acc=0.9375 train/location_loss=1.695 train/location_acc=0.4375 train/token_losses=16.57 throughput/total_tokens=2,785,280 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:15:20.516 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=171/600, eta=2 hours, 56 minutes] train/CrossEntropyLoss=0.0451 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.149 train/patch_acc=0.5938 train/subpatch_loss=0.3964 train/subpatch_acc=0.8750 train/location_loss=1.592 train/location_acc=0.3750 train/token_losses=17.15 throughput/total_tokens=2,801,664 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:15:45.042 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=172/600, eta=2 hours, 56 minutes] train/CrossEntropyLoss=0.0082 train/ZLoss=0.0004 train/Accuracy=0.9939 train/patch_loss=1.276 train/patch_acc=0.5625 train/subpatch_loss=0.3910 train/subpatch_acc=0.8125 train/location_loss=1.544 train/location_acc=0.3750 train/token_losses=17.94 throughput/total_tokens=2,818,048 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:16:09.560 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=173/600, eta=2 hours, 55 minutes] train/CrossEntropyLoss=0.0532 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6955 train/patch_acc=0.8125 train/subpatch_loss=0.1646 train/subpatch_acc=0.9375 train/location_loss=1.452 train/location_acc=0.4375 train/token_losses=12.03 throughput/total_tokens=2,834,432 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:16:34.090 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=174/600, eta=2 hours, 55 minutes] train/CrossEntropyLoss=0.0329 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.090 train/patch_acc=0.7188 train/subpatch_loss=0.2271 train/subpatch_acc=0.9375 train/location_loss=1.771 train/location_acc=0.2500 train/token_losses=16.71 throughput/total_tokens=2,850,816 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:16:58.602 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=175/600, eta=2 hours, 55 minutes] train/CrossEntropyLoss=0.0124 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.091 train/patch_acc=0.6875 train/subpatch_loss=0.2746 train/subpatch_acc=1.0000 train/location_loss=1.659 train/location_acc=0.3750 train/token_losses=16.46 throughput/total_tokens=2,867,200 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:17:23.141 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=176/600, eta=2 hours, 54 minutes] train/CrossEntropyLoss=0.0240 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.338 train/patch_acc=0.6875 train/subpatch_loss=0.2365 train/subpatch_acc=0.8750 train/location_loss=1.468 train/location_acc=0.5000 train/token_losses=17.52 throughput/total_tokens=2,883,584 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:17:47.667 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=177/600, eta=2 hours, 54 minutes] train/CrossEntropyLoss=0.0295 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9045 train/patch_acc=0.6875 train/subpatch_loss=0.2576 train/subpatch_acc=0.9375 train/location_loss=1.500 train/location_acc=0.5000 train/token_losses=14.27 throughput/total_tokens=2,899,968 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:18:12.208 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=178/600, eta=2 hours, 53 minutes] train/CrossEntropyLoss=0.0234 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.113 train/patch_acc=0.7188 train/subpatch_loss=0.3337 train/subpatch_acc=0.8750 train/location_loss=1.221 train/location_acc=0.6250 train/token_losses=15.12 throughput/total_tokens=2,916,352 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:18:36.735 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=179/600, eta=2 hours, 53 minutes] train/CrossEntropyLoss=0.0345 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.193 train/patch_acc=0.6250 train/subpatch_loss=0.4259 train/subpatch_acc=0.7500 train/location_loss=1.797 train/location_acc=0.3125 train/token_losses=18.44 throughput/total_tokens=2,932,736 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:19:01.278 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=180/600, eta=2 hours, 52 minutes] train/CrossEntropyLoss=0.0233 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.132 train/patch_acc=0.7188 train/subpatch_loss=0.4378 train/subpatch_acc=0.8125 train/location_loss=1.980 train/location_acc=0.1875 train/token_losses=18.73 throughput/total_tokens=2,949,120 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:19:25.792 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=181/600, eta=2 hours, 52 minutes] train/CrossEntropyLoss=0.0150 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9730 train/patch_acc=0.7188 train/subpatch_loss=0.3283 train/subpatch_acc=0.9375 train/location_loss=1.438 train/location_acc=0.4375 train/token_losses=14.85 throughput/total_tokens=2,965,504 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:19:50.311 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=182/600, eta=2 hours, 52 minutes] train/CrossEntropyLoss=0.0263 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.335 train/patch_acc=0.6250 train/subpatch_loss=0.1778 train/subpatch_acc=1.0000 train/location_loss=1.782 train/location_acc=0.3125 train/token_losses=18.52 throughput/total_tokens=2,981,888 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:20:14.838 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=183/600, eta=2 hours, 51 minutes] train/CrossEntropyLoss=0.0169 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.426 train/patch_acc=0.6562 train/subpatch_loss=0.6100 train/subpatch_acc=0.8125 train/location_loss=1.278 train/location_acc=0.3750 train/token_losses=18.96 throughput/total_tokens=2,998,272 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:20:39.358 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=184/600, eta=2 hours, 51 minutes] train/CrossEntropyLoss=0.0600 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.185 train/patch_acc=0.6875 train/subpatch_loss=0.3371 train/subpatch_acc=0.9375 train/location_loss=1.671 train/location_acc=0.3750 train/token_losses=17.52 throughput/total_tokens=3,014,656 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:21:03.885 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=185/600, eta=2 hours, 50 minutes] train/CrossEntropyLoss=0.0675 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.302 train/patch_acc=0.6562 train/subpatch_loss=0.6097 train/subpatch_acc=0.6250 train/location_loss=1.661 train/location_acc=0.4375 train/token_losses=19.50 throughput/total_tokens=3,031,040 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:21:28.413 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=186/600, eta=2 hours, 50 minutes] train/CrossEntropyLoss=0.0247 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9202 train/patch_acc=0.7188 train/subpatch_loss=0.2108 train/subpatch_acc=0.9375 train/location_loss=1.261 train/location_acc=0.6250 train/token_losses=13.25 throughput/total_tokens=3,047,424 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:21:52.950 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=187/600, eta=2 hours, 50 minutes] train/CrossEntropyLoss=0.0259 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.9579 train/patch_acc=0.6875 train/subpatch_loss=0.3063 train/subpatch_acc=0.8750 train/location_loss=1.871 train/location_acc=0.1250 train/token_losses=16.37 throughput/total_tokens=3,063,808 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:22:17.463 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=188/600, eta=2 hours, 49 minutes] train/CrossEntropyLoss=0.0374 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.109 train/patch_acc=0.6875 train/subpatch_loss=0.2732 train/subpatch_acc=0.8750 train/location_loss=1.625 train/location_acc=0.3125 train/token_losses=16.46 throughput/total_tokens=3,080,192 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:22:42.005 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=189/600, eta=2 hours, 49 minutes] train/CrossEntropyLoss=0.0438 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9955 train/patch_acc=0.7188 train/subpatch_loss=0.2958 train/subpatch_acc=0.8750 train/location_loss=1.362 train/location_acc=0.6250 train/token_losses=14.60 throughput/total_tokens=3,096,576 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:23:06.527 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=190/600, eta=2 hours, 48 minutes] train/CrossEntropyLoss=0.0068 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.233 train/patch_acc=0.6250 train/subpatch_loss=0.3008 train/subpatch_acc=0.8750 train/location_loss=1.876 train/location_acc=0.3125 train/token_losses=18.58 throughput/total_tokens=3,112,960 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:23:31.109 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=191/600, eta=2 hours, 48 minutes] train/CrossEntropyLoss=0.0285 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.189 train/patch_acc=0.6562 train/subpatch_loss=0.6557 train/subpatch_acc=0.8125 train/location_loss=1.218 train/location_acc=0.6250 train/token_losses=17.01 throughput/total_tokens=3,129,344 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:23:55.663 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=192/600, eta=2 hours, 48 minutes] train/CrossEntropyLoss=0.0254 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8855 train/patch_acc=0.7812 train/subpatch_loss=0.2937 train/subpatch_acc=0.9375 train/location_loss=1.316 train/location_acc=0.5625 train/token_losses=13.52 throughput/total_tokens=3,145,728 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:24:20.186 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=193/600, eta=2 hours, 47 minutes] train/CrossEntropyLoss=0.0634 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9928 train/patch_acc=0.7500 train/subpatch_loss=0.3289 train/subpatch_acc=0.8750 train/location_loss=1.473 train/location_acc=0.3125 train/token_losses=15.15 throughput/total_tokens=3,162,112 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:24:44.698 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=194/600, eta=2 hours, 47 minutes] train/CrossEntropyLoss=0.0244 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.053 train/patch_acc=0.6562 train/subpatch_loss=0.4282 train/subpatch_acc=0.8125 train/location_loss=1.652 train/location_acc=0.3750 train/token_losses=16.74 throughput/total_tokens=3,178,496 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:25:09.224 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=195/600, eta=2 hours, 46 minutes] train/CrossEntropyLoss=0.1309 train/ZLoss=0.0005 train/Accuracy=0.9695 train/patch_loss=1.219 train/patch_acc=0.6562 train/subpatch_loss=0.3268 train/subpatch_acc=0.8750 train/location_loss=1.750 train/location_acc=0.3750 train/token_losses=18.06 throughput/total_tokens=3,194,880 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:25:33.777 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=196/600, eta=2 hours, 46 minutes] train/CrossEntropyLoss=0.0285 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.229 train/patch_acc=0.6562 train/subpatch_loss=0.2189 train/subpatch_acc=0.9375 train/location_loss=1.428 train/location_acc=0.3750 train/token_losses=16.42 throughput/total_tokens=3,211,264 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:25:58.309 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=197/600, eta=2 hours, 45 minutes] train/CrossEntropyLoss=0.0283 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.8187 train/patch_acc=0.8125 train/subpatch_loss=0.2523 train/subpatch_acc=0.9375 train/location_loss=1.410 train/location_acc=0.6875 train/token_losses=13.20 throughput/total_tokens=3,227,648 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:26:22.846 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=198/600, eta=2 hours, 45 minutes] train/CrossEntropyLoss=0.0526 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.059 train/patch_acc=0.6875 train/subpatch_loss=0.2388 train/subpatch_acc=1.0000 train/location_loss=1.254 train/location_acc=0.5625 train/token_losses=14.44 throughput/total_tokens=3,244,032 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:26:47.394 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=199/600, eta=2 hours, 45 minutes] train/CrossEntropyLoss=0.0289 train/ZLoss=0.0005 train/Accuracy=0.9909 train/patch_loss=1.007 train/patch_acc=0.7188 train/subpatch_loss=0.2643 train/subpatch_acc=0.9375 train/location_loss=1.346 train/location_acc=0.6250 train/token_losses=14.50 throughput/total_tokens=3,260,416 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:27:11.948 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=200/600, eta=2 hours, 44 minutes] train/CrossEntropyLoss=0.0600 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=1.159 train/patch_acc=0.6562 train/subpatch_loss=0.2863 train/subpatch_acc=0.8125 train/location_loss=1.838 train/location_acc=0.2500 train/token_losses=17.77 throughput/total_tokens=3,276,800 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:27:11.949 2012bc4f83ae:0 olmo.train.trainer:1415 INFO Saving checkpoint... 2026-07-27 13:27:14.021 2012bc4f83ae:0 olmo.train.trainer:1420 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step200 2026-07-27 13:27:38.825 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=201/600, eta=2 hours, 44 minutes] train/CrossEntropyLoss=0.0293 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.051 train/patch_acc=0.6875 train/subpatch_loss=0.3679 train/subpatch_acc=0.8125 train/location_loss=1.443 train/location_acc=0.3750 train/token_losses=15.65 throughput/total_tokens=3,293,184 throughput/device/tokens_per_second=660.6 throughput/device/batches_per_second=0.0403 2026-07-27 13:28:03.333 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=202/600, eta=2 hours, 44 minutes] train/CrossEntropyLoss=0.0306 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.054 train/patch_acc=0.7500 train/subpatch_loss=0.2129 train/subpatch_acc=0.8750 train/location_loss=1.518 train/location_acc=0.3125 train/token_losses=15.35 throughput/total_tokens=3,309,568 throughput/device/tokens_per_second=664.5 throughput/device/batches_per_second=0.0406 2026-07-27 13:28:27.845 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=203/600, eta=2 hours, 43 minutes] train/CrossEntropyLoss=0.0073 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8504 train/patch_acc=0.7812 train/subpatch_loss=0.4100 train/subpatch_acc=0.8125 train/location_loss=1.256 train/location_acc=0.6250 train/token_losses=13.47 throughput/total_tokens=3,325,952 throughput/device/tokens_per_second=665.8 throughput/device/batches_per_second=0.0406 2026-07-27 13:28:52.367 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=204/600, eta=2 hours, 43 minutes] train/CrossEntropyLoss=0.0616 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.9215 train/patch_acc=0.7500 train/subpatch_loss=0.3042 train/subpatch_acc=0.9375 train/location_loss=1.487 train/location_acc=0.5625 train/token_losses=14.54 throughput/total_tokens=3,342,336 throughput/device/tokens_per_second=666.4 throughput/device/batches_per_second=0.0407 2026-07-27 13:29:16.910 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=205/600, eta=2 hours, 42 minutes] train/CrossEntropyLoss=0.0230 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.298 train/patch_acc=0.6250 train/subpatch_loss=0.0624 train/subpatch_acc=1.0000 train/location_loss=1.345 train/location_acc=0.5000 train/token_losses=16.02 throughput/total_tokens=3,358,720 throughput/device/tokens_per_second=666.6 throughput/device/batches_per_second=0.0407 2026-07-27 13:29:41.457 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=206/600, eta=2 hours, 42 minutes] train/CrossEntropyLoss=0.0925 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.8903 train/patch_acc=0.7188 train/subpatch_loss=0.1460 train/subpatch_acc=1.0000 train/location_loss=1.410 train/location_acc=0.6250 train/token_losses=13.35 throughput/total_tokens=3,375,104 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 13:30:06.044 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=207/600, eta=2 hours, 41 minutes] train/CrossEntropyLoss=0.0411 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8739 train/patch_acc=0.7188 train/subpatch_loss=0.3182 train/subpatch_acc=0.9375 train/location_loss=1.522 train/location_acc=0.5000 train/token_losses=14.35 throughput/total_tokens=3,391,488 throughput/device/tokens_per_second=666.7 throughput/device/batches_per_second=0.0407 2026-07-27 13:30:30.612 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=208/600, eta=2 hours, 41 minutes] train/CrossEntropyLoss=0.0214 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.8943 train/patch_acc=0.7500 train/subpatch_loss=0.2751 train/subpatch_acc=0.8750 train/location_loss=1.418 train/location_acc=0.4375 train/token_losses=13.93 throughput/total_tokens=3,407,872 throughput/device/tokens_per_second=666.7 throughput/device/batches_per_second=0.0407 2026-07-27 13:30:55.163 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=209/600, eta=2 hours, 41 minutes] train/CrossEntropyLoss=0.0188 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.078 train/patch_acc=0.6562 train/subpatch_loss=0.2879 train/subpatch_acc=0.9375 train/location_loss=1.392 train/location_acc=0.5625 train/token_losses=15.34 throughput/total_tokens=3,424,256 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 13:31:19.711 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=210/600, eta=2 hours, 40 minutes] train/CrossEntropyLoss=0.0155 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.275 train/patch_acc=0.5312 train/subpatch_loss=0.3120 train/subpatch_acc=0.8750 train/location_loss=1.386 train/location_acc=0.5000 train/token_losses=17.00 throughput/total_tokens=3,440,640 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:31:44.270 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=211/600, eta=2 hours, 40 minutes] train/CrossEntropyLoss=0.0205 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.103 train/patch_acc=0.7500 train/subpatch_loss=0.2393 train/subpatch_acc=0.9375 train/location_loss=1.606 train/location_acc=0.5000 train/token_losses=16.21 throughput/total_tokens=3,457,024 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 13:32:08.832 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=212/600, eta=2 hours, 39 minutes] train/CrossEntropyLoss=0.0853 train/ZLoss=0.0007 train/Accuracy=0.9756 train/patch_loss=0.8682 train/patch_acc=0.7812 train/subpatch_loss=0.1177 train/subpatch_acc=1.0000 train/location_loss=1.465 train/location_acc=0.6250 train/token_losses=13.28 throughput/total_tokens=3,473,408 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 13:32:33.365 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=213/600, eta=2 hours, 39 minutes] train/CrossEntropyLoss=0.0334 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.273 train/patch_acc=0.6875 train/subpatch_loss=0.1824 train/subpatch_acc=0.9375 train/location_loss=1.257 train/location_acc=0.5625 train/token_losses=15.94 throughput/total_tokens=3,489,792 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 13:32:57.926 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=214/600, eta=2 hours, 39 minutes] train/CrossEntropyLoss=0.1748 train/ZLoss=0.0008 train/Accuracy=0.9573 train/patch_loss=1.301 train/patch_acc=0.6875 train/subpatch_loss=0.5103 train/subpatch_acc=0.7500 train/location_loss=1.428 train/location_acc=0.4375 train/token_losses=18.16 throughput/total_tokens=3,506,176 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 13:33:22.468 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=215/600, eta=2 hours, 38 minutes] train/CrossEntropyLoss=0.0172 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9626 train/patch_acc=0.7188 train/subpatch_loss=0.2434 train/subpatch_acc=0.9375 train/location_loss=1.455 train/location_acc=0.3750 train/token_losses=14.49 throughput/total_tokens=3,522,560 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 13:33:47.011 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=216/600, eta=2 hours, 38 minutes] train/CrossEntropyLoss=0.0399 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.217 train/patch_acc=0.6562 train/subpatch_loss=0.2705 train/subpatch_acc=0.8750 train/location_loss=1.464 train/location_acc=0.5000 train/token_losses=16.68 throughput/total_tokens=3,538,944 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 13:34:11.551 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=217/600, eta=2 hours, 37 minutes] train/CrossEntropyLoss=0.0432 train/ZLoss=0.0008 train/Accuracy=0.9878 train/patch_loss=0.9462 train/patch_acc=0.7188 train/subpatch_loss=0.2386 train/subpatch_acc=0.8750 train/location_loss=1.495 train/location_acc=0.2500 train/token_losses=14.50 throughput/total_tokens=3,555,328 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 13:34:36.078 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=218/600, eta=2 hours, 37 minutes] train/CrossEntropyLoss=0.0701 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.220 train/patch_acc=0.7188 train/subpatch_loss=0.1752 train/subpatch_acc=1.0000 train/location_loss=1.483 train/location_acc=0.5000 train/token_losses=16.39 throughput/total_tokens=3,571,712 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 13:35:00.625 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=219/600, eta=2 hours, 37 minutes] train/CrossEntropyLoss=0.1799 train/ZLoss=0.0006 train/Accuracy=0.9512 train/patch_loss=1.607 train/patch_acc=0.5938 train/subpatch_loss=0.4187 train/subpatch_acc=0.9375 train/location_loss=1.806 train/location_acc=0.2500 train/token_losses=21.75 throughput/total_tokens=3,588,096 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 13:35:25.146 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=220/600, eta=2 hours, 36 minutes] train/CrossEntropyLoss=0.0537 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7372 train/patch_acc=0.8125 train/subpatch_loss=0.3014 train/subpatch_acc=0.9375 train/location_loss=1.344 train/location_acc=0.4375 train/token_losses=12.48 throughput/total_tokens=3,604,480 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:35:49.675 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=221/600, eta=2 hours, 36 minutes] train/CrossEntropyLoss=0.0350 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.025 train/patch_acc=0.6875 train/subpatch_loss=0.2243 train/subpatch_acc=1.0000 train/location_loss=1.018 train/location_acc=0.6875 train/token_losses=13.17 throughput/total_tokens=3,620,864 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 13:36:14.204 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=222/600, eta=2 hours, 35 minutes] train/CrossEntropyLoss=0.0978 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.8802 train/patch_acc=0.7812 train/subpatch_loss=0.1338 train/subpatch_acc=1.0000 train/location_loss=1.189 train/location_acc=0.3750 train/token_losses=12.33 throughput/total_tokens=3,637,248 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 13:36:38.754 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=223/600, eta=2 hours, 35 minutes] train/CrossEntropyLoss=0.0557 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.046 train/patch_acc=0.7500 train/subpatch_loss=0.2319 train/subpatch_acc=0.9375 train/location_loss=1.626 train/location_acc=0.4375 train/token_losses=15.80 throughput/total_tokens=3,653,632 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 13:37:03.264 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=224/600, eta=2 hours, 34 minutes] train/CrossEntropyLoss=0.0472 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.042 train/patch_acc=0.6250 train/subpatch_loss=0.2576 train/subpatch_acc=0.9375 train/location_loss=1.410 train/location_acc=0.4375 train/token_losses=15.00 throughput/total_tokens=3,670,016 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 13:37:27.795 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=225/600, eta=2 hours, 34 minutes] train/CrossEntropyLoss=0.0534 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.149 train/patch_acc=0.6250 train/subpatch_loss=0.2417 train/subpatch_acc=0.9375 train/location_loss=1.706 train/location_acc=0.3125 train/token_losses=16.98 throughput/total_tokens=3,686,400 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 13:37:52.329 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=226/600, eta=2 hours, 34 minutes] train/CrossEntropyLoss=0.0523 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.309 train/patch_acc=0.5938 train/subpatch_loss=0.2771 train/subpatch_acc=0.9375 train/location_loss=1.254 train/location_acc=0.5000 train/token_losses=16.59 throughput/total_tokens=3,702,784 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 13:38:16.830 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=227/600, eta=2 hours, 33 minutes] train/CrossEntropyLoss=0.0206 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.9175 train/patch_acc=0.7812 train/subpatch_loss=0.2195 train/subpatch_acc=0.9375 train/location_loss=1.121 train/location_acc=0.5625 train/token_losses=12.70 throughput/total_tokens=3,719,168 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:38:41.343 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=228/600, eta=2 hours, 33 minutes] train/CrossEntropyLoss=0.1867 train/ZLoss=0.0007 train/Accuracy=0.9573 train/patch_loss=1.037 train/patch_acc=0.7188 train/subpatch_loss=0.2476 train/subpatch_acc=0.9375 train/location_loss=1.236 train/location_acc=0.6875 train/token_losses=14.23 throughput/total_tokens=3,735,552 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:39:05.891 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=229/600, eta=2 hours, 32 minutes] train/CrossEntropyLoss=0.0551 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.5599 train/patch_acc=0.8750 train/subpatch_loss=0.1258 train/subpatch_acc=1.0000 train/location_loss=1.237 train/location_acc=0.4375 train/token_losses=9.929 throughput/total_tokens=3,751,936 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:39:30.417 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=230/600, eta=2 hours, 32 minutes] train/CrossEntropyLoss=0.0035 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.8505 train/patch_acc=0.7500 train/subpatch_loss=0.1754 train/subpatch_acc=0.8750 train/location_loss=1.046 train/location_acc=0.7500 train/token_losses=11.69 throughput/total_tokens=3,768,320 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:39:54.960 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=231/600, eta=2 hours, 32 minutes] train/CrossEntropyLoss=0.0323 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.8545 train/patch_acc=0.7812 train/subpatch_loss=0.3340 train/subpatch_acc=0.8750 train/location_loss=1.460 train/location_acc=0.6250 train/token_losses=14.01 throughput/total_tokens=3,784,704 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:40:19.488 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=232/600, eta=2 hours, 31 minutes] train/CrossEntropyLoss=0.0479 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.033 train/patch_acc=0.7188 train/subpatch_loss=0.1189 train/subpatch_acc=0.9375 train/location_loss=1.682 train/location_acc=0.3125 train/token_losses=15.47 throughput/total_tokens=3,801,088 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:40:43.992 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=233/600, eta=2 hours, 31 minutes] train/CrossEntropyLoss=0.0547 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.8717 train/patch_acc=0.7500 train/subpatch_loss=0.2355 train/subpatch_acc=0.9375 train/location_loss=1.269 train/location_acc=0.6875 train/token_losses=12.99 throughput/total_tokens=3,817,472 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:41:08.517 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=234/600, eta=2 hours, 30 minutes] train/CrossEntropyLoss=0.0361 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.9063 train/patch_acc=0.7500 train/subpatch_loss=0.0744 train/subpatch_acc=1.0000 train/location_loss=1.626 train/location_acc=0.3125 train/token_losses=14.05 throughput/total_tokens=3,833,856 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:41:33.040 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=235/600, eta=2 hours, 30 minutes] train/CrossEntropyLoss=0.0025 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=1.220 train/patch_acc=0.6875 train/subpatch_loss=0.1315 train/subpatch_acc=1.0000 train/location_loss=1.569 train/location_acc=0.4375 train/token_losses=16.56 throughput/total_tokens=3,850,240 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:41:57.567 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=236/600, eta=2 hours, 30 minutes] train/CrossEntropyLoss=0.0628 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8635 train/patch_acc=0.7500 train/subpatch_loss=0.1229 train/subpatch_acc=1.0000 train/location_loss=1.443 train/location_acc=0.4375 train/token_losses=13.17 throughput/total_tokens=3,866,624 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:42:22.114 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=237/600, eta=2 hours, 29 minutes] train/CrossEntropyLoss=0.0222 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.098 train/patch_acc=0.7500 train/subpatch_loss=0.2108 train/subpatch_acc=1.0000 train/location_loss=1.544 train/location_acc=0.4375 train/token_losses=15.80 throughput/total_tokens=3,883,008 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:42:46.651 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=238/600, eta=2 hours, 29 minutes] train/CrossEntropyLoss=0.0095 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8421 train/patch_acc=0.7812 train/subpatch_loss=0.2148 train/subpatch_acc=0.9375 train/location_loss=1.535 train/location_acc=0.2500 train/token_losses=13.74 throughput/total_tokens=3,899,392 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:43:11.211 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=239/600, eta=2 hours, 28 minutes] train/CrossEntropyLoss=0.3109 train/ZLoss=0.0011 train/Accuracy=0.9177 train/patch_loss=1.624 train/patch_acc=0.5312 train/subpatch_loss=0.3211 train/subpatch_acc=0.8750 train/location_loss=1.556 train/location_acc=0.4375 train/token_losses=20.50 throughput/total_tokens=3,915,776 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:43:35.734 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=240/600, eta=2 hours, 28 minutes] train/CrossEntropyLoss=0.0193 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9187 train/patch_acc=0.8125 train/subpatch_loss=0.1558 train/subpatch_acc=1.0000 train/location_loss=1.304 train/location_acc=0.4375 train/token_losses=13.19 throughput/total_tokens=3,932,160 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:44:00.252 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=241/600, eta=2 hours, 28 minutes] train/CrossEntropyLoss=0.0043 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.9915 train/patch_acc=0.6875 train/subpatch_loss=0.2114 train/subpatch_acc=0.9375 train/location_loss=1.313 train/location_acc=0.5000 train/token_losses=14.03 throughput/total_tokens=3,948,544 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:44:24.793 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=242/600, eta=2 hours, 27 minutes] train/CrossEntropyLoss=0.0011 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.8858 train/patch_acc=0.7812 train/subpatch_loss=0.2204 train/subpatch_acc=0.9375 train/location_loss=1.245 train/location_acc=0.6250 train/token_losses=12.95 throughput/total_tokens=3,964,928 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:44:49.321 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=243/600, eta=2 hours, 27 minutes] train/CrossEntropyLoss=0.0215 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9519 train/patch_acc=0.7188 train/subpatch_loss=0.1854 train/subpatch_acc=0.9375 train/location_loss=0.8762 train/location_acc=0.8125 train/token_losses=11.86 throughput/total_tokens=3,981,312 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 13:45:13.869 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=244/600, eta=2 hours, 26 minutes] train/CrossEntropyLoss=0.1967 train/ZLoss=0.0007 train/Accuracy=0.9573 train/patch_loss=1.119 train/patch_acc=0.6875 train/subpatch_loss=0.2741 train/subpatch_acc=0.8750 train/location_loss=1.164 train/location_acc=0.5625 train/token_losses=14.70 throughput/total_tokens=3,997,696 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:45:38.386 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=245/600, eta=2 hours, 26 minutes] train/CrossEntropyLoss=0.0808 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=1.004 train/patch_acc=0.7812 train/subpatch_loss=0.2528 train/subpatch_acc=0.8750 train/location_loss=1.253 train/location_acc=0.5625 train/token_losses=14.06 throughput/total_tokens=4,014,080 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:46:02.943 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=246/600, eta=2 hours, 25 minutes] train/CrossEntropyLoss=0.0040 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.8777 train/patch_acc=0.7812 train/subpatch_loss=0.0873 train/subpatch_acc=1.0000 train/location_loss=1.353 train/location_acc=0.5000 train/token_losses=12.78 throughput/total_tokens=4,030,464 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:46:27.483 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=247/600, eta=2 hours, 25 minutes] train/CrossEntropyLoss=0.0429 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.8240 train/patch_acc=0.7812 train/subpatch_loss=0.3038 train/subpatch_acc=0.8750 train/location_loss=1.192 train/location_acc=0.6250 train/token_losses=12.57 throughput/total_tokens=4,046,848 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:46:52.034 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=248/600, eta=2 hours, 25 minutes] train/CrossEntropyLoss=0.0131 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.345 train/patch_acc=0.6875 train/subpatch_loss=0.5328 train/subpatch_acc=0.8125 train/location_loss=1.558 train/location_acc=0.2500 train/token_losses=19.12 throughput/total_tokens=4,063,232 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:47:16.564 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=249/600, eta=2 hours, 24 minutes] train/CrossEntropyLoss=0.0241 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7069 train/patch_acc=0.8125 train/subpatch_loss=0.2217 train/subpatch_acc=0.9375 train/location_loss=1.622 train/location_acc=0.5000 train/token_losses=13.03 throughput/total_tokens=4,079,616 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:47:41.106 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=250/600, eta=2 hours, 24 minutes] train/CrossEntropyLoss=0.2137 train/ZLoss=0.0005 train/Accuracy=0.9543 train/patch_loss=1.249 train/patch_acc=0.6875 train/subpatch_loss=0.3959 train/subpatch_acc=0.8750 train/location_loss=1.353 train/location_acc=0.6250 train/token_losses=16.99 throughput/total_tokens=4,096,000 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:48:05.634 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=251/600, eta=2 hours, 23 minutes] train/CrossEntropyLoss=0.0515 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7189 train/patch_acc=0.9062 train/subpatch_loss=0.2479 train/subpatch_acc=0.8750 train/location_loss=1.613 train/location_acc=0.4375 train/token_losses=13.20 throughput/total_tokens=4,112,384 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:48:30.161 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=252/600, eta=2 hours, 23 minutes] train/CrossEntropyLoss=0.1297 train/ZLoss=0.0008 train/Accuracy=0.9634 train/patch_loss=0.7759 train/patch_acc=0.7812 train/subpatch_loss=0.1923 train/subpatch_acc=0.8750 train/location_loss=1.050 train/location_acc=0.7500 train/token_losses=11.18 throughput/total_tokens=4,128,768 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:48:54.711 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=253/600, eta=2 hours, 23 minutes] train/CrossEntropyLoss=0.0535 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.9858 train/patch_acc=0.7812 train/subpatch_loss=0.2582 train/subpatch_acc=0.8750 train/location_loss=1.177 train/location_acc=0.5625 train/token_losses=13.63 throughput/total_tokens=4,145,152 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:49:19.246 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=254/600, eta=2 hours, 22 minutes] train/CrossEntropyLoss=0.0135 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.034 train/patch_acc=0.7188 train/subpatch_loss=0.2770 train/subpatch_acc=0.9375 train/location_loss=1.378 train/location_acc=0.3750 train/token_losses=14.89 throughput/total_tokens=4,161,536 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 13:49:43.753 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=255/600, eta=2 hours, 22 minutes] train/CrossEntropyLoss=0.0803 train/ZLoss=0.0008 train/Accuracy=0.9787 train/patch_loss=0.9407 train/patch_acc=0.7500 train/subpatch_loss=0.1897 train/subpatch_acc=0.9375 train/location_loss=1.147 train/location_acc=0.6250 train/token_losses=12.87 throughput/total_tokens=4,177,920 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:50:08.264 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=256/600, eta=2 hours, 21 minutes] train/CrossEntropyLoss=0.0515 train/ZLoss=0.0004 train/Accuracy=0.9878 train/patch_loss=0.6302 train/patch_acc=0.8125 train/subpatch_loss=0.0882 train/subpatch_acc=1.0000 train/location_loss=1.074 train/location_acc=0.6875 train/token_losses=9.691 throughput/total_tokens=4,194,304 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:50:32.811 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=257/600, eta=2 hours, 21 minutes] train/CrossEntropyLoss=0.0159 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7382 train/patch_acc=0.8125 train/subpatch_loss=0.3982 train/subpatch_acc=0.8750 train/location_loss=1.150 train/location_acc=0.5000 train/token_losses=12.10 throughput/total_tokens=4,210,688 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:50:57.356 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=258/600, eta=2 hours, 21 minutes] train/CrossEntropyLoss=0.1636 train/ZLoss=0.0007 train/Accuracy=0.9573 train/patch_loss=1.554 train/patch_acc=0.5625 train/subpatch_loss=0.4832 train/subpatch_acc=0.8750 train/location_loss=1.421 train/location_acc=0.6250 train/token_losses=20.05 throughput/total_tokens=4,227,072 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:51:21.882 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=259/600, eta=2 hours, 20 minutes] train/CrossEntropyLoss=0.0355 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.8291 train/patch_acc=0.7188 train/subpatch_loss=0.1790 train/subpatch_acc=1.0000 train/location_loss=1.386 train/location_acc=0.5625 train/token_losses=12.89 throughput/total_tokens=4,243,456 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:51:46.398 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=260/600, eta=2 hours, 20 minutes] train/CrossEntropyLoss=0.0893 train/ZLoss=0.0006 train/Accuracy=0.9787 train/patch_loss=1.039 train/patch_acc=0.6250 train/subpatch_loss=0.1693 train/subpatch_acc=1.0000 train/location_loss=1.249 train/location_acc=0.6875 train/token_losses=13.98 throughput/total_tokens=4,259,840 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:52:10.935 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=261/600, eta=2 hours, 19 minutes] train/CrossEntropyLoss=0.0612 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.090 train/patch_acc=0.6562 train/subpatch_loss=0.2390 train/subpatch_acc=0.9375 train/location_loss=1.391 train/location_acc=0.5000 train/token_losses=15.24 throughput/total_tokens=4,276,224 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:52:35.477 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=262/600, eta=2 hours, 19 minutes] train/CrossEntropyLoss=0.0375 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8209 train/patch_acc=0.8125 train/subpatch_loss=0.3192 train/subpatch_acc=0.9375 train/location_loss=1.142 train/location_acc=0.5625 train/token_losses=12.41 throughput/total_tokens=4,292,608 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:53:00.022 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=263/600, eta=2 hours, 18 minutes] train/CrossEntropyLoss=0.0042 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=1.010 train/patch_acc=0.7500 train/subpatch_loss=0.1934 train/subpatch_acc=0.9375 train/location_loss=1.286 train/location_acc=0.6250 train/token_losses=13.99 throughput/total_tokens=4,308,992 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:53:24.533 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=264/600, eta=2 hours, 18 minutes] train/CrossEntropyLoss=0.0446 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.082 train/patch_acc=0.7188 train/subpatch_loss=0.2024 train/subpatch_acc=1.0000 train/location_loss=1.399 train/location_acc=0.3125 train/token_losses=15.06 throughput/total_tokens=4,325,376 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:53:49.071 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=265/600, eta=2 hours, 18 minutes] train/CrossEntropyLoss=0.0089 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.150 train/patch_acc=0.6875 train/subpatch_loss=0.0925 train/subpatch_acc=1.0000 train/location_loss=1.067 train/location_acc=0.7500 train/token_losses=13.84 throughput/total_tokens=4,341,760 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:54:13.580 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=266/600, eta=2 hours, 17 minutes] train/CrossEntropyLoss=0.0018 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=1.131 train/patch_acc=0.6250 train/subpatch_loss=0.1726 train/subpatch_acc=0.9375 train/location_loss=1.333 train/location_acc=0.5625 train/token_losses=15.07 throughput/total_tokens=4,358,144 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:54:38.118 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=267/600, eta=2 hours, 17 minutes] train/CrossEntropyLoss=0.0014 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.5914 train/patch_acc=0.8125 train/subpatch_loss=0.2046 train/subpatch_acc=0.9375 train/location_loss=1.494 train/location_acc=0.4375 train/token_losses=11.53 throughput/total_tokens=4,374,528 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:55:02.665 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=268/600, eta=2 hours, 16 minutes] train/CrossEntropyLoss=0.0215 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7138 train/patch_acc=0.8438 train/subpatch_loss=0.1758 train/subpatch_acc=1.0000 train/location_loss=1.276 train/location_acc=0.4375 train/token_losses=11.52 throughput/total_tokens=4,390,912 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:55:27.210 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=269/600, eta=2 hours, 16 minutes] train/CrossEntropyLoss=0.0554 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.9182 train/patch_acc=0.7500 train/subpatch_loss=0.2269 train/subpatch_acc=0.9375 train/location_loss=1.532 train/location_acc=0.5000 train/token_losses=14.38 throughput/total_tokens=4,407,296 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:55:51.754 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=270/600, eta=2 hours, 16 minutes] train/CrossEntropyLoss=0.1530 train/ZLoss=0.0006 train/Accuracy=0.9665 train/patch_loss=0.9809 train/patch_acc=0.7812 train/subpatch_loss=0.1028 train/subpatch_acc=1.0000 train/location_loss=1.162 train/location_acc=0.6250 train/token_losses=12.91 throughput/total_tokens=4,423,680 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 13:56:16.293 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=271/600, eta=2 hours, 15 minutes] train/CrossEntropyLoss=0.0847 train/ZLoss=0.0004 train/Accuracy=0.9726 train/patch_loss=0.6748 train/patch_acc=0.8125 train/subpatch_loss=0.1608 train/subpatch_acc=0.9375 train/location_loss=1.132 train/location_acc=0.6250 train/token_losses=10.57 throughput/total_tokens=4,440,064 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:56:40.823 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=272/600, eta=2 hours, 15 minutes] train/CrossEntropyLoss=0.0452 train/ZLoss=0.0001 train/Accuracy=0.9787 train/patch_loss=0.8922 train/patch_acc=0.7500 train/subpatch_loss=0.1380 train/subpatch_acc=1.0000 train/location_loss=1.225 train/location_acc=0.5625 train/token_losses=12.59 throughput/total_tokens=4,456,448 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:57:05.345 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=273/600, eta=2 hours, 14 minutes] train/CrossEntropyLoss=0.0259 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.9467 train/patch_acc=0.7188 train/subpatch_loss=0.0916 train/subpatch_acc=1.0000 train/location_loss=1.033 train/location_acc=0.6250 train/token_losses=12.07 throughput/total_tokens=4,472,832 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:57:29.883 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=274/600, eta=2 hours, 14 minutes] train/CrossEntropyLoss=0.0201 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9665 train/patch_acc=0.7500 train/subpatch_loss=0.1336 train/subpatch_acc=1.0000 train/location_loss=1.290 train/location_acc=0.4375 train/token_losses=13.43 throughput/total_tokens=4,489,216 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:57:54.406 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=275/600, eta=2 hours, 14 minutes] train/CrossEntropyLoss=0.0213 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.241 train/patch_acc=0.6562 train/subpatch_loss=0.1724 train/subpatch_acc=0.9375 train/location_loss=1.397 train/location_acc=0.5000 train/token_losses=16.21 throughput/total_tokens=4,505,600 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:58:18.925 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=276/600, eta=2 hours, 13 minutes] train/CrossEntropyLoss=0.0753 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9129 train/patch_acc=0.7500 train/subpatch_loss=0.3990 train/subpatch_acc=0.8125 train/location_loss=1.445 train/location_acc=0.5000 train/token_losses=14.68 throughput/total_tokens=4,521,984 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:58:43.466 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=277/600, eta=2 hours, 13 minutes] train/CrossEntropyLoss=0.0272 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9636 train/patch_acc=0.6562 train/subpatch_loss=0.2264 train/subpatch_acc=0.9375 train/location_loss=1.244 train/location_acc=0.5000 train/token_losses=13.59 throughput/total_tokens=4,538,368 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 13:59:08.019 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=278/600, eta=2 hours, 12 minutes] train/CrossEntropyLoss=0.0171 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.8862 train/patch_acc=0.8125 train/subpatch_loss=0.1262 train/subpatch_acc=1.0000 train/location_loss=1.399 train/location_acc=0.5000 train/token_losses=13.19 throughput/total_tokens=4,554,752 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:59:32.548 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=279/600, eta=2 hours, 12 minutes] train/CrossEntropyLoss=0.0354 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9921 train/patch_acc=0.7500 train/subpatch_loss=0.1984 train/subpatch_acc=0.9375 train/location_loss=1.725 train/location_acc=0.2500 train/token_losses=15.63 throughput/total_tokens=4,571,136 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 13:59:57.062 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=280/600, eta=2 hours, 12 minutes] train/CrossEntropyLoss=0.0543 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.6057 train/patch_acc=0.8125 train/subpatch_loss=0.2687 train/subpatch_acc=0.9375 train/location_loss=1.150 train/location_acc=0.5625 train/token_losses=10.52 throughput/total_tokens=4,587,520 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:00:21.605 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=281/600, eta=2 hours, 11 minutes] train/CrossEntropyLoss=0.1121 train/ZLoss=0.0002 train/Accuracy=0.9756 train/patch_loss=0.6370 train/patch_acc=0.8125 train/subpatch_loss=0.1307 train/subpatch_acc=0.9375 train/location_loss=1.144 train/location_acc=0.5625 train/token_losses=10.19 throughput/total_tokens=4,603,904 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:00:46.163 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=282/600, eta=2 hours, 11 minutes] train/CrossEntropyLoss=0.0452 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.069 train/patch_acc=0.6562 train/subpatch_loss=0.3253 train/subpatch_acc=0.9375 train/location_loss=1.467 train/location_acc=0.5000 train/token_losses=15.72 throughput/total_tokens=4,620,288 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:01:10.712 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=283/600, eta=2 hours, 10 minutes] train/CrossEntropyLoss=0.0306 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9368 train/patch_acc=0.7812 train/subpatch_loss=0.1957 train/subpatch_acc=0.8750 train/location_loss=1.143 train/location_acc=0.6250 train/token_losses=12.85 throughput/total_tokens=4,636,672 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:01:35.247 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=284/600, eta=2 hours, 10 minutes] train/CrossEntropyLoss=0.1712 train/ZLoss=0.0007 train/Accuracy=0.9573 train/patch_loss=1.285 train/patch_acc=0.6562 train/subpatch_loss=0.1342 train/subpatch_acc=1.0000 train/location_loss=1.081 train/location_acc=0.5625 train/token_losses=15.14 throughput/total_tokens=4,653,056 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:01:59.803 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=285/600, eta=2 hours, 9 minutes] train/CrossEntropyLoss=0.0157 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.9416 train/patch_acc=0.7812 train/subpatch_loss=0.1288 train/subpatch_acc=1.0000 train/location_loss=1.321 train/location_acc=0.3750 train/token_losses=13.33 throughput/total_tokens=4,669,440 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 14:02:24.340 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=286/600, eta=2 hours, 9 minutes] train/CrossEntropyLoss=0.0110 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.7429 train/patch_acc=0.7812 train/subpatch_loss=0.1221 train/subpatch_acc=1.0000 train/location_loss=1.037 train/location_acc=0.5000 train/token_losses=10.58 throughput/total_tokens=4,685,824 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 14:02:48.893 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=287/600, eta=2 hours, 9 minutes] train/CrossEntropyLoss=0.0421 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.9527 train/patch_acc=0.7812 train/subpatch_loss=0.1770 train/subpatch_acc=1.0000 train/location_loss=1.216 train/location_acc=0.6875 train/token_losses=13.19 throughput/total_tokens=4,702,208 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 14:03:13.433 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=288/600, eta=2 hours, 8 minutes] train/CrossEntropyLoss=0.1889 train/ZLoss=0.0006 train/Accuracy=0.9543 train/patch_loss=1.077 train/patch_acc=0.7500 train/subpatch_loss=0.1086 train/subpatch_acc=1.0000 train/location_loss=1.142 train/location_acc=0.6875 train/token_losses=13.61 throughput/total_tokens=4,718,592 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 14:03:37.950 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=289/600, eta=2 hours, 8 minutes] train/CrossEntropyLoss=0.0659 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.4705 train/patch_acc=0.9375 train/subpatch_loss=0.1327 train/subpatch_acc=1.0000 train/location_loss=0.7408 train/location_acc=0.8750 train/token_losses=7.258 throughput/total_tokens=4,734,976 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 14:04:02.473 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=290/600, eta=2 hours, 7 minutes] train/CrossEntropyLoss=0.0125 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.9708 train/patch_acc=0.7188 train/subpatch_loss=0.0878 train/subpatch_acc=1.0000 train/location_loss=1.168 train/location_acc=0.6250 train/token_losses=12.79 throughput/total_tokens=4,751,360 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:04:26.986 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=291/600, eta=2 hours, 7 minutes] train/CrossEntropyLoss=0.0445 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9665 train/patch_acc=0.7500 train/subpatch_loss=0.2025 train/subpatch_acc=0.9375 train/location_loss=1.010 train/location_acc=0.7500 train/token_losses=12.58 throughput/total_tokens=4,767,744 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:04:51.527 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=292/600, eta=2 hours, 7 minutes] train/CrossEntropyLoss=0.1540 train/ZLoss=0.0007 train/Accuracy=0.9665 train/patch_loss=1.121 train/patch_acc=0.7500 train/subpatch_loss=0.4523 train/subpatch_acc=0.6875 train/location_loss=1.193 train/location_acc=0.6250 train/token_losses=15.55 throughput/total_tokens=4,784,128 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:05:16.061 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=293/600, eta=2 hours, 6 minutes] train/CrossEntropyLoss=0.0396 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.199 train/patch_acc=0.7500 train/subpatch_loss=0.1894 train/subpatch_acc=1.0000 train/location_loss=1.013 train/location_acc=0.6875 train/token_losses=14.41 throughput/total_tokens=4,800,512 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:05:40.596 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=294/600, eta=2 hours, 6 minutes] train/CrossEntropyLoss=0.0357 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.132 train/patch_acc=0.7188 train/subpatch_loss=0.1881 train/subpatch_acc=1.0000 train/location_loss=1.282 train/location_acc=0.6250 train/token_losses=14.94 throughput/total_tokens=4,816,896 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:06:05.113 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=295/600, eta=2 hours, 5 minutes] train/CrossEntropyLoss=0.0176 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8736 train/patch_acc=0.7812 train/subpatch_loss=0.2572 train/subpatch_acc=0.9375 train/location_loss=1.279 train/location_acc=0.4375 train/token_losses=13.13 throughput/total_tokens=4,833,280 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:06:29.637 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=296/600, eta=2 hours, 5 minutes] train/CrossEntropyLoss=0.0205 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.8626 train/patch_acc=0.7188 train/subpatch_loss=0.0957 train/subpatch_acc=1.0000 train/location_loss=0.9775 train/location_acc=0.6875 train/token_losses=11.19 throughput/total_tokens=4,849,664 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:06:54.179 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=297/600, eta=2 hours, 5 minutes] train/CrossEntropyLoss=0.0339 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7433 train/patch_acc=0.7812 train/subpatch_loss=0.2801 train/subpatch_acc=0.8750 train/location_loss=1.049 train/location_acc=0.6875 train/token_losses=11.26 throughput/total_tokens=4,866,048 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:07:18.715 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=298/600, eta=2 hours, 4 minutes] train/CrossEntropyLoss=0.0331 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.294 train/patch_acc=0.6875 train/subpatch_loss=0.1719 train/subpatch_acc=1.0000 train/location_loss=1.285 train/location_acc=0.4375 train/token_losses=16.18 throughput/total_tokens=4,882,432 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:07:43.250 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=299/600, eta=2 hours, 4 minutes] train/CrossEntropyLoss=0.0247 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.000 train/patch_acc=0.7500 train/subpatch_loss=0.1110 train/subpatch_acc=1.0000 train/location_loss=0.8691 train/location_acc=0.6875 train/token_losses=11.92 throughput/total_tokens=4,898,816 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:08:07.783 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=300/600, eta=2 hours, 3 minutes] train/CrossEntropyLoss=0.0511 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.8662 train/patch_acc=0.8125 train/subpatch_loss=0.0519 train/subpatch_acc=1.0000 train/location_loss=1.091 train/location_acc=0.6875 train/token_losses=11.50 throughput/total_tokens=4,915,200 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:08:07.785 2012bc4f83ae:0 olmo.train.trainer:1415 INFO Saving checkpoint... 2026-07-27 14:08:09.925 2012bc4f83ae:0 olmo.train.trainer:1420 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step300 2026-07-27 14:08:35.097 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=301/600, eta=2 hours, 3 minutes] train/CrossEntropyLoss=0.1504 train/ZLoss=0.0007 train/Accuracy=0.9543 train/patch_loss=0.6194 train/patch_acc=0.8125 train/subpatch_loss=0.1394 train/subpatch_acc=1.0000 train/location_loss=1.380 train/location_acc=0.5000 train/token_losses=11.03 throughput/total_tokens=4,931,584 throughput/device/tokens_per_second=651.0 throughput/device/batches_per_second=0.0397 2026-07-27 14:08:59.610 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=302/600, eta=2 hours, 3 minutes] train/CrossEntropyLoss=0.0087 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.7440 train/patch_acc=0.8750 train/subpatch_loss=0.1336 train/subpatch_acc=1.0000 train/location_loss=1.096 train/location_acc=0.5000 train/token_losses=10.87 throughput/total_tokens=4,947,968 throughput/device/tokens_per_second=659.5 throughput/device/batches_per_second=0.0403 2026-07-27 14:09:24.135 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=303/600, eta=2 hours, 2 minutes] train/CrossEntropyLoss=0.0422 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8593 train/patch_acc=0.7188 train/subpatch_loss=0.0955 train/subpatch_acc=0.9375 train/location_loss=0.8936 train/location_acc=0.8125 train/token_losses=10.83 throughput/total_tokens=4,964,352 throughput/device/tokens_per_second=662.4 throughput/device/batches_per_second=0.0404 2026-07-27 14:09:48.664 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=304/600, eta=2 hours, 2 minutes] train/CrossEntropyLoss=0.0928 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.7629 train/patch_acc=0.8438 train/subpatch_loss=0.2155 train/subpatch_acc=0.9375 train/location_loss=0.9941 train/location_acc=0.7500 train/token_losses=10.94 throughput/total_tokens=4,980,736 throughput/device/tokens_per_second=663.7 throughput/device/batches_per_second=0.0405 2026-07-27 14:10:13.177 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=305/600, eta=2 hours, 1 minute] train/CrossEntropyLoss=0.0449 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.9480 train/patch_acc=0.7188 train/subpatch_loss=0.1819 train/subpatch_acc=0.9375 train/location_loss=1.075 train/location_acc=0.5625 train/token_losses=12.61 throughput/total_tokens=4,997,120 throughput/device/tokens_per_second=664.7 throughput/device/batches_per_second=0.0406 2026-07-27 14:10:37.724 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=306/600, eta=2 hours, 1 minute] train/CrossEntropyLoss=0.0392 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.013 train/patch_acc=0.7188 train/subpatch_loss=0.2021 train/subpatch_acc=0.8750 train/location_loss=1.253 train/location_acc=0.7500 train/token_losses=13.92 throughput/total_tokens=5,013,504 throughput/device/tokens_per_second=665.1 throughput/device/batches_per_second=0.0406 2026-07-27 14:11:02.262 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=307/600, eta=2 hours] train/CrossEntropyLoss=0.0413 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8773 train/patch_acc=0.7812 train/subpatch_loss=0.3322 train/subpatch_acc=0.8750 train/location_loss=1.295 train/location_acc=0.6250 train/token_losses=13.53 throughput/total_tokens=5,029,888 throughput/device/tokens_per_second=665.5 throughput/device/batches_per_second=0.0406 2026-07-27 14:11:26.781 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=308/600, eta=2 hours] train/CrossEntropyLoss=0.0496 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.8507 train/patch_acc=0.7812 train/subpatch_loss=0.1386 train/subpatch_acc=1.0000 train/location_loss=1.331 train/location_acc=0.5000 train/token_losses=12.68 throughput/total_tokens=5,046,272 throughput/device/tokens_per_second=665.8 throughput/device/batches_per_second=0.0406 2026-07-27 14:11:51.328 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=309/600, eta=2 hours] train/CrossEntropyLoss=0.0294 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8443 train/patch_acc=0.6875 train/subpatch_loss=0.1552 train/subpatch_acc=1.0000 train/location_loss=1.149 train/location_acc=0.5000 train/token_losses=11.97 throughput/total_tokens=5,062,656 throughput/device/tokens_per_second=666.0 throughput/device/batches_per_second=0.0407 2026-07-27 14:12:15.850 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=310/600, eta=1 hour, 59 minutes] train/CrossEntropyLoss=0.0439 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.191 train/patch_acc=0.6250 train/subpatch_loss=0.2023 train/subpatch_acc=0.9375 train/location_loss=1.482 train/location_acc=0.6250 train/token_losses=16.26 throughput/total_tokens=5,079,040 throughput/device/tokens_per_second=666.2 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:12:40.372 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=311/600, eta=1 hour, 59 minutes] train/CrossEntropyLoss=0.0412 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.7569 train/patch_acc=0.7812 train/subpatch_loss=0.1726 train/subpatch_acc=0.9375 train/location_loss=1.239 train/location_acc=0.6250 train/token_losses=11.70 throughput/total_tokens=5,095,424 throughput/device/tokens_per_second=666.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:13:04.935 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=312/600, eta=1 hour, 58 minutes] train/CrossEntropyLoss=0.0651 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9967 train/patch_acc=0.6875 train/subpatch_loss=0.2212 train/subpatch_acc=0.9375 train/location_loss=1.318 train/location_acc=0.3750 train/token_losses=14.13 throughput/total_tokens=5,111,808 throughput/device/tokens_per_second=666.5 throughput/device/batches_per_second=0.0407 2026-07-27 14:13:29.460 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=313/600, eta=1 hour, 58 minutes] train/CrossEntropyLoss=0.0235 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.113 train/patch_acc=0.7188 train/subpatch_loss=0.1366 train/subpatch_acc=1.0000 train/location_loss=1.191 train/location_acc=0.6250 train/token_losses=14.22 throughput/total_tokens=5,128,192 throughput/device/tokens_per_second=666.6 throughput/device/batches_per_second=0.0407 2026-07-27 14:13:53.984 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=314/600, eta=1 hour, 58 minutes] train/CrossEntropyLoss=0.1291 train/ZLoss=0.0006 train/Accuracy=0.9604 train/patch_loss=1.010 train/patch_acc=0.7500 train/subpatch_loss=0.1287 train/subpatch_acc=1.0000 train/location_loss=1.311 train/location_acc=0.5625 train/token_losses=13.84 throughput/total_tokens=5,144,576 throughput/device/tokens_per_second=666.7 throughput/device/batches_per_second=0.0407 2026-07-27 14:14:18.499 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=315/600, eta=1 hour, 57 minutes] train/CrossEntropyLoss=0.0371 train/ZLoss=0.0005 train/Accuracy=0.9878 train/patch_loss=1.146 train/patch_acc=0.5938 train/subpatch_loss=0.0987 train/subpatch_acc=1.0000 train/location_loss=1.247 train/location_acc=0.5625 train/token_losses=14.55 throughput/total_tokens=5,160,960 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 14:14:43.043 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=316/600, eta=1 hour, 57 minutes] train/CrossEntropyLoss=0.0219 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.017 train/patch_acc=0.7188 train/subpatch_loss=0.1595 train/subpatch_acc=0.9375 train/location_loss=1.458 train/location_acc=0.3750 train/token_losses=14.60 throughput/total_tokens=5,177,344 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 14:15:07.594 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=317/600, eta=1 hour, 56 minutes] train/CrossEntropyLoss=0.0340 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7337 train/patch_acc=0.7812 train/subpatch_loss=0.2857 train/subpatch_acc=0.8750 train/location_loss=0.9099 train/location_acc=0.8125 train/token_losses=10.65 throughput/total_tokens=5,193,728 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 14:15:32.138 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=318/600, eta=1 hour, 56 minutes] train/CrossEntropyLoss=0.0675 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9511 train/patch_acc=0.7500 train/subpatch_loss=0.2293 train/subpatch_acc=0.8750 train/location_loss=1.194 train/location_acc=0.5625 train/token_losses=13.30 throughput/total_tokens=5,210,112 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 14:15:56.675 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=319/600, eta=1 hour, 56 minutes] train/CrossEntropyLoss=0.0099 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.083 train/patch_acc=0.6250 train/subpatch_loss=0.3430 train/subpatch_acc=0.8750 train/location_loss=1.618 train/location_acc=0.4375 train/token_losses=16.51 throughput/total_tokens=5,226,496 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 14:16:21.190 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=320/600, eta=1 hour, 55 minutes] train/CrossEntropyLoss=0.0384 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.004 train/patch_acc=0.7188 train/subpatch_loss=0.2516 train/subpatch_acc=0.8750 train/location_loss=1.039 train/location_acc=0.7500 train/token_losses=13.20 throughput/total_tokens=5,242,880 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:16:45.692 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=321/600, eta=1 hour, 55 minutes] train/CrossEntropyLoss=0.2115 train/ZLoss=0.0006 train/Accuracy=0.9451 train/patch_loss=0.8418 train/patch_acc=0.7188 train/subpatch_loss=0.2313 train/subpatch_acc=0.9375 train/location_loss=1.369 train/location_acc=0.3750 train/token_losses=13.13 throughput/total_tokens=5,259,264 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:17:10.202 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=322/600, eta=1 hour, 54 minutes] train/CrossEntropyLoss=0.0184 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8040 train/patch_acc=0.7500 train/subpatch_loss=0.0753 train/subpatch_acc=1.0000 train/location_loss=0.9972 train/location_acc=0.8750 train/token_losses=10.72 throughput/total_tokens=5,275,648 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:17:34.714 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=323/600, eta=1 hour, 54 minutes] train/CrossEntropyLoss=0.0421 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.8985 train/patch_acc=0.8125 train/subpatch_loss=0.1278 train/subpatch_acc=1.0000 train/location_loss=1.097 train/location_acc=0.6250 train/token_losses=12.09 throughput/total_tokens=5,292,032 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:17:59.224 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=324/600, eta=1 hour, 54 minutes] train/CrossEntropyLoss=0.0032 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=0.9348 train/patch_acc=0.8125 train/subpatch_loss=0.0931 train/subpatch_acc=1.0000 train/location_loss=1.279 train/location_acc=0.6250 train/token_losses=12.97 throughput/total_tokens=5,308,416 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:18:23.737 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=325/600, eta=1 hour, 53 minutes] train/CrossEntropyLoss=0.0425 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6744 train/patch_acc=0.7188 train/subpatch_loss=0.1788 train/subpatch_acc=1.0000 train/location_loss=1.152 train/location_acc=0.6250 train/token_losses=10.72 throughput/total_tokens=5,324,800 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:18:48.333 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=326/600, eta=1 hour, 53 minutes] train/CrossEntropyLoss=0.0442 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.039 train/patch_acc=0.7188 train/subpatch_loss=0.0777 train/subpatch_acc=1.0000 train/location_loss=1.150 train/location_acc=0.6250 train/token_losses=13.22 throughput/total_tokens=5,341,184 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:19:12.845 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=327/600, eta=1 hour, 52 minutes] train/CrossEntropyLoss=0.0584 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.6000 train/patch_acc=0.8750 train/subpatch_loss=0.1293 train/subpatch_acc=0.9375 train/location_loss=1.172 train/location_acc=0.5000 train/token_losses=10.01 throughput/total_tokens=5,357,568 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:19:37.348 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=328/600, eta=1 hour, 52 minutes] train/CrossEntropyLoss=0.0263 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7558 train/patch_acc=0.7188 train/subpatch_loss=0.1653 train/subpatch_acc=1.0000 train/location_loss=1.130 train/location_acc=0.5000 train/token_losses=11.23 throughput/total_tokens=5,373,952 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:20:01.869 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=329/600, eta=1 hour, 51 minutes] train/CrossEntropyLoss=0.1208 train/ZLoss=0.0008 train/Accuracy=0.9726 train/patch_loss=1.302 train/patch_acc=0.6562 train/subpatch_loss=0.1052 train/subpatch_acc=1.0000 train/location_loss=1.282 train/location_acc=0.5000 train/token_losses=15.96 throughput/total_tokens=5,390,336 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:20:26.413 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=330/600, eta=1 hour, 51 minutes] train/CrossEntropyLoss=0.1679 train/ZLoss=0.0011 train/Accuracy=0.9543 train/patch_loss=1.168 train/patch_acc=0.6562 train/subpatch_loss=0.0813 train/subpatch_acc=1.0000 train/location_loss=1.107 train/location_acc=0.6875 train/token_losses=14.10 throughput/total_tokens=5,406,720 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:20:50.940 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=331/600, eta=1 hour, 51 minutes] train/CrossEntropyLoss=0.0432 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.080 train/patch_acc=0.6562 train/subpatch_loss=0.1471 train/subpatch_acc=0.9375 train/location_loss=1.119 train/location_acc=0.5625 train/token_losses=13.70 throughput/total_tokens=5,423,104 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:21:15.464 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=332/600, eta=1 hour, 50 minutes] train/CrossEntropyLoss=0.0202 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6699 train/patch_acc=0.8438 train/subpatch_loss=0.0482 train/subpatch_acc=1.0000 train/location_loss=1.089 train/location_acc=0.6250 train/token_losses=9.906 throughput/total_tokens=5,439,488 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:21:39.997 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=333/600, eta=1 hour, 50 minutes] train/CrossEntropyLoss=0.0213 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8133 train/patch_acc=0.8125 train/subpatch_loss=0.1321 train/subpatch_acc=1.0000 train/location_loss=0.9430 train/location_acc=0.8125 train/token_losses=10.81 throughput/total_tokens=5,455,872 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:22:04.539 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=334/600, eta=1 hour, 49 minutes] train/CrossEntropyLoss=0.0507 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.8311 train/patch_acc=0.7188 train/subpatch_loss=0.1533 train/subpatch_acc=0.9375 train/location_loss=0.9267 train/location_acc=0.7500 train/token_losses=10.97 throughput/total_tokens=5,472,256 throughput/device/tokens_per_second=668.0 throughput/device/batches_per_second=0.0408 2026-07-27 14:22:29.094 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=335/600, eta=1 hour, 49 minutes] train/CrossEntropyLoss=0.0103 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9921 train/patch_acc=0.7188 train/subpatch_loss=0.1703 train/subpatch_acc=1.0000 train/location_loss=1.372 train/location_acc=0.4375 train/token_losses=14.10 throughput/total_tokens=5,488,640 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:22:53.634 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=336/600, eta=1 hour, 49 minutes] train/CrossEntropyLoss=0.0315 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9215 train/patch_acc=0.7188 train/subpatch_loss=0.1838 train/subpatch_acc=1.0000 train/location_loss=1.044 train/location_acc=0.6250 train/token_losses=12.28 throughput/total_tokens=5,505,024 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:23:18.182 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=337/600, eta=1 hour, 48 minutes] train/CrossEntropyLoss=0.0424 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.7049 train/patch_acc=0.8125 train/subpatch_loss=0.1373 train/subpatch_acc=0.9375 train/location_loss=1.119 train/location_acc=0.5625 train/token_losses=10.66 throughput/total_tokens=5,521,408 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:23:42.741 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=338/600, eta=1 hour, 48 minutes] train/CrossEntropyLoss=0.0230 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.016 train/patch_acc=0.7500 train/subpatch_loss=0.2663 train/subpatch_acc=0.8750 train/location_loss=1.031 train/location_acc=0.6250 train/token_losses=13.32 throughput/total_tokens=5,537,792 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:24:07.295 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=339/600, eta=1 hour, 47 minutes] train/CrossEntropyLoss=0.1185 train/ZLoss=0.0008 train/Accuracy=0.9756 train/patch_loss=0.8718 train/patch_acc=0.7500 train/subpatch_loss=0.2158 train/subpatch_acc=0.9375 train/location_loss=1.139 train/location_acc=0.5000 train/token_losses=12.39 throughput/total_tokens=5,554,176 throughput/device/tokens_per_second=667.9 throughput/device/batches_per_second=0.0408 2026-07-27 14:24:31.857 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=340/600, eta=1 hour, 47 minutes] train/CrossEntropyLoss=0.0297 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=1.053 train/patch_acc=0.6875 train/subpatch_loss=0.1734 train/subpatch_acc=0.9375 train/location_loss=1.026 train/location_acc=0.7500 train/token_losses=13.22 throughput/total_tokens=5,570,560 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:24:56.406 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=341/600, eta=1 hour, 47 minutes] train/CrossEntropyLoss=0.0302 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.7152 train/patch_acc=0.7812 train/subpatch_loss=0.1408 train/subpatch_acc=0.9375 train/location_loss=1.084 train/location_acc=0.8125 train/token_losses=10.62 throughput/total_tokens=5,586,944 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 14:25:20.962 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=342/600, eta=1 hour, 46 minutes] train/CrossEntropyLoss=0.0382 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.8378 train/patch_acc=0.7812 train/subpatch_loss=0.1011 train/subpatch_acc=1.0000 train/location_loss=0.8535 train/location_acc=0.8125 train/token_losses=10.52 throughput/total_tokens=5,603,328 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 14:25:45.535 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=343/600, eta=1 hour, 46 minutes] train/CrossEntropyLoss=0.0304 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.3953 train/patch_acc=0.9375 train/subpatch_loss=0.0955 train/subpatch_acc=1.0000 train/location_loss=0.9114 train/location_acc=0.6250 train/token_losses=7.190 throughput/total_tokens=5,619,712 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 14:26:10.075 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=344/600, eta=1 hour, 45 minutes] train/CrossEntropyLoss=0.0957 train/ZLoss=0.0005 train/Accuracy=0.9848 train/patch_loss=1.036 train/patch_acc=0.7500 train/subpatch_loss=0.2708 train/subpatch_acc=0.8750 train/location_loss=1.162 train/location_acc=0.6875 train/token_losses=14.02 throughput/total_tokens=5,636,096 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 14:26:34.626 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=345/600, eta=1 hour, 45 minutes] train/CrossEntropyLoss=0.0157 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8330 train/patch_acc=0.7188 train/subpatch_loss=0.1904 train/subpatch_acc=0.9375 train/location_loss=0.7190 train/location_acc=0.8125 train/token_losses=10.30 throughput/total_tokens=5,652,480 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 14:26:59.200 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=346/600, eta=1 hour, 45 minutes] train/CrossEntropyLoss=0.0287 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.021 train/patch_acc=0.7812 train/subpatch_loss=0.1808 train/subpatch_acc=0.9375 train/location_loss=1.150 train/location_acc=0.5000 train/token_losses=13.49 throughput/total_tokens=5,668,864 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 14:27:23.772 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=347/600, eta=1 hour, 44 minutes] train/CrossEntropyLoss=0.0432 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.046 train/patch_acc=0.7188 train/subpatch_loss=0.1321 train/subpatch_acc=0.9375 train/location_loss=1.291 train/location_acc=0.5625 train/token_losses=14.06 throughput/total_tokens=5,685,248 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 14:27:48.339 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=348/600, eta=1 hour, 44 minutes] train/CrossEntropyLoss=0.0080 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6053 train/patch_acc=0.8125 train/subpatch_loss=0.0901 train/subpatch_acc=1.0000 train/location_loss=0.8419 train/location_acc=0.6875 train/token_losses=8.571 throughput/total_tokens=5,701,632 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:28:12.886 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=349/600, eta=1 hour, 43 minutes] train/CrossEntropyLoss=0.0220 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7056 train/patch_acc=0.8125 train/subpatch_loss=0.1371 train/subpatch_acc=1.0000 train/location_loss=0.8736 train/location_acc=0.6875 train/token_losses=9.688 throughput/total_tokens=5,718,016 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:28:37.427 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=350/600, eta=1 hour, 43 minutes] train/CrossEntropyLoss=0.0242 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.4931 train/patch_acc=0.9062 train/subpatch_loss=0.0503 train/subpatch_acc=1.0000 train/location_loss=0.9375 train/location_acc=0.5000 train/token_losses=7.897 throughput/total_tokens=5,734,400 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:29:01.997 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=351/600, eta=1 hour, 42 minutes] train/CrossEntropyLoss=0.0692 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.8682 train/patch_acc=0.7500 train/subpatch_loss=0.0847 train/subpatch_acc=1.0000 train/location_loss=1.074 train/location_acc=0.6250 train/token_losses=11.58 throughput/total_tokens=5,750,784 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:29:26.550 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=352/600, eta=1 hour, 42 minutes] train/CrossEntropyLoss=0.0858 train/ZLoss=0.0005 train/Accuracy=0.9726 train/patch_loss=0.9692 train/patch_acc=0.7188 train/subpatch_loss=0.2187 train/subpatch_acc=0.9375 train/location_loss=1.129 train/location_acc=0.6875 train/token_losses=13.14 throughput/total_tokens=5,767,168 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:29:51.085 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=353/600, eta=1 hour, 42 minutes] train/CrossEntropyLoss=0.0006 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.7114 train/patch_acc=0.8438 train/subpatch_loss=0.1272 train/subpatch_acc=0.9375 train/location_loss=1.154 train/location_acc=0.5000 train/token_losses=10.81 throughput/total_tokens=5,783,552 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:30:15.709 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=354/600, eta=1 hour, 41 minutes] train/CrossEntropyLoss=0.0413 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.8765 train/patch_acc=0.7812 train/subpatch_loss=0.1639 train/subpatch_acc=0.9375 train/location_loss=1.178 train/location_acc=0.6875 train/token_losses=12.38 throughput/total_tokens=5,799,936 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:30:40.283 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=355/600, eta=1 hour, 41 minutes] train/CrossEntropyLoss=0.0175 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.4612 train/patch_acc=0.8750 train/subpatch_loss=0.0488 train/subpatch_acc=1.0000 train/location_loss=0.7984 train/location_acc=0.8750 train/token_losses=7.078 throughput/total_tokens=5,816,320 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:31:04.841 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=356/600, eta=1 hour, 40 minutes] train/CrossEntropyLoss=0.0334 train/ZLoss=0.0006 train/Accuracy=0.9939 train/patch_loss=1.035 train/patch_acc=0.8125 train/subpatch_loss=0.0741 train/subpatch_acc=1.0000 train/location_loss=1.339 train/location_acc=0.5625 train/token_losses=13.93 throughput/total_tokens=5,832,704 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:31:29.386 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=357/600, eta=1 hour, 40 minutes] train/CrossEntropyLoss=0.0587 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.7686 train/patch_acc=0.8125 train/subpatch_loss=0.2499 train/subpatch_acc=0.9375 train/location_loss=1.157 train/location_acc=0.6250 train/token_losses=11.78 throughput/total_tokens=5,849,088 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:31:53.920 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=358/600, eta=1 hour, 40 minutes] train/CrossEntropyLoss=0.0291 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.8134 train/patch_acc=0.7500 train/subpatch_loss=0.2464 train/subpatch_acc=0.8125 train/location_loss=1.003 train/location_acc=0.6250 train/token_losses=11.50 throughput/total_tokens=5,865,472 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:32:18.456 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=359/600, eta=1 hour, 39 minutes] train/CrossEntropyLoss=0.0082 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9491 train/patch_acc=0.8125 train/subpatch_loss=0.2010 train/subpatch_acc=1.0000 train/location_loss=1.122 train/location_acc=0.6250 train/token_losses=12.88 throughput/total_tokens=5,881,856 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:32:43.016 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=360/600, eta=1 hour, 39 minutes] train/CrossEntropyLoss=0.0099 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9552 train/patch_acc=0.7188 train/subpatch_loss=0.0609 train/subpatch_acc=1.0000 train/location_loss=1.086 train/location_acc=0.7500 train/token_losses=12.23 throughput/total_tokens=5,898,240 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:33:07.582 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=361/600, eta=1 hour, 38 minutes] train/CrossEntropyLoss=0.1003 train/ZLoss=0.0007 train/Accuracy=0.9756 train/patch_loss=0.8210 train/patch_acc=0.7500 train/subpatch_loss=0.1032 train/subpatch_acc=1.0000 train/location_loss=1.145 train/location_acc=0.5625 train/token_losses=11.56 throughput/total_tokens=5,914,624 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:33:32.135 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=362/600, eta=1 hour, 38 minutes] train/CrossEntropyLoss=0.0785 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.8524 train/patch_acc=0.7812 train/subpatch_loss=0.1155 train/subpatch_acc=1.0000 train/location_loss=1.034 train/location_acc=0.5625 train/token_losses=11.42 throughput/total_tokens=5,931,008 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:33:56.676 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=363/600, eta=1 hour, 38 minutes] train/CrossEntropyLoss=0.0799 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.9092 train/patch_acc=0.7188 train/subpatch_loss=0.1551 train/subpatch_acc=1.0000 train/location_loss=0.8875 train/location_acc=0.8125 train/token_losses=11.44 throughput/total_tokens=5,947,392 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:34:21.225 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=364/600, eta=1 hour, 37 minutes] train/CrossEntropyLoss=0.0323 train/ZLoss=0.0011 train/Accuracy=0.9909 train/patch_loss=1.083 train/patch_acc=0.7188 train/subpatch_loss=0.1099 train/subpatch_acc=1.0000 train/location_loss=1.051 train/location_acc=0.5625 train/token_losses=13.31 throughput/total_tokens=5,963,776 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:34:45.780 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=365/600, eta=1 hour, 37 minutes] train/CrossEntropyLoss=0.0045 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=0.6102 train/patch_acc=0.8750 train/subpatch_loss=0.0445 train/subpatch_acc=1.0000 train/location_loss=0.7487 train/location_acc=0.8125 train/token_losses=8.054 throughput/total_tokens=5,980,160 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:35:10.354 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=366/600, eta=1 hour, 36 minutes] train/CrossEntropyLoss=0.0191 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8128 train/patch_acc=0.7812 train/subpatch_loss=0.1747 train/subpatch_acc=0.9375 train/location_loss=0.8191 train/location_acc=0.8125 train/token_losses=10.48 throughput/total_tokens=5,996,544 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:35:34.893 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=367/600, eta=1 hour, 36 minutes] train/CrossEntropyLoss=0.2460 train/ZLoss=0.0006 train/Accuracy=0.9421 train/patch_loss=0.8669 train/patch_acc=0.7812 train/subpatch_loss=0.0899 train/subpatch_acc=1.0000 train/location_loss=1.272 train/location_acc=0.6250 train/token_losses=12.38 throughput/total_tokens=6,012,928 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:35:59.441 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=368/600, eta=1 hour, 35 minutes] train/CrossEntropyLoss=0.0058 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=1.071 train/patch_acc=0.6250 train/subpatch_loss=0.2343 train/subpatch_acc=0.8750 train/location_loss=1.131 train/location_acc=0.5625 train/token_losses=14.03 throughput/total_tokens=6,029,312 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:36:23.990 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=369/600, eta=1 hour, 35 minutes] train/CrossEntropyLoss=0.0079 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6570 train/patch_acc=0.7812 train/subpatch_loss=0.1164 train/subpatch_acc=1.0000 train/location_loss=0.8039 train/location_acc=0.7500 train/token_losses=8.937 throughput/total_tokens=6,045,696 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:36:48.530 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=370/600, eta=1 hour, 35 minutes] train/CrossEntropyLoss=0.0521 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.6311 train/patch_acc=0.8125 train/subpatch_loss=0.0938 train/subpatch_acc=1.0000 train/location_loss=0.9289 train/location_acc=0.6875 train/token_losses=9.140 throughput/total_tokens=6,062,080 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:37:13.082 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=371/600, eta=1 hour, 34 minutes] train/CrossEntropyLoss=0.0089 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=1.172 train/patch_acc=0.6250 train/subpatch_loss=0.0535 train/subpatch_acc=1.0000 train/location_loss=0.8906 train/location_acc=0.6875 train/token_losses=13.15 throughput/total_tokens=6,078,464 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:37:37.620 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=372/600, eta=1 hour, 34 minutes] train/CrossEntropyLoss=0.0307 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=1.297 train/patch_acc=0.6250 train/subpatch_loss=0.1250 train/subpatch_acc=1.0000 train/location_loss=0.9459 train/location_acc=0.6875 train/token_losses=14.66 throughput/total_tokens=6,094,848 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:38:02.167 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=373/600, eta=1 hour, 33 minutes] train/CrossEntropyLoss=0.0202 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=1.170 train/patch_acc=0.6875 train/subpatch_loss=0.1222 train/subpatch_acc=1.0000 train/location_loss=1.263 train/location_acc=0.6250 train/token_losses=14.90 throughput/total_tokens=6,111,232 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:38:26.731 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=374/600, eta=1 hour, 33 minutes] train/CrossEntropyLoss=0.0156 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.3230 train/patch_acc=0.9688 train/subpatch_loss=0.0938 train/subpatch_acc=1.0000 train/location_loss=0.5464 train/location_acc=0.9375 train/token_losses=5.145 throughput/total_tokens=6,127,616 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:38:51.309 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=375/600, eta=1 hour, 33 minutes] train/CrossEntropyLoss=0.0127 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.6190 train/patch_acc=0.8438 train/subpatch_loss=0.0396 train/subpatch_acc=1.0000 train/location_loss=1.168 train/location_acc=0.6250 train/token_losses=9.784 throughput/total_tokens=6,144,000 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:39:15.851 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=376/600, eta=1 hour, 32 minutes] train/CrossEntropyLoss=0.0171 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=1.079 train/patch_acc=0.6562 train/subpatch_loss=0.1995 train/subpatch_acc=1.0000 train/location_loss=0.9650 train/location_acc=0.7500 train/token_losses=13.29 throughput/total_tokens=6,160,384 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:39:40.402 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=377/600, eta=1 hour, 32 minutes] train/CrossEntropyLoss=0.0324 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.6221 train/patch_acc=0.8438 train/subpatch_loss=0.2021 train/subpatch_acc=0.9375 train/location_loss=0.7669 train/location_acc=0.8125 train/token_losses=8.852 throughput/total_tokens=6,176,768 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:40:04.952 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=378/600, eta=1 hour, 31 minutes] train/CrossEntropyLoss=0.0394 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6855 train/patch_acc=0.8125 train/subpatch_loss=0.0569 train/subpatch_acc=1.0000 train/location_loss=1.085 train/location_acc=0.5625 train/token_losses=10.05 throughput/total_tokens=6,193,152 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:40:29.484 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=379/600, eta=1 hour, 31 minutes] train/CrossEntropyLoss=0.0175 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9403 train/patch_acc=0.7500 train/subpatch_loss=0.1711 train/subpatch_acc=0.8750 train/location_loss=0.7769 train/location_acc=0.8750 train/token_losses=11.31 throughput/total_tokens=6,209,536 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:40:54.044 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=380/600, eta=1 hour, 31 minutes] train/CrossEntropyLoss=0.0206 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.8908 train/patch_acc=0.8438 train/subpatch_loss=0.1397 train/subpatch_acc=1.0000 train/location_loss=1.211 train/location_acc=0.6250 train/token_losses=12.53 throughput/total_tokens=6,225,920 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:41:18.612 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=381/600, eta=1 hour, 30 minutes] train/CrossEntropyLoss=0.0586 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.6052 train/patch_acc=0.8438 train/subpatch_loss=0.1185 train/subpatch_acc=1.0000 train/location_loss=0.7060 train/location_acc=0.8125 train/token_losses=8.140 throughput/total_tokens=6,242,304 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:41:43.157 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=382/600, eta=1 hour, 30 minutes] train/CrossEntropyLoss=0.0495 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9669 train/patch_acc=0.7500 train/subpatch_loss=0.2232 train/subpatch_acc=0.8750 train/location_loss=0.8818 train/location_acc=0.6875 train/token_losses=12.15 throughput/total_tokens=6,258,688 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:42:07.728 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=383/600, eta=1 hour, 29 minutes] train/CrossEntropyLoss=0.0271 train/ZLoss=0.0004 train/Accuracy=0.9878 train/patch_loss=0.8754 train/patch_acc=0.7500 train/subpatch_loss=0.1482 train/subpatch_acc=1.0000 train/location_loss=1.057 train/location_acc=0.6250 train/token_losses=11.82 throughput/total_tokens=6,275,072 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:42:32.284 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=384/600, eta=1 hour, 29 minutes] train/CrossEntropyLoss=0.0421 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7511 train/patch_acc=0.8125 train/subpatch_loss=0.0557 train/subpatch_acc=1.0000 train/location_loss=1.036 train/location_acc=0.6875 train/token_losses=10.38 throughput/total_tokens=6,291,456 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:42:56.821 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=385/600, eta=1 hour, 29 minutes] train/CrossEntropyLoss=0.0333 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.200 train/patch_acc=0.6250 train/subpatch_loss=0.0318 train/subpatch_acc=1.0000 train/location_loss=0.8447 train/location_acc=0.7500 train/token_losses=13.11 throughput/total_tokens=6,307,840 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:43:21.376 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=386/600, eta=1 hour, 28 minutes] train/CrossEntropyLoss=0.0198 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.5353 train/patch_acc=0.8750 train/subpatch_loss=0.1601 train/subpatch_acc=0.9375 train/location_loss=0.7699 train/location_acc=0.8125 train/token_losses=8.002 throughput/total_tokens=6,324,224 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:43:45.903 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=387/600, eta=1 hour, 28 minutes] train/CrossEntropyLoss=0.1494 train/ZLoss=0.0006 train/Accuracy=0.9665 train/patch_loss=0.9262 train/patch_acc=0.7188 train/subpatch_loss=0.1992 train/subpatch_acc=0.9375 train/location_loss=1.056 train/location_acc=0.6250 train/token_losses=12.43 throughput/total_tokens=6,340,608 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 14:44:10.478 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=388/600, eta=1 hour, 27 minutes] train/CrossEntropyLoss=0.0312 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.9358 train/patch_acc=0.7188 train/subpatch_loss=0.0860 train/subpatch_acc=1.0000 train/location_loss=0.7661 train/location_acc=0.7500 train/token_losses=10.89 throughput/total_tokens=6,356,992 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:44:35.020 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=389/600, eta=1 hour, 27 minutes] train/CrossEntropyLoss=0.0321 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7853 train/patch_acc=0.7812 train/subpatch_loss=0.1053 train/subpatch_acc=1.0000 train/location_loss=0.7358 train/location_acc=0.8125 train/token_losses=9.647 throughput/total_tokens=6,373,376 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:44:59.598 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=390/600, eta=1 hour, 26 minutes] train/CrossEntropyLoss=0.0126 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.9140 train/patch_acc=0.7500 train/subpatch_loss=0.0871 train/subpatch_acc=1.0000 train/location_loss=0.9279 train/location_acc=0.6250 train/token_losses=11.37 throughput/total_tokens=6,389,760 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:45:24.157 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=391/600, eta=1 hour, 26 minutes] train/CrossEntropyLoss=0.0403 train/ZLoss=0.0007 train/Accuracy=0.9878 train/patch_loss=0.9248 train/patch_acc=0.6562 train/subpatch_loss=0.1124 train/subpatch_acc=1.0000 train/location_loss=1.133 train/location_acc=0.5625 train/token_losses=12.38 throughput/total_tokens=6,406,144 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:45:48.723 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=392/600, eta=1 hour, 26 minutes] train/CrossEntropyLoss=0.0047 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=0.8147 train/patch_acc=0.6875 train/subpatch_loss=0.0479 train/subpatch_acc=1.0000 train/location_loss=0.5358 train/location_acc=0.8125 train/token_losses=8.853 throughput/total_tokens=6,422,528 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:46:13.280 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=393/600, eta=1 hour, 25 minutes] train/CrossEntropyLoss=0.0145 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.7660 train/patch_acc=0.8438 train/subpatch_loss=0.1249 train/subpatch_acc=1.0000 train/location_loss=1.378 train/location_acc=0.3750 train/token_losses=12.14 throughput/total_tokens=6,438,912 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:46:37.854 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=394/600, eta=1 hour, 25 minutes] train/CrossEntropyLoss=0.0291 train/ZLoss=0.0003 train/Accuracy=0.9939 train/patch_loss=0.7523 train/patch_acc=0.7188 train/subpatch_loss=0.0809 train/subpatch_acc=1.0000 train/location_loss=1.061 train/location_acc=0.7500 train/token_losses=10.58 throughput/total_tokens=6,455,296 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:47:02.396 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=395/600, eta=1 hour, 24 minutes] train/CrossEntropyLoss=0.0064 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.8084 train/patch_acc=0.8125 train/subpatch_loss=0.0668 train/subpatch_acc=1.0000 train/location_loss=1.005 train/location_acc=0.6250 train/token_losses=10.75 throughput/total_tokens=6,471,680 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:47:26.919 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=396/600, eta=1 hour, 24 minutes] train/CrossEntropyLoss=0.0467 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7176 train/patch_acc=0.7188 train/subpatch_loss=0.1693 train/subpatch_acc=1.0000 train/location_loss=1.308 train/location_acc=0.5000 train/token_losses=11.65 throughput/total_tokens=6,488,064 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:47:51.473 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=397/600, eta=1 hour, 24 minutes] train/CrossEntropyLoss=0.0231 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8129 train/patch_acc=0.7500 train/subpatch_loss=0.1048 train/subpatch_acc=1.0000 train/location_loss=1.216 train/location_acc=0.5000 train/token_losses=11.79 throughput/total_tokens=6,504,448 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:48:16.017 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=398/600, eta=1 hour, 23 minutes] train/CrossEntropyLoss=0.0447 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.7317 train/patch_acc=0.7500 train/subpatch_loss=0.1439 train/subpatch_acc=1.0000 train/location_loss=1.335 train/location_acc=0.5000 train/token_losses=11.77 throughput/total_tokens=6,520,832 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:48:40.558 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=399/600, eta=1 hour, 23 minutes] train/CrossEntropyLoss=0.0524 train/ZLoss=0.0002 train/Accuracy=0.9787 train/patch_loss=0.6636 train/patch_acc=0.7500 train/subpatch_loss=0.1996 train/subpatch_acc=0.8750 train/location_loss=1.300 train/location_acc=0.5625 train/token_losses=11.31 throughput/total_tokens=6,537,216 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 14:49:05.103 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=400/600, eta=1 hour, 22 minutes] train/CrossEntropyLoss=0.0489 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=1.086 train/patch_acc=0.7188 train/subpatch_loss=0.0300 train/subpatch_acc=1.0000 train/location_loss=0.8495 train/location_acc=0.6875 train/token_losses=12.20 throughput/total_tokens=6,553,600 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:49:05.104 2012bc4f83ae:0 olmo.train.trainer:1415 INFO Saving checkpoint... 2026-07-27 14:49:12.852 2012bc4f83ae:0 olmo.train.trainer:1420 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step400 2026-07-27 14:49:38.494 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=401/600, eta=1 hour, 22 minutes] train/CrossEntropyLoss=0.0069 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.7918 train/patch_acc=0.6875 train/subpatch_loss=0.2056 train/subpatch_acc=1.0000 train/location_loss=1.037 train/location_acc=0.6875 train/token_losses=11.31 throughput/total_tokens=6,569,984 throughput/device/tokens_per_second=639.1 throughput/device/batches_per_second=0.0390 2026-07-27 14:50:03.031 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=402/600, eta=1 hour, 22 minutes] train/CrossEntropyLoss=0.0435 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7456 train/patch_acc=0.7812 train/subpatch_loss=0.0785 train/subpatch_acc=1.0000 train/location_loss=0.7108 train/location_acc=0.8125 train/token_losses=9.122 throughput/total_tokens=6,586,368 throughput/device/tokens_per_second=653.1 throughput/device/batches_per_second=0.0399 2026-07-27 14:50:27.582 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=403/600, eta=1 hour, 21 minutes] train/CrossEntropyLoss=0.0216 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6839 train/patch_acc=0.8125 train/subpatch_loss=0.1257 train/subpatch_acc=1.0000 train/location_loss=0.9212 train/location_acc=0.6250 train/token_losses=9.658 throughput/total_tokens=6,602,752 throughput/device/tokens_per_second=657.8 throughput/device/batches_per_second=0.0401 2026-07-27 14:50:52.146 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=404/600, eta=1 hour, 21 minutes] train/CrossEntropyLoss=0.0219 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.7496 train/patch_acc=0.8125 train/subpatch_loss=0.0437 train/subpatch_acc=1.0000 train/location_loss=0.6998 train/location_acc=0.8125 train/token_losses=8.971 throughput/total_tokens=6,619,136 throughput/device/tokens_per_second=660.1 throughput/device/batches_per_second=0.0403 2026-07-27 14:51:16.686 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=405/600, eta=1 hour, 20 minutes] train/CrossEntropyLoss=0.0156 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9175 train/patch_acc=0.7188 train/subpatch_loss=0.0572 train/subpatch_acc=1.0000 train/location_loss=1.011 train/location_acc=0.6875 train/token_losses=11.61 throughput/total_tokens=6,635,520 throughput/device/tokens_per_second=661.6 throughput/device/batches_per_second=0.0404 2026-07-27 14:51:41.239 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=406/600, eta=1 hour, 20 minutes] train/CrossEntropyLoss=0.0722 train/ZLoss=0.0007 train/Accuracy=0.9817 train/patch_loss=0.9934 train/patch_acc=0.7188 train/subpatch_loss=0.1907 train/subpatch_acc=0.8750 train/location_loss=0.7134 train/location_acc=0.8750 train/token_losses=11.56 throughput/total_tokens=6,651,904 throughput/device/tokens_per_second=662.5 throughput/device/batches_per_second=0.0404 2026-07-27 14:52:05.792 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=407/600, eta=1 hour, 20 minutes] train/CrossEntropyLoss=0.0190 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9561 train/patch_acc=0.7188 train/subpatch_loss=0.0896 train/subpatch_acc=0.9375 train/location_loss=0.8081 train/location_acc=0.7500 train/token_losses=11.24 throughput/total_tokens=6,668,288 throughput/device/tokens_per_second=663.2 throughput/device/batches_per_second=0.0405 2026-07-27 14:52:30.334 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=408/600, eta=1 hour, 19 minutes] train/CrossEntropyLoss=0.0848 train/ZLoss=0.0001 train/Accuracy=0.9787 train/patch_loss=0.6812 train/patch_acc=0.8125 train/subpatch_loss=0.1229 train/subpatch_acc=1.0000 train/location_loss=0.9424 train/location_acc=0.6250 train/token_losses=9.710 throughput/total_tokens=6,684,672 throughput/device/tokens_per_second=663.7 throughput/device/batches_per_second=0.0405 2026-07-27 14:52:54.896 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=409/600, eta=1 hour, 19 minutes] train/CrossEntropyLoss=0.1862 train/ZLoss=0.0007 train/Accuracy=0.9482 train/patch_loss=1.079 train/patch_acc=0.7188 train/subpatch_loss=0.1028 train/subpatch_acc=1.0000 train/location_loss=0.9627 train/location_acc=0.6250 train/token_losses=12.89 throughput/total_tokens=6,701,056 throughput/device/tokens_per_second=664.1 throughput/device/batches_per_second=0.0405 2026-07-27 14:53:19.456 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=410/600, eta=1 hour, 18 minutes] train/CrossEntropyLoss=0.0375 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8700 train/patch_acc=0.7188 train/subpatch_loss=0.1005 train/subpatch_acc=1.0000 train/location_loss=0.8973 train/location_acc=0.6875 train/token_losses=10.95 throughput/total_tokens=6,717,440 throughput/device/tokens_per_second=664.4 throughput/device/batches_per_second=0.0406 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:53:44.035 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=411/600, eta=1 hour, 18 minutes] train/CrossEntropyLoss=0.0341 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9256 train/patch_acc=0.6562 train/subpatch_loss=0.1408 train/subpatch_acc=0.9375 train/location_loss=1.077 train/location_acc=0.5625 train/token_losses=12.28 throughput/total_tokens=6,733,824 throughput/device/tokens_per_second=664.6 throughput/device/batches_per_second=0.0406 2026-07-27 14:54:08.622 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=412/600, eta=1 hour, 18 minutes] train/CrossEntropyLoss=0.0505 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.7329 train/patch_acc=0.7188 train/subpatch_loss=0.0981 train/subpatch_acc=1.0000 train/location_loss=0.8506 train/location_acc=0.7500 train/token_losses=9.658 throughput/total_tokens=6,750,208 throughput/device/tokens_per_second=664.7 throughput/device/batches_per_second=0.0406 2026-07-27 14:54:33.190 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=413/600, eta=1 hour, 17 minutes] train/CrossEntropyLoss=0.0158 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6177 train/patch_acc=0.7812 train/subpatch_loss=0.0502 train/subpatch_acc=1.0000 train/location_loss=0.8859 train/location_acc=0.6875 train/token_losses=8.686 throughput/total_tokens=6,766,592 throughput/device/tokens_per_second=664.9 throughput/device/batches_per_second=0.0406 2026-07-27 14:54:57.723 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=414/600, eta=1 hour, 17 minutes] train/CrossEntropyLoss=0.2676 train/ZLoss=0.0006 train/Accuracy=0.9390 train/patch_loss=1.181 train/patch_acc=0.6875 train/subpatch_loss=0.2297 train/subpatch_acc=0.8750 train/location_loss=1.111 train/location_acc=0.5625 train/token_losses=14.81 throughput/total_tokens=6,782,976 throughput/device/tokens_per_second=665.1 throughput/device/batches_per_second=0.0406 2026-07-27 14:55:22.287 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=415/600, eta=1 hour, 16 minutes] train/CrossEntropyLoss=0.0099 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9563 train/patch_acc=0.7188 train/subpatch_loss=0.0995 train/subpatch_acc=1.0000 train/location_loss=0.7696 train/location_acc=0.7500 train/token_losses=11.13 throughput/total_tokens=6,799,360 throughput/device/tokens_per_second=665.2 throughput/device/batches_per_second=0.0406 2026-07-27 14:55:46.845 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=416/600, eta=1 hour, 16 minutes] train/CrossEntropyLoss=0.0063 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6717 train/patch_acc=0.8125 train/subpatch_loss=0.0924 train/subpatch_acc=1.0000 train/location_loss=0.7291 train/location_acc=0.8750 train/token_losses=8.660 throughput/total_tokens=6,815,744 throughput/device/tokens_per_second=665.4 throughput/device/batches_per_second=0.0406 2026-07-27 14:56:11.410 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=417/600, eta=1 hour, 15 minutes] train/CrossEntropyLoss=0.0382 train/ZLoss=0.0006 train/Accuracy=0.9909 train/patch_loss=0.6628 train/patch_acc=0.8438 train/subpatch_loss=0.2013 train/subpatch_acc=0.9375 train/location_loss=1.129 train/location_acc=0.6250 train/token_losses=10.62 throughput/total_tokens=6,832,128 throughput/device/tokens_per_second=665.5 throughput/device/batches_per_second=0.0406 2026-07-27 14:56:35.974 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=418/600, eta=1 hour, 15 minutes] train/CrossEntropyLoss=0.0131 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9146 train/patch_acc=0.7500 train/subpatch_loss=0.0661 train/subpatch_acc=1.0000 train/location_loss=0.9992 train/location_acc=0.6250 train/token_losses=11.58 throughput/total_tokens=6,848,512 throughput/device/tokens_per_second=665.5 throughput/device/batches_per_second=0.0406 2026-07-27 14:57:00.543 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=419/600, eta=1 hour, 15 minutes] train/CrossEntropyLoss=0.0223 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7061 train/patch_acc=0.8125 train/subpatch_loss=0.1429 train/subpatch_acc=0.9375 train/location_loss=0.8645 train/location_acc=0.7500 train/token_losses=9.678 throughput/total_tokens=6,864,896 throughput/device/tokens_per_second=665.6 throughput/device/batches_per_second=0.0406 2026-07-27 14:57:25.103 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=420/600, eta=1 hour, 14 minutes] train/CrossEntropyLoss=0.0543 train/ZLoss=0.0006 train/Accuracy=0.9817 train/patch_loss=0.8363 train/patch_acc=0.7812 train/subpatch_loss=0.0411 train/subpatch_acc=1.0000 train/location_loss=0.7516 train/location_acc=0.8750 train/token_losses=9.861 throughput/total_tokens=6,881,280 throughput/device/tokens_per_second=665.7 throughput/device/batches_per_second=0.0406 System/Peak GPU Memory (MB)=41,776 2026-07-27 14:57:49.657 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=421/600, eta=1 hour, 14 minutes] train/CrossEntropyLoss=0.0480 train/ZLoss=0.0008 train/Accuracy=0.9848 train/patch_loss=0.8301 train/patch_acc=0.7812 train/subpatch_loss=0.0961 train/subpatch_acc=0.9375 train/location_loss=0.9265 train/location_acc=0.7500 train/token_losses=10.73 throughput/total_tokens=6,897,664 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:58:14.191 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=422/600, eta=1 hour, 13 minutes] train/CrossEntropyLoss=0.0568 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.5608 train/patch_acc=0.7812 train/subpatch_loss=0.1055 train/subpatch_acc=1.0000 train/location_loss=0.9370 train/location_acc=0.6875 train/token_losses=8.657 throughput/total_tokens=6,914,048 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 14:58:38.759 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=423/600, eta=1 hour, 13 minutes] train/CrossEntropyLoss=0.0256 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7156 train/patch_acc=0.7812 train/subpatch_loss=0.3009 train/subpatch_acc=0.8750 train/location_loss=1.002 train/location_acc=0.6875 train/token_losses=10.94 throughput/total_tokens=6,930,432 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:59:03.348 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=424/600, eta=1 hour, 13 minutes] train/CrossEntropyLoss=0.1453 train/ZLoss=0.0008 train/Accuracy=0.9573 train/patch_loss=0.8683 train/patch_acc=0.7812 train/subpatch_loss=0.1113 train/subpatch_acc=1.0000 train/location_loss=0.8833 train/location_acc=0.6875 train/token_losses=10.93 throughput/total_tokens=6,946,816 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 14:59:27.943 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=425/600, eta=1 hour, 12 minutes] train/CrossEntropyLoss=0.0115 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.7304 train/patch_acc=0.8125 train/subpatch_loss=0.0412 train/subpatch_acc=1.0000 train/location_loss=0.9847 train/location_acc=0.6250 train/token_losses=9.947 throughput/total_tokens=6,963,200 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 14:59:52.519 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=426/600, eta=1 hour, 12 minutes] train/CrossEntropyLoss=0.0019 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.7911 train/patch_acc=0.7188 train/subpatch_loss=0.0665 train/subpatch_acc=1.0000 train/location_loss=0.7590 train/location_acc=0.6875 train/token_losses=9.631 throughput/total_tokens=6,979,584 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:00:17.090 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=427/600, eta=1 hour, 11 minutes] train/CrossEntropyLoss=0.0374 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6778 train/patch_acc=0.7812 train/subpatch_loss=0.1775 train/subpatch_acc=0.8750 train/location_loss=1.057 train/location_acc=0.6250 train/token_losses=10.36 throughput/total_tokens=6,995,968 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:00:41.718 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=428/600, eta=1 hour, 11 minutes] train/CrossEntropyLoss=0.0578 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9269 train/patch_acc=0.7500 train/subpatch_loss=0.1018 train/subpatch_acc=1.0000 train/location_loss=0.8423 train/location_acc=0.8125 train/token_losses=11.19 throughput/total_tokens=7,012,352 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:01:06.294 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=429/600, eta=1 hour, 11 minutes] train/CrossEntropyLoss=0.0347 train/ZLoss=0.0005 train/Accuracy=0.9878 train/patch_loss=0.6049 train/patch_acc=0.8125 train/subpatch_loss=0.0722 train/subpatch_acc=1.0000 train/location_loss=0.7160 train/location_acc=0.9375 train/token_losses=7.992 throughput/total_tokens=7,028,736 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 15:01:30.882 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=430/600, eta=1 hour, 10 minutes] train/CrossEntropyLoss=0.0219 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.6645 train/patch_acc=0.7812 train/subpatch_loss=0.0737 train/subpatch_acc=1.0000 train/location_loss=0.8201 train/location_acc=0.8750 train/token_losses=8.891 throughput/total_tokens=7,045,120 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:01:55.444 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=431/600, eta=1 hour, 10 minutes] train/CrossEntropyLoss=0.0564 train/ZLoss=0.0005 train/Accuracy=0.9878 train/patch_loss=0.7289 train/patch_acc=0.8125 train/subpatch_loss=0.1728 train/subpatch_acc=0.9375 train/location_loss=1.459 train/location_acc=0.3750 train/token_losses=12.36 throughput/total_tokens=7,061,504 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 15:02:19.989 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=432/600, eta=1 hour, 9 minutes] train/CrossEntropyLoss=0.0257 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.000 train/patch_acc=0.7500 train/subpatch_loss=0.1077 train/subpatch_acc=1.0000 train/location_loss=1.053 train/location_acc=0.6875 train/token_losses=12.65 throughput/total_tokens=7,077,888 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:02:44.536 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=433/600, eta=1 hour, 9 minutes] train/CrossEntropyLoss=0.0047 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8599 train/patch_acc=0.6875 train/subpatch_loss=0.1220 train/subpatch_acc=0.9375 train/location_loss=1.061 train/location_acc=0.5625 train/token_losses=11.61 throughput/total_tokens=7,094,272 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:03:09.126 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=434/600, eta=1 hour, 9 minutes] train/CrossEntropyLoss=0.0697 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.6423 train/patch_acc=0.7812 train/subpatch_loss=0.1098 train/subpatch_acc=0.9375 train/location_loss=1.083 train/location_acc=0.5000 train/token_losses=9.910 throughput/total_tokens=7,110,656 throughput/device/tokens_per_second=666.8 throughput/device/batches_per_second=0.0407 2026-07-27 15:03:33.663 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=435/600, eta=1 hour, 8 minutes] train/CrossEntropyLoss=0.0068 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.9622 train/patch_acc=0.7188 train/subpatch_loss=0.2504 train/subpatch_acc=0.9375 train/location_loss=1.088 train/location_acc=0.5625 train/token_losses=13.05 throughput/total_tokens=7,127,040 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:03:58.198 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=436/600, eta=1 hour, 8 minutes] train/CrossEntropyLoss=0.0082 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8694 train/patch_acc=0.8125 train/subpatch_loss=0.0515 train/subpatch_acc=1.0000 train/location_loss=0.9557 train/location_acc=0.6250 train/token_losses=10.98 throughput/total_tokens=7,143,424 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:04:22.740 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=437/600, eta=1 hour, 7 minutes] train/CrossEntropyLoss=0.0663 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.5832 train/patch_acc=0.7500 train/subpatch_loss=0.1038 train/subpatch_acc=1.0000 train/location_loss=0.8387 train/location_acc=0.6875 train/token_losses=8.436 throughput/total_tokens=7,159,808 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:04:47.289 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=438/600, eta=1 hour, 7 minutes] train/CrossEntropyLoss=0.0127 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.9841 train/patch_acc=0.6875 train/subpatch_loss=0.1270 train/subpatch_acc=1.0000 train/location_loss=1.049 train/location_acc=0.6250 train/token_losses=12.58 throughput/total_tokens=7,176,192 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:05:11.835 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=439/600, eta=1 hour, 6 minutes] train/CrossEntropyLoss=0.0056 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6356 train/patch_acc=0.9062 train/subpatch_loss=0.0818 train/subpatch_acc=1.0000 train/location_loss=0.7644 train/location_acc=0.8750 train/token_losses=8.470 throughput/total_tokens=7,192,576 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:05:36.400 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=440/600, eta=1 hour, 6 minutes] train/CrossEntropyLoss=0.3321 train/ZLoss=0.0010 train/Accuracy=0.9085 train/patch_loss=0.8960 train/patch_acc=0.6875 train/subpatch_loss=0.1883 train/subpatch_acc=0.9375 train/location_loss=1.058 train/location_acc=0.5625 train/token_losses=12.15 throughput/total_tokens=7,208,960 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:06:00.959 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=441/600, eta=1 hour, 6 minutes] train/CrossEntropyLoss=0.0229 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.6184 train/patch_acc=0.8750 train/subpatch_loss=0.0747 train/subpatch_acc=1.0000 train/location_loss=1.031 train/location_acc=0.6250 train/token_losses=9.369 throughput/total_tokens=7,225,344 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:06:25.506 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=442/600, eta=1 hour, 5 minutes] train/CrossEntropyLoss=0.0602 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.6613 train/patch_acc=0.8750 train/subpatch_loss=0.2190 train/subpatch_acc=0.9375 train/location_loss=0.9787 train/location_acc=0.8125 train/token_losses=10.08 throughput/total_tokens=7,241,728 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:06:50.048 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=443/600, eta=1 hour, 5 minutes] train/CrossEntropyLoss=0.0790 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.8992 train/patch_acc=0.6875 train/subpatch_loss=0.0985 train/subpatch_acc=1.0000 train/location_loss=0.7601 train/location_acc=0.7500 train/token_losses=10.63 throughput/total_tokens=7,258,112 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:07:14.579 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=444/600, eta=1 hour, 4 minutes] train/CrossEntropyLoss=0.0375 train/ZLoss=0.0001 train/Accuracy=0.9787 train/patch_loss=0.8333 train/patch_acc=0.7188 train/subpatch_loss=0.0927 train/subpatch_acc=1.0000 train/location_loss=1.112 train/location_acc=0.5625 train/token_losses=11.48 throughput/total_tokens=7,274,496 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:07:39.157 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=445/600, eta=1 hour, 4 minutes] train/CrossEntropyLoss=0.1293 train/ZLoss=0.0006 train/Accuracy=0.9695 train/patch_loss=0.8254 train/patch_acc=0.8125 train/subpatch_loss=0.2270 train/subpatch_acc=0.9375 train/location_loss=1.336 train/location_acc=0.5625 train/token_losses=12.85 throughput/total_tokens=7,290,880 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:08:03.695 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=446/600, eta=1 hour, 4 minutes] train/CrossEntropyLoss=0.0347 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.4632 train/patch_acc=0.8750 train/subpatch_loss=0.0674 train/subpatch_acc=1.0000 train/location_loss=0.5818 train/location_acc=0.9375 train/token_losses=6.303 throughput/total_tokens=7,307,264 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:08:28.237 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=447/600, eta=1 hour, 3 minutes] train/CrossEntropyLoss=0.0176 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.7406 train/patch_acc=0.7812 train/subpatch_loss=0.1186 train/subpatch_acc=0.9375 train/location_loss=0.8447 train/location_acc=0.8125 train/token_losses=9.778 throughput/total_tokens=7,323,648 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:08:52.769 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=448/600, eta=1 hour, 3 minutes] train/CrossEntropyLoss=0.0294 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.6940 train/patch_acc=0.7812 train/subpatch_loss=0.1023 train/subpatch_acc=1.0000 train/location_loss=0.9211 train/location_acc=0.6250 train/token_losses=9.646 throughput/total_tokens=7,340,032 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:09:17.300 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=449/600, eta=1 hour, 2 minutes] train/CrossEntropyLoss=0.0614 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.5965 train/patch_acc=0.8125 train/subpatch_loss=0.1180 train/subpatch_acc=1.0000 train/location_loss=1.155 train/location_acc=0.4375 train/token_losses=9.865 throughput/total_tokens=7,356,416 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:09:41.860 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=450/600, eta=1 hour, 2 minutes] train/CrossEntropyLoss=0.0306 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.7588 train/patch_acc=0.7812 train/subpatch_loss=0.1281 train/subpatch_acc=1.0000 train/location_loss=0.9021 train/location_acc=0.7500 train/token_losses=10.19 throughput/total_tokens=7,372,800 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:10:06.433 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=451/600, eta=1 hour, 2 minutes] train/CrossEntropyLoss=0.0345 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.8339 train/patch_acc=0.6875 train/subpatch_loss=0.1496 train/subpatch_acc=0.9375 train/location_loss=1.033 train/location_acc=0.8125 train/token_losses=11.40 throughput/total_tokens=7,389,184 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:10:30.990 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=452/600, eta=1 hour, 1 minute] train/CrossEntropyLoss=0.0242 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.8769 train/patch_acc=0.7500 train/subpatch_loss=0.0576 train/subpatch_acc=1.0000 train/location_loss=0.9490 train/location_acc=0.6875 train/token_losses=11.04 throughput/total_tokens=7,405,568 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:10:55.542 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=453/600, eta=1 hour, 1 minute] train/CrossEntropyLoss=0.0409 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7026 train/patch_acc=0.8125 train/subpatch_loss=0.1432 train/subpatch_acc=0.9375 train/location_loss=0.8377 train/location_acc=0.6875 train/token_losses=9.545 throughput/total_tokens=7,421,952 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:11:20.085 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=454/600, eta=1 hour] train/CrossEntropyLoss=0.0327 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7268 train/patch_acc=0.7500 train/subpatch_loss=0.0772 train/subpatch_acc=1.0000 train/location_loss=0.8477 train/location_acc=0.8125 train/token_losses=9.514 throughput/total_tokens=7,438,336 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:11:44.623 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=455/600, eta=1 hour] train/CrossEntropyLoss=0.0580 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.8193 train/patch_acc=0.7812 train/subpatch_loss=0.1486 train/subpatch_acc=0.9375 train/location_loss=0.7605 train/location_acc=0.6875 train/token_losses=10.19 throughput/total_tokens=7,454,720 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:12:09.187 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=456/600, eta=1 hour] train/CrossEntropyLoss=0.0027 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.9614 train/patch_acc=0.7812 train/subpatch_loss=0.0983 train/subpatch_acc=1.0000 train/location_loss=0.8565 train/location_acc=0.7500 train/token_losses=11.51 throughput/total_tokens=7,471,104 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:12:33.788 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=457/600, eta=59 minutes] train/CrossEntropyLoss=0.0166 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=1.001 train/patch_acc=0.6562 train/subpatch_loss=0.1108 train/subpatch_acc=1.0000 train/location_loss=1.045 train/location_acc=0.6250 train/token_losses=12.64 throughput/total_tokens=7,487,488 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:12:58.361 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=458/600, eta=59 minutes] train/CrossEntropyLoss=0.0049 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.7869 train/patch_acc=0.6875 train/subpatch_loss=0.0946 train/subpatch_acc=1.0000 train/location_loss=0.9444 train/location_acc=0.6250 train/token_losses=10.45 throughput/total_tokens=7,503,872 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:13:22.926 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=459/600, eta=58 minutes] train/CrossEntropyLoss=0.0038 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.4307 train/patch_acc=0.8750 train/subpatch_loss=0.0656 train/subpatch_acc=1.0000 train/location_loss=0.7051 train/location_acc=0.8125 train/token_losses=6.528 throughput/total_tokens=7,520,256 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:13:47.489 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=460/600, eta=58 minutes] train/CrossEntropyLoss=0.1728 train/ZLoss=0.0007 train/Accuracy=0.9604 train/patch_loss=0.9205 train/patch_acc=0.7500 train/subpatch_loss=0.1514 train/subpatch_acc=0.9375 train/location_loss=1.077 train/location_acc=0.7500 train/token_losses=12.28 throughput/total_tokens=7,536,640 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:14:12.049 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=461/600, eta=57 minutes] train/CrossEntropyLoss=0.0371 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.038 train/patch_acc=0.6250 train/subpatch_loss=0.1419 train/subpatch_acc=0.9375 train/location_loss=0.9045 train/location_acc=0.6250 train/token_losses=12.49 throughput/total_tokens=7,553,024 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:14:36.598 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=462/600, eta=57 minutes] train/CrossEntropyLoss=0.0276 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8621 train/patch_acc=0.7500 train/subpatch_loss=0.1333 train/subpatch_acc=0.9375 train/location_loss=0.9416 train/location_acc=0.6250 train/token_losses=11.20 throughput/total_tokens=7,569,408 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:15:01.158 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=463/600, eta=57 minutes] train/CrossEntropyLoss=0.0478 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.7947 train/patch_acc=0.7812 train/subpatch_loss=0.0904 train/subpatch_acc=1.0000 train/location_loss=0.8288 train/location_acc=0.7500 train/token_losses=10.03 throughput/total_tokens=7,585,792 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:15:25.736 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=464/600, eta=56 minutes] train/CrossEntropyLoss=0.0226 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.8882 train/patch_acc=0.7500 train/subpatch_loss=0.1084 train/subpatch_acc=1.0000 train/location_loss=0.8556 train/location_acc=0.8125 train/token_losses=10.96 throughput/total_tokens=7,602,176 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:15:50.291 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=465/600, eta=56 minutes] train/CrossEntropyLoss=0.0402 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6907 train/patch_acc=0.7812 train/subpatch_loss=0.0962 train/subpatch_acc=1.0000 train/location_loss=0.8174 train/location_acc=0.7500 train/token_losses=9.180 throughput/total_tokens=7,618,560 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:16:14.854 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=466/600, eta=55 minutes] train/CrossEntropyLoss=0.1558 train/ZLoss=0.0007 train/Accuracy=0.9604 train/patch_loss=0.7580 train/patch_acc=0.7188 train/subpatch_loss=0.1060 train/subpatch_acc=1.0000 train/location_loss=0.9698 train/location_acc=0.6875 train/token_losses=10.37 throughput/total_tokens=7,634,944 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:16:39.415 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=467/600, eta=55 minutes] train/CrossEntropyLoss=0.0583 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.5193 train/patch_acc=0.8750 train/subpatch_loss=0.0745 train/subpatch_acc=1.0000 train/location_loss=0.9777 train/location_acc=0.6875 train/token_losses=8.363 throughput/total_tokens=7,651,328 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:17:03.973 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=468/600, eta=55 minutes] train/CrossEntropyLoss=0.0461 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.9686 train/patch_acc=0.7188 train/subpatch_loss=0.0884 train/subpatch_acc=1.0000 train/location_loss=0.7868 train/location_acc=0.7500 train/token_losses=11.25 throughput/total_tokens=7,667,712 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:17:28.540 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=469/600, eta=54 minutes] train/CrossEntropyLoss=0.0054 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.5542 train/patch_acc=0.8750 train/subpatch_loss=0.0796 train/subpatch_acc=1.0000 train/location_loss=0.5957 train/location_acc=0.8750 train/token_losses=7.134 throughput/total_tokens=7,684,096 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:17:53.090 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=470/600, eta=54 minutes] train/CrossEntropyLoss=0.0435 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.7890 train/patch_acc=0.8125 train/subpatch_loss=0.1305 train/subpatch_acc=1.0000 train/location_loss=1.140 train/location_acc=0.5000 train/token_losses=11.39 throughput/total_tokens=7,700,480 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:18:17.642 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=471/600, eta=53 minutes] train/CrossEntropyLoss=0.0325 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.5993 train/patch_acc=0.9375 train/subpatch_loss=0.0540 train/subpatch_acc=1.0000 train/location_loss=1.203 train/location_acc=0.6250 train/token_losses=9.821 throughput/total_tokens=7,716,864 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:18:42.219 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=472/600, eta=53 minutes] train/CrossEntropyLoss=0.0212 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7179 train/patch_acc=0.8125 train/subpatch_loss=0.0299 train/subpatch_acc=1.0000 train/location_loss=0.8080 train/location_acc=0.6875 train/token_losses=9.095 throughput/total_tokens=7,733,248 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:19:06.792 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=473/600, eta=53 minutes] train/CrossEntropyLoss=0.0136 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=1.002 train/patch_acc=0.7188 train/subpatch_loss=0.0897 train/subpatch_acc=1.0000 train/location_loss=0.8574 train/location_acc=0.6250 train/token_losses=11.80 throughput/total_tokens=7,749,632 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:19:31.365 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=474/600, eta=52 minutes] train/CrossEntropyLoss=0.0164 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.7197 train/patch_acc=0.7500 train/subpatch_loss=0.0503 train/subpatch_acc=1.0000 train/location_loss=0.9600 train/location_acc=0.6250 train/token_losses=9.799 throughput/total_tokens=7,766,016 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:19:55.951 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=475/600, eta=52 minutes] train/CrossEntropyLoss=0.0589 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7920 train/patch_acc=0.8438 train/subpatch_loss=0.0775 train/subpatch_acc=1.0000 train/location_loss=0.8149 train/location_acc=0.7500 train/token_losses=9.906 throughput/total_tokens=7,782,400 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:20:20.534 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=476/600, eta=51 minutes] train/CrossEntropyLoss=0.0092 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6888 train/patch_acc=0.7812 train/subpatch_loss=0.1061 train/subpatch_acc=1.0000 train/location_loss=0.6831 train/location_acc=0.8750 train/token_losses=8.667 throughput/total_tokens=7,798,784 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:20:45.099 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=477/600, eta=51 minutes] train/CrossEntropyLoss=0.0660 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.5519 train/patch_acc=0.8438 train/subpatch_loss=0.1082 train/subpatch_acc=1.0000 train/location_loss=0.6813 train/location_acc=0.8750 train/token_losses=7.573 throughput/total_tokens=7,815,168 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:21:09.653 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=478/600, eta=50 minutes] train/CrossEntropyLoss=0.0583 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.6171 train/patch_acc=0.8750 train/subpatch_loss=0.1035 train/subpatch_acc=1.0000 train/location_loss=0.9645 train/location_acc=0.7500 train/token_losses=9.209 throughput/total_tokens=7,831,552 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:21:34.222 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=479/600, eta=50 minutes] train/CrossEntropyLoss=0.1175 train/ZLoss=0.0008 train/Accuracy=0.9604 train/patch_loss=0.6898 train/patch_acc=0.7812 train/subpatch_loss=0.0759 train/subpatch_acc=0.9375 train/location_loss=0.9610 train/location_acc=0.7500 train/token_losses=9.666 throughput/total_tokens=7,847,936 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:21:58.791 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=480/600, eta=50 minutes] train/CrossEntropyLoss=0.0368 train/ZLoss=0.0006 train/Accuracy=0.9909 train/patch_loss=0.6523 train/patch_acc=0.8438 train/subpatch_loss=0.0869 train/subpatch_acc=1.0000 train/location_loss=0.9042 train/location_acc=0.7500 train/token_losses=9.183 throughput/total_tokens=7,864,320 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:22:23.354 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=481/600, eta=49 minutes] train/CrossEntropyLoss=0.0485 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.6479 train/patch_acc=0.7812 train/subpatch_loss=0.1167 train/subpatch_acc=0.9375 train/location_loss=0.9887 train/location_acc=0.6250 train/token_losses=9.605 throughput/total_tokens=7,880,704 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:22:47.911 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=482/600, eta=49 minutes] train/CrossEntropyLoss=0.0410 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.6231 train/patch_acc=0.8125 train/subpatch_loss=0.0521 train/subpatch_acc=1.0000 train/location_loss=0.9455 train/location_acc=0.6250 train/token_losses=8.975 throughput/total_tokens=7,897,088 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:23:12.516 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=483/600, eta=48 minutes] train/CrossEntropyLoss=0.0401 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.6249 train/patch_acc=0.7812 train/subpatch_loss=0.1628 train/subpatch_acc=0.9375 train/location_loss=0.7639 train/location_acc=0.6875 train/token_losses=8.706 throughput/total_tokens=7,913,472 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:23:37.071 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=484/600, eta=48 minutes] train/CrossEntropyLoss=0.0573 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.9664 train/patch_acc=0.7500 train/subpatch_loss=0.0913 train/subpatch_acc=1.0000 train/location_loss=1.073 train/location_acc=0.5000 train/token_losses=12.39 throughput/total_tokens=7,929,856 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:24:01.623 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=485/600, eta=48 minutes] train/CrossEntropyLoss=0.1525 train/ZLoss=0.0006 train/Accuracy=0.9634 train/patch_loss=0.9878 train/patch_acc=0.7500 train/subpatch_loss=0.1693 train/subpatch_acc=0.9375 train/location_loss=1.109 train/location_acc=0.6250 train/token_losses=13.02 throughput/total_tokens=7,946,240 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:24:26.173 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=486/600, eta=47 minutes] train/CrossEntropyLoss=0.0086 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8732 train/patch_acc=0.7812 train/subpatch_loss=0.1046 train/subpatch_acc=1.0000 train/location_loss=1.093 train/location_acc=0.4375 train/token_losses=11.78 throughput/total_tokens=7,962,624 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:24:50.742 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=487/600, eta=47 minutes] train/CrossEntropyLoss=0.0226 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.5989 train/patch_acc=0.9375 train/subpatch_loss=0.0600 train/subpatch_acc=1.0000 train/location_loss=0.8121 train/location_acc=0.7500 train/token_losses=8.280 throughput/total_tokens=7,979,008 throughput/device/tokens_per_second=666.9 throughput/device/batches_per_second=0.0407 2026-07-27 15:25:15.273 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=488/600, eta=46 minutes] train/CrossEntropyLoss=0.0228 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.4839 train/patch_acc=0.8750 train/subpatch_loss=0.0589 train/subpatch_acc=1.0000 train/location_loss=0.5779 train/location_acc=0.7500 train/token_losses=6.418 throughput/total_tokens=7,995,392 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:25:39.817 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=489/600, eta=46 minutes] train/CrossEntropyLoss=0.0292 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.001 train/patch_acc=0.6875 train/subpatch_loss=0.0956 train/subpatch_acc=1.0000 train/location_loss=1.050 train/location_acc=0.6250 train/token_losses=12.59 throughput/total_tokens=8,011,776 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:26:04.370 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=490/600, eta=46 minutes] train/CrossEntropyLoss=0.0770 train/ZLoss=0.0001 train/Accuracy=0.9817 train/patch_loss=0.6076 train/patch_acc=0.8750 train/subpatch_loss=0.1037 train/subpatch_acc=1.0000 train/location_loss=0.5755 train/location_acc=0.9375 train/token_losses=7.577 throughput/total_tokens=8,028,160 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:26:28.918 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=491/600, eta=45 minutes] train/CrossEntropyLoss=0.0381 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.8913 train/patch_acc=0.7812 train/subpatch_loss=0.1071 train/subpatch_acc=1.0000 train/location_loss=1.145 train/location_acc=0.5625 train/token_losses=12.14 throughput/total_tokens=8,044,544 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:26:53.460 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=492/600, eta=45 minutes] train/CrossEntropyLoss=0.0198 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9596 train/patch_acc=0.7500 train/subpatch_loss=0.0964 train/subpatch_acc=1.0000 train/location_loss=0.9559 train/location_acc=0.6875 train/token_losses=11.89 throughput/total_tokens=8,060,928 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:27:18.003 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=493/600, eta=44 minutes] train/CrossEntropyLoss=0.0295 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9356 train/patch_acc=0.7812 train/subpatch_loss=0.0889 train/subpatch_acc=1.0000 train/location_loss=0.8212 train/location_acc=0.6875 train/token_losses=11.13 throughput/total_tokens=8,077,312 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:27:42.590 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=494/600, eta=44 minutes] train/CrossEntropyLoss=0.0315 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.6014 train/patch_acc=0.8125 train/subpatch_loss=0.1227 train/subpatch_acc=1.0000 train/location_loss=0.6589 train/location_acc=0.9375 train/token_losses=7.937 throughput/total_tokens=8,093,696 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:28:07.126 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=495/600, eta=44 minutes] train/CrossEntropyLoss=0.0489 train/ZLoss=0.0005 train/Accuracy=0.9909 train/patch_loss=0.9568 train/patch_acc=0.7500 train/subpatch_loss=0.1692 train/subpatch_acc=0.8750 train/location_loss=1.085 train/location_acc=0.5625 train/token_losses=12.67 throughput/total_tokens=8,110,080 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:28:31.673 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=496/600, eta=43 minutes] train/CrossEntropyLoss=0.0311 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.9590 train/patch_acc=0.7188 train/subpatch_loss=0.0493 train/subpatch_acc=1.0000 train/location_loss=1.021 train/location_acc=0.7500 train/token_losses=11.95 throughput/total_tokens=8,126,464 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:28:56.225 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=497/600, eta=43 minutes] train/CrossEntropyLoss=0.0148 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6332 train/patch_acc=0.8125 train/subpatch_loss=0.1208 train/subpatch_acc=0.9375 train/location_loss=0.6475 train/location_acc=0.8750 train/token_losses=8.139 throughput/total_tokens=8,142,848 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:29:20.793 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=498/600, eta=42 minutes] train/CrossEntropyLoss=0.0575 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6437 train/patch_acc=0.8125 train/subpatch_loss=0.0533 train/subpatch_acc=1.0000 train/location_loss=0.7760 train/location_acc=0.8125 train/token_losses=8.467 throughput/total_tokens=8,159,232 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:29:45.347 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=499/600, eta=42 minutes] train/CrossEntropyLoss=0.0286 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6729 train/patch_acc=0.9062 train/subpatch_loss=0.0751 train/subpatch_acc=1.0000 train/location_loss=1.006 train/location_acc=0.6875 train/token_losses=9.706 throughput/total_tokens=8,175,616 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:30:09.905 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=500/600, eta=41 minutes] train/CrossEntropyLoss=0.0048 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.8884 train/patch_acc=0.7812 train/subpatch_loss=0.0320 train/subpatch_acc=1.0000 train/location_loss=1.028 train/location_acc=0.3750 train/token_losses=11.35 throughput/total_tokens=8,192,000 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:30:09.907 2012bc4f83ae:0 olmo.train.trainer:1415 INFO Saving checkpoint... 2026-07-27 15:30:13.606 2012bc4f83ae:0 olmo.train.trainer:1420 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step500 2026-07-27 15:30:39.852 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=501/600, eta=41 minutes] train/CrossEntropyLoss=0.0298 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.5618 train/patch_acc=0.8438 train/subpatch_loss=0.1336 train/subpatch_acc=0.9375 train/location_loss=0.7197 train/location_acc=0.8750 train/token_losses=7.907 throughput/total_tokens=8,208,384 throughput/device/tokens_per_second=624.4 throughput/device/batches_per_second=0.0381 2026-07-27 15:31:04.373 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=502/600, eta=41 minutes] train/CrossEntropyLoss=0.0265 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.6853 train/patch_acc=0.8125 train/subpatch_loss=0.1202 train/subpatch_acc=1.0000 train/location_loss=0.6040 train/location_acc=1.0000 train/token_losses=8.379 throughput/total_tokens=8,224,768 throughput/device/tokens_per_second=645.5 throughput/device/batches_per_second=0.0394 2026-07-27 15:31:28.900 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=503/600, eta=40 minutes] train/CrossEntropyLoss=0.0302 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7491 train/patch_acc=0.7188 train/subpatch_loss=0.1454 train/subpatch_acc=0.9375 train/location_loss=0.8142 train/location_acc=0.8125 train/token_losses=9.831 throughput/total_tokens=8,241,152 throughput/device/tokens_per_second=652.8 throughput/device/batches_per_second=0.0398 2026-07-27 15:31:53.432 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=504/600, eta=40 minutes] train/CrossEntropyLoss=0.0172 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.6600 train/patch_acc=0.7812 train/subpatch_loss=0.0967 train/subpatch_acc=1.0000 train/location_loss=0.8456 train/location_acc=0.8125 train/token_losses=9.049 throughput/total_tokens=8,257,536 throughput/device/tokens_per_second=656.5 throughput/device/batches_per_second=0.0401 2026-07-27 15:32:18.006 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=505/600, eta=39 minutes] train/CrossEntropyLoss=0.0473 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.7885 train/patch_acc=0.7812 train/subpatch_loss=0.0548 train/subpatch_acc=1.0000 train/location_loss=0.6937 train/location_acc=0.8125 train/token_losses=9.302 throughput/total_tokens=8,273,920 throughput/device/tokens_per_second=658.5 throughput/device/batches_per_second=0.0402 2026-07-27 15:32:42.561 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=506/600, eta=39 minutes] train/CrossEntropyLoss=0.0341 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.5657 train/patch_acc=0.8125 train/subpatch_loss=0.0714 train/subpatch_acc=1.0000 train/location_loss=0.6653 train/location_acc=0.8750 train/token_losses=7.472 throughput/total_tokens=8,290,304 throughput/device/tokens_per_second=660.0 throughput/device/batches_per_second=0.0403 2026-07-27 15:33:07.125 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=507/600, eta=39 minutes] train/CrossEntropyLoss=0.0285 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.7445 train/patch_acc=0.8125 train/subpatch_loss=0.1115 train/subpatch_acc=1.0000 train/location_loss=0.6832 train/location_acc=0.7500 train/token_losses=9.134 throughput/total_tokens=8,306,688 throughput/device/tokens_per_second=661.0 throughput/device/batches_per_second=0.0403 2026-07-27 15:33:31.667 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=508/600, eta=38 minutes] train/CrossEntropyLoss=0.0214 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.6137 train/patch_acc=0.7500 train/subpatch_loss=0.0342 train/subpatch_acc=1.0000 train/location_loss=0.9023 train/location_acc=0.6875 train/token_losses=8.655 throughput/total_tokens=8,323,072 throughput/device/tokens_per_second=661.8 throughput/device/batches_per_second=0.0404 2026-07-27 15:33:56.214 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=509/600, eta=38 minutes] train/CrossEntropyLoss=0.0349 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.8265 train/patch_acc=0.7500 train/subpatch_loss=0.1254 train/subpatch_acc=0.9375 train/location_loss=0.9228 train/location_acc=0.7500 train/token_losses=10.81 throughput/total_tokens=8,339,456 throughput/device/tokens_per_second=662.4 throughput/device/batches_per_second=0.0404 2026-07-27 15:34:20.772 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=510/600, eta=37 minutes] train/CrossEntropyLoss=0.0092 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.7083 train/patch_acc=0.7812 train/subpatch_loss=0.0436 train/subpatch_acc=1.0000 train/location_loss=0.9967 train/location_acc=0.5625 train/token_losses=9.828 throughput/total_tokens=8,355,840 throughput/device/tokens_per_second=662.9 throughput/device/batches_per_second=0.0405 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:34:45.331 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=511/600, eta=37 minutes] train/CrossEntropyLoss=0.0293 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.8079 train/patch_acc=0.8125 train/subpatch_loss=0.1013 train/subpatch_acc=1.0000 train/location_loss=1.227 train/location_acc=0.4375 train/token_losses=11.77 throughput/total_tokens=8,372,224 throughput/device/tokens_per_second=663.3 throughput/device/batches_per_second=0.0405 2026-07-27 15:35:09.889 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=512/600, eta=37 minutes] train/CrossEntropyLoss=0.0482 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.6447 train/patch_acc=0.8438 train/subpatch_loss=0.1109 train/subpatch_acc=1.0000 train/location_loss=0.8628 train/location_acc=0.7500 train/token_losses=9.052 throughput/total_tokens=8,388,608 throughput/device/tokens_per_second=663.6 throughput/device/batches_per_second=0.0405 2026-07-27 15:35:34.446 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=513/600, eta=36 minutes] train/CrossEntropyLoss=0.0108 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.7362 train/patch_acc=0.8125 train/subpatch_loss=0.0396 train/subpatch_acc=1.0000 train/location_loss=0.7168 train/location_acc=0.8125 train/token_losses=8.916 throughput/total_tokens=8,404,992 throughput/device/tokens_per_second=663.9 throughput/device/batches_per_second=0.0405 2026-07-27 15:35:58.993 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=514/600, eta=36 minutes] train/CrossEntropyLoss=0.0511 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.9167 train/patch_acc=0.6875 train/subpatch_loss=0.0528 train/subpatch_acc=1.0000 train/location_loss=0.8715 train/location_acc=0.6250 train/token_losses=11.03 throughput/total_tokens=8,421,376 throughput/device/tokens_per_second=664.1 throughput/device/batches_per_second=0.0405 2026-07-27 15:36:23.534 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=515/600, eta=35 minutes] train/CrossEntropyLoss=0.0449 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.6894 train/patch_acc=0.7500 train/subpatch_loss=0.1657 train/subpatch_acc=0.8750 train/location_loss=0.8861 train/location_acc=0.6250 train/token_losses=9.722 throughput/total_tokens=8,437,760 throughput/device/tokens_per_second=664.4 throughput/device/batches_per_second=0.0405 2026-07-27 15:36:48.078 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=516/600, eta=35 minutes] train/CrossEntropyLoss=0.0473 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.4726 train/patch_acc=0.9688 train/subpatch_loss=0.0858 train/subpatch_acc=1.0000 train/location_loss=0.4710 train/location_acc=0.8750 train/token_losses=6.008 throughput/total_tokens=8,454,144 throughput/device/tokens_per_second=664.6 throughput/device/batches_per_second=0.0406 2026-07-27 15:37:12.638 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=517/600, eta=35 minutes] train/CrossEntropyLoss=0.0652 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.6961 train/patch_acc=0.8125 train/subpatch_loss=0.0855 train/subpatch_acc=1.0000 train/location_loss=0.7024 train/location_acc=0.7500 train/token_losses=8.721 throughput/total_tokens=8,470,528 throughput/device/tokens_per_second=664.7 throughput/device/batches_per_second=0.0406 2026-07-27 15:37:37.252 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=518/600, eta=34 minutes] train/CrossEntropyLoss=0.0173 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.9318 train/patch_acc=0.7812 train/subpatch_loss=0.1435 train/subpatch_acc=0.9375 train/location_loss=0.7206 train/location_acc=0.8750 train/token_losses=10.91 throughput/total_tokens=8,486,912 throughput/device/tokens_per_second=664.8 throughput/device/batches_per_second=0.0406 2026-07-27 15:38:01.833 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=519/600, eta=34 minutes] train/CrossEntropyLoss=0.2052 train/ZLoss=0.0006 train/Accuracy=0.9421 train/patch_loss=0.8302 train/patch_acc=0.7812 train/subpatch_loss=0.1161 train/subpatch_acc=1.0000 train/location_loss=0.9584 train/location_acc=0.7500 train/token_losses=10.94 throughput/total_tokens=8,503,296 throughput/device/tokens_per_second=664.8 throughput/device/batches_per_second=0.0406 2026-07-27 15:38:26.424 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=520/600, eta=33 minutes] train/CrossEntropyLoss=0.0277 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6004 train/patch_acc=0.8125 train/subpatch_loss=0.1963 train/subpatch_acc=1.0000 train/location_loss=0.7482 train/location_acc=0.8125 train/token_losses=8.581 throughput/total_tokens=8,519,680 throughput/device/tokens_per_second=664.9 throughput/device/batches_per_second=0.0406 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:38:51.026 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=521/600, eta=33 minutes] train/CrossEntropyLoss=0.0379 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.8750 train/patch_acc=0.7500 train/subpatch_loss=0.1209 train/subpatch_acc=1.0000 train/location_loss=0.8024 train/location_acc=0.7500 train/token_losses=10.69 throughput/total_tokens=8,536,064 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:39:15.585 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=522/600, eta=32 minutes] train/CrossEntropyLoss=0.0699 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.4185 train/patch_acc=0.9375 train/subpatch_loss=0.0486 train/subpatch_acc=1.0000 train/location_loss=0.8511 train/location_acc=0.7500 train/token_losses=6.947 throughput/total_tokens=8,552,448 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:39:40.138 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=523/600, eta=32 minutes] train/CrossEntropyLoss=0.0463 train/ZLoss=0.0006 train/Accuracy=0.9848 train/patch_loss=0.6065 train/patch_acc=0.8438 train/subpatch_loss=0.0800 train/subpatch_acc=0.9375 train/location_loss=0.9782 train/location_acc=0.8125 train/token_losses=9.085 throughput/total_tokens=8,568,832 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:40:04.691 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=524/600, eta=32 minutes] train/CrossEntropyLoss=0.0742 train/ZLoss=0.0002 train/Accuracy=0.9787 train/patch_loss=1.012 train/patch_acc=0.6875 train/subpatch_loss=0.1102 train/subpatch_acc=0.9375 train/location_loss=1.106 train/location_acc=0.5000 train/token_losses=12.96 throughput/total_tokens=8,585,216 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:40:29.235 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=525/600, eta=31 minutes] train/CrossEntropyLoss=0.0066 train/ZLoss=0.0004 train/Accuracy=0.9970 train/patch_loss=0.9182 train/patch_acc=0.7500 train/subpatch_loss=0.0637 train/subpatch_acc=1.0000 train/location_loss=1.071 train/location_acc=0.6875 train/token_losses=11.89 throughput/total_tokens=8,601,600 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:40:53.794 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=526/600, eta=31 minutes] train/CrossEntropyLoss=0.0230 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.6879 train/patch_acc=0.8438 train/subpatch_loss=0.1540 train/subpatch_acc=0.9375 train/location_loss=0.7884 train/location_acc=0.8750 train/token_losses=9.272 throughput/total_tokens=8,617,984 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:41:18.344 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=527/600, eta=30 minutes] train/CrossEntropyLoss=0.0093 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6435 train/patch_acc=0.8125 train/subpatch_loss=0.0718 train/subpatch_acc=1.0000 train/location_loss=0.8904 train/location_acc=0.6875 train/token_losses=8.997 throughput/total_tokens=8,634,368 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:41:42.925 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=528/600, eta=30 minutes] train/CrossEntropyLoss=0.0170 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.8956 train/patch_acc=0.7188 train/subpatch_loss=0.0724 train/subpatch_acc=1.0000 train/location_loss=0.5983 train/location_acc=0.8750 train/token_losses=9.848 throughput/total_tokens=8,650,752 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:42:07.480 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=529/600, eta=30 minutes] train/CrossEntropyLoss=0.0269 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.8112 train/patch_acc=0.7500 train/subpatch_loss=0.1180 train/subpatch_acc=1.0000 train/location_loss=0.7718 train/location_acc=0.9375 train/token_losses=10.05 throughput/total_tokens=8,667,136 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:42:32.034 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=530/600, eta=29 minutes] train/CrossEntropyLoss=0.0049 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.7034 train/patch_acc=0.7812 train/subpatch_loss=0.0238 train/subpatch_acc=1.0000 train/location_loss=0.6707 train/location_acc=0.8750 train/token_losses=8.406 throughput/total_tokens=8,683,520 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:42:56.608 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=531/600, eta=29 minutes] train/CrossEntropyLoss=0.0219 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.7132 train/patch_acc=0.7812 train/subpatch_loss=0.1099 train/subpatch_acc=1.0000 train/location_loss=1.006 train/location_acc=0.6250 train/token_losses=10.17 throughput/total_tokens=8,699,904 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:43:21.149 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=532/600, eta=28 minutes] train/CrossEntropyLoss=0.0385 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.5749 train/patch_acc=0.8438 train/subpatch_loss=0.0549 train/subpatch_acc=1.0000 train/location_loss=0.5667 train/location_acc=0.8750 train/token_losses=7.086 throughput/total_tokens=8,716,288 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:43:45.703 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=533/600, eta=28 minutes] train/CrossEntropyLoss=0.0185 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.5597 train/patch_acc=0.8438 train/subpatch_loss=0.0853 train/subpatch_acc=1.0000 train/location_loss=0.5794 train/location_acc=0.8750 train/token_losses=7.137 throughput/total_tokens=8,732,672 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:44:10.262 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=534/600, eta=28 minutes] train/CrossEntropyLoss=0.0991 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.7860 train/patch_acc=0.7188 train/subpatch_loss=0.1378 train/subpatch_acc=0.9375 train/location_loss=0.5340 train/location_acc=0.8750 train/token_losses=8.975 throughput/total_tokens=8,749,056 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:44:34.809 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=535/600, eta=27 minutes] train/CrossEntropyLoss=0.2089 train/ZLoss=0.0007 train/Accuracy=0.9512 train/patch_loss=0.8731 train/patch_acc=0.7188 train/subpatch_loss=0.1104 train/subpatch_acc=0.9375 train/location_loss=0.8998 train/location_acc=0.6875 train/token_losses=11.03 throughput/total_tokens=8,765,440 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:44:59.357 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=536/600, eta=27 minutes] train/CrossEntropyLoss=0.0233 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.4925 train/patch_acc=0.9375 train/subpatch_loss=0.0872 train/subpatch_acc=1.0000 train/location_loss=0.5886 train/location_acc=0.8750 train/token_losses=6.643 throughput/total_tokens=8,781,824 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:45:23.903 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=537/600, eta=26 minutes] train/CrossEntropyLoss=0.0520 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6404 train/patch_acc=0.7812 train/subpatch_loss=0.1417 train/subpatch_acc=0.9375 train/location_loss=0.5521 train/location_acc=0.9375 train/token_losses=7.898 throughput/total_tokens=8,798,208 throughput/device/tokens_per_second=667.0 throughput/device/batches_per_second=0.0407 2026-07-27 15:45:48.446 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=538/600, eta=26 minutes] train/CrossEntropyLoss=0.0360 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.9087 train/patch_acc=0.7812 train/subpatch_loss=0.0584 train/subpatch_acc=1.0000 train/location_loss=0.8110 train/location_acc=0.6875 train/token_losses=10.75 throughput/total_tokens=8,814,592 throughput/device/tokens_per_second=667.1 throughput/device/batches_per_second=0.0407 2026-07-27 15:46:13.006 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=539/600, eta=26 minutes] train/CrossEntropyLoss=0.0079 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.5850 train/patch_acc=0.8438 train/subpatch_loss=0.0461 train/subpatch_acc=1.0000 train/location_loss=0.7996 train/location_acc=0.7500 train/token_losses=8.063 throughput/total_tokens=8,830,976 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 2026-07-27 15:46:37.571 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=540/600, eta=25 minutes] train/CrossEntropyLoss=0.0304 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.8531 train/patch_acc=0.7812 train/subpatch_loss=0.1258 train/subpatch_acc=1.0000 train/location_loss=1.198 train/location_acc=0.6875 train/token_losses=12.12 throughput/total_tokens=8,847,360 throughput/device/tokens_per_second=667.2 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:47:02.105 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=541/600, eta=25 minutes] train/CrossEntropyLoss=0.0232 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.8486 train/patch_acc=0.7500 train/subpatch_loss=0.0609 train/subpatch_acc=1.0000 train/location_loss=1.041 train/location_acc=0.6875 train/token_losses=11.20 throughput/total_tokens=8,863,744 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:47:26.642 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=542/600, eta=24 minutes] train/CrossEntropyLoss=0.0258 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7667 train/patch_acc=0.7188 train/subpatch_loss=0.2075 train/subpatch_acc=0.8750 train/location_loss=0.7128 train/location_acc=0.8750 train/token_losses=9.814 throughput/total_tokens=8,880,128 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:47:51.194 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=543/600, eta=24 minutes] train/CrossEntropyLoss=0.0230 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.4858 train/patch_acc=0.8438 train/subpatch_loss=0.0441 train/subpatch_acc=1.0000 train/location_loss=1.049 train/location_acc=0.6875 train/token_losses=8.257 throughput/total_tokens=8,896,512 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:48:15.748 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=544/600, eta=23 minutes] train/CrossEntropyLoss=0.0715 train/ZLoss=0.0001 train/Accuracy=0.9787 train/patch_loss=0.9265 train/patch_acc=0.7500 train/subpatch_loss=0.1066 train/subpatch_acc=1.0000 train/location_loss=1.026 train/location_acc=0.6875 train/token_losses=11.94 throughput/total_tokens=8,912,896 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:48:40.307 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=545/600, eta=23 minutes] train/CrossEntropyLoss=0.0186 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8085 train/patch_acc=0.6875 train/subpatch_loss=0.0943 train/subpatch_acc=1.0000 train/location_loss=1.000 train/location_acc=0.5625 train/token_losses=10.85 throughput/total_tokens=8,929,280 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:49:04.846 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=546/600, eta=23 minutes] train/CrossEntropyLoss=0.0400 train/ZLoss=0.0003 train/Accuracy=0.9848 train/patch_loss=0.7937 train/patch_acc=0.7812 train/subpatch_loss=0.0446 train/subpatch_acc=1.0000 train/location_loss=1.002 train/location_acc=0.5000 train/token_losses=10.53 throughput/total_tokens=8,945,664 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:49:29.392 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=547/600, eta=22 minutes] train/CrossEntropyLoss=0.0389 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6598 train/patch_acc=0.7812 train/subpatch_loss=0.0309 train/subpatch_acc=1.0000 train/location_loss=0.7700 train/location_acc=0.6875 train/token_losses=8.482 throughput/total_tokens=8,962,048 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:49:53.948 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=548/600, eta=22 minutes] train/CrossEntropyLoss=0.0172 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=1.040 train/patch_acc=0.7188 train/subpatch_loss=0.0616 train/subpatch_acc=1.0000 train/location_loss=0.9460 train/location_acc=0.7500 train/token_losses=12.35 throughput/total_tokens=8,978,432 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:50:18.510 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=549/600, eta=21 minutes] train/CrossEntropyLoss=0.1276 train/ZLoss=0.0006 train/Accuracy=0.9695 train/patch_loss=0.8785 train/patch_acc=0.6875 train/subpatch_loss=0.1738 train/subpatch_acc=0.9375 train/location_loss=1.105 train/location_acc=0.6250 train/token_losses=12.14 throughput/total_tokens=8,994,816 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:50:43.066 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=550/600, eta=21 minutes] train/CrossEntropyLoss=0.0485 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.5384 train/patch_acc=0.8750 train/subpatch_loss=0.0819 train/subpatch_acc=1.0000 train/location_loss=0.5901 train/location_acc=0.7500 train/token_losses=6.996 throughput/total_tokens=9,011,200 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:51:07.602 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=551/600, eta=21 minutes] train/CrossEntropyLoss=0.0102 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.8669 train/patch_acc=0.6875 train/subpatch_loss=0.1996 train/subpatch_acc=0.9375 train/location_loss=0.6256 train/location_acc=0.9375 train/token_losses=10.24 throughput/total_tokens=9,027,584 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:51:32.152 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=552/600, eta=20 minutes] train/CrossEntropyLoss=0.0313 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.6038 train/patch_acc=0.8125 train/subpatch_loss=0.0851 train/subpatch_acc=1.0000 train/location_loss=0.9766 train/location_acc=0.8125 train/token_losses=9.077 throughput/total_tokens=9,043,968 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:51:56.704 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=553/600, eta=20 minutes] train/CrossEntropyLoss=0.0161 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.7408 train/patch_acc=0.8125 train/subpatch_loss=0.0368 train/subpatch_acc=1.0000 train/location_loss=0.9783 train/location_acc=0.6250 train/token_losses=9.986 throughput/total_tokens=9,060,352 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:52:21.251 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=554/600, eta=19 minutes] train/CrossEntropyLoss=0.0414 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.5605 train/patch_acc=0.8750 train/subpatch_loss=0.0869 train/subpatch_acc=1.0000 train/location_loss=0.5817 train/location_acc=0.8125 train/token_losses=7.158 throughput/total_tokens=9,076,736 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:52:45.817 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=555/600, eta=19 minutes] train/CrossEntropyLoss=0.0559 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.7241 train/patch_acc=0.7812 train/subpatch_loss=0.1820 train/subpatch_acc=0.9375 train/location_loss=0.8879 train/location_acc=0.7500 train/token_losses=10.07 throughput/total_tokens=9,093,120 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:53:10.380 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=556/600, eta=19 minutes] train/CrossEntropyLoss=0.0513 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.5740 train/patch_acc=0.8438 train/subpatch_loss=0.0392 train/subpatch_acc=1.0000 train/location_loss=0.9152 train/location_acc=0.5625 train/token_losses=8.410 throughput/total_tokens=9,109,504 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:53:34.913 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=557/600, eta=18 minutes] train/CrossEntropyLoss=0.0185 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.6256 train/patch_acc=0.8438 train/subpatch_loss=0.0332 train/subpatch_acc=1.0000 train/location_loss=0.5626 train/location_acc=0.7500 train/token_losses=7.388 throughput/total_tokens=9,125,888 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:53:59.473 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=558/600, eta=18 minutes] train/CrossEntropyLoss=0.2341 train/ZLoss=0.0013 train/Accuracy=0.9390 train/patch_loss=0.9278 train/patch_acc=0.7500 train/subpatch_loss=0.1044 train/subpatch_acc=1.0000 train/location_loss=1.050 train/location_acc=0.6875 train/token_losses=12.04 throughput/total_tokens=9,142,272 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:54:23.994 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=559/600, eta=17 minutes] train/CrossEntropyLoss=0.0201 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.9587 train/patch_acc=0.7188 train/subpatch_loss=0.0693 train/subpatch_acc=1.0000 train/location_loss=0.7721 train/location_acc=0.6875 train/token_losses=11.03 throughput/total_tokens=9,158,656 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:54:48.530 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=560/600, eta=17 minutes] train/CrossEntropyLoss=0.0111 train/ZLoss=0.0004 train/Accuracy=0.9970 train/patch_loss=0.6091 train/patch_acc=0.8125 train/subpatch_loss=0.0782 train/subpatch_acc=1.0000 train/location_loss=0.5067 train/location_acc=0.9375 train/token_losses=7.213 throughput/total_tokens=9,175,040 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:55:13.074 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=561/600, eta=16 minutes] train/CrossEntropyLoss=0.0431 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.5746 train/patch_acc=0.8125 train/subpatch_loss=0.1241 train/subpatch_acc=0.9375 train/location_loss=1.086 train/location_acc=0.6250 train/token_losses=9.438 throughput/total_tokens=9,191,424 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:55:37.608 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=562/600, eta=16 minutes] train/CrossEntropyLoss=0.0114 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6448 train/patch_acc=0.8125 train/subpatch_loss=0.0700 train/subpatch_acc=1.0000 train/location_loss=0.6725 train/location_acc=0.8750 train/token_losses=8.128 throughput/total_tokens=9,207,808 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:56:02.168 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=563/600, eta=16 minutes] train/CrossEntropyLoss=0.0124 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.7340 train/patch_acc=0.7188 train/subpatch_loss=0.0249 train/subpatch_acc=1.0000 train/location_loss=0.9777 train/location_acc=0.7500 train/token_losses=9.883 throughput/total_tokens=9,224,192 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:56:26.728 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=564/600, eta=15 minutes] train/CrossEntropyLoss=0.0139 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.7291 train/patch_acc=0.7500 train/subpatch_loss=0.1106 train/subpatch_acc=0.9375 train/location_loss=0.9497 train/location_acc=0.7500 train/token_losses=10.07 throughput/total_tokens=9,240,576 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:56:51.323 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=565/600, eta=15 minutes] train/CrossEntropyLoss=0.0303 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6144 train/patch_acc=0.8125 train/subpatch_loss=0.0730 train/subpatch_acc=1.0000 train/location_loss=0.7526 train/location_acc=0.8750 train/token_losses=8.218 throughput/total_tokens=9,256,960 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:57:15.875 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=566/600, eta=14 minutes] train/CrossEntropyLoss=0.0125 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.6551 train/patch_acc=0.8750 train/subpatch_loss=0.0719 train/subpatch_acc=1.0000 train/location_loss=1.097 train/location_acc=0.5625 train/token_losses=9.916 throughput/total_tokens=9,273,344 throughput/device/tokens_per_second=667.3 throughput/device/batches_per_second=0.0407 2026-07-27 15:57:40.413 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=567/600, eta=14 minutes] train/CrossEntropyLoss=0.0094 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6342 train/patch_acc=0.8125 train/subpatch_loss=0.0443 train/subpatch_acc=1.0000 train/location_loss=0.7294 train/location_acc=0.8125 train/token_losses=8.168 throughput/total_tokens=9,289,728 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:58:04.967 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=568/600, eta=14 minutes] train/CrossEntropyLoss=0.0437 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.5158 train/patch_acc=0.8438 train/subpatch_loss=0.1227 train/subpatch_acc=0.9375 train/location_loss=0.9515 train/location_acc=0.6250 train/token_losses=8.423 throughput/total_tokens=9,306,112 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:58:29.517 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=569/600, eta=13 minutes] train/CrossEntropyLoss=0.0131 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6919 train/patch_acc=0.8125 train/subpatch_loss=0.0859 train/subpatch_acc=1.0000 train/location_loss=0.6673 train/location_acc=0.6875 train/token_losses=8.549 throughput/total_tokens=9,322,496 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:58:54.034 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=570/600, eta=13 minutes] train/CrossEntropyLoss=0.0376 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.6314 train/patch_acc=0.7812 train/subpatch_loss=0.0833 train/subpatch_acc=1.0000 train/location_loss=1.076 train/location_acc=0.5625 train/token_losses=9.690 throughput/total_tokens=9,338,880 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 15:59:18.565 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=571/600, eta=12 minutes] train/CrossEntropyLoss=0.0107 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.7059 train/patch_acc=0.7188 train/subpatch_loss=0.0933 train/subpatch_acc=1.0000 train/location_loss=0.6769 train/location_acc=0.8750 train/token_losses=8.728 throughput/total_tokens=9,355,264 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 15:59:43.098 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=572/600, eta=12 minutes] train/CrossEntropyLoss=0.0263 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6615 train/patch_acc=0.7812 train/subpatch_loss=0.1552 train/subpatch_acc=0.9375 train/location_loss=0.7474 train/location_acc=0.8125 train/token_losses=8.903 throughput/total_tokens=9,371,648 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 16:00:07.649 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=573/600, eta=12 minutes] train/CrossEntropyLoss=0.0329 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8871 train/patch_acc=0.7500 train/subpatch_loss=0.0523 train/subpatch_acc=1.0000 train/location_loss=0.7794 train/location_acc=0.6875 train/token_losses=10.42 throughput/total_tokens=9,388,032 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 16:00:32.253 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=574/600, eta=11 minutes] train/CrossEntropyLoss=0.2209 train/ZLoss=0.0005 train/Accuracy=0.9451 train/patch_loss=1.032 train/patch_acc=0.6875 train/subpatch_loss=0.1516 train/subpatch_acc=0.9375 train/location_loss=1.317 train/location_acc=0.5000 train/token_losses=14.13 throughput/total_tokens=9,404,416 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:00:56.809 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=575/600, eta=11 minutes] train/CrossEntropyLoss=0.0305 train/ZLoss=0.0001 train/Accuracy=0.9878 train/patch_loss=0.5358 train/patch_acc=0.8125 train/subpatch_loss=0.1999 train/subpatch_acc=0.8750 train/location_loss=0.6366 train/location_acc=0.8125 train/token_losses=7.632 throughput/total_tokens=9,420,800 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:01:21.357 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=576/600, eta=10 minutes] train/CrossEntropyLoss=0.0699 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.5291 train/patch_acc=0.8438 train/subpatch_loss=0.0760 train/subpatch_acc=1.0000 train/location_loss=0.7021 train/location_acc=0.8125 train/token_losses=7.346 throughput/total_tokens=9,437,184 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:01:45.894 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=577/600, eta=10 minutes] train/CrossEntropyLoss=0.0245 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.6863 train/patch_acc=0.8438 train/subpatch_loss=0.0595 train/subpatch_acc=1.0000 train/location_loss=0.7199 train/location_acc=0.7500 train/token_losses=8.608 throughput/total_tokens=9,453,568 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:02:10.439 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=578/600, eta=10 minutes] train/CrossEntropyLoss=0.0511 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.7662 train/patch_acc=0.7500 train/subpatch_loss=0.1177 train/subpatch_acc=0.9375 train/location_loss=0.7206 train/location_acc=0.8125 train/token_losses=9.483 throughput/total_tokens=9,469,952 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:02:34.968 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=579/600, eta=9 minutes] train/CrossEntropyLoss=0.0278 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.7493 train/patch_acc=0.8125 train/subpatch_loss=0.0450 train/subpatch_acc=1.0000 train/location_loss=0.6891 train/location_acc=0.8750 train/token_losses=8.930 throughput/total_tokens=9,486,336 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:02:59.503 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=580/600, eta=9 minutes] train/CrossEntropyLoss=0.0535 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.8401 train/patch_acc=0.7500 train/subpatch_loss=0.0901 train/subpatch_acc=1.0000 train/location_loss=0.8209 train/location_acc=0.6250 train/token_losses=10.36 throughput/total_tokens=9,502,720 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 16:03:24.048 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=581/600, eta=8 minutes] train/CrossEntropyLoss=0.0136 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6458 train/patch_acc=0.7812 train/subpatch_loss=0.0680 train/subpatch_acc=0.9375 train/location_loss=0.6202 train/location_acc=0.8750 train/token_losses=7.920 throughput/total_tokens=9,519,104 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:03:48.589 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=582/600, eta=8 minutes] train/CrossEntropyLoss=0.0057 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.4736 train/patch_acc=0.8438 train/subpatch_loss=0.0811 train/subpatch_acc=1.0000 train/location_loss=1.103 train/location_acc=0.7500 train/token_losses=8.524 throughput/total_tokens=9,535,488 throughput/device/tokens_per_second=667.4 throughput/device/batches_per_second=0.0407 2026-07-27 16:04:13.113 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=583/600, eta=7 minutes] train/CrossEntropyLoss=0.0562 train/ZLoss=0.0002 train/Accuracy=0.9909 train/patch_loss=0.7869 train/patch_acc=0.7812 train/subpatch_loss=0.1270 train/subpatch_acc=1.0000 train/location_loss=0.8574 train/location_acc=0.8750 train/token_losses=10.23 throughput/total_tokens=9,551,872 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 16:04:37.646 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=584/600, eta=7 minutes] train/CrossEntropyLoss=0.0424 train/ZLoss=0.0001 train/Accuracy=0.9848 train/patch_loss=0.8771 train/patch_acc=0.8125 train/subpatch_loss=0.1662 train/subpatch_acc=0.9375 train/location_loss=0.8377 train/location_acc=0.8125 train/token_losses=11.03 throughput/total_tokens=9,568,256 throughput/device/tokens_per_second=667.5 throughput/device/batches_per_second=0.0407 2026-07-27 16:05:02.191 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=585/600, eta=7 minutes] train/CrossEntropyLoss=0.0029 train/ZLoss=0.0002 train/Accuracy=1.0000 train/patch_loss=0.9436 train/patch_acc=0.7188 train/subpatch_loss=0.0308 train/subpatch_acc=1.0000 train/location_loss=0.6829 train/location_acc=0.6875 train/token_losses=10.40 throughput/total_tokens=9,584,640 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 16:05:26.730 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=586/600, eta=6 minutes] train/CrossEntropyLoss=0.0900 train/ZLoss=0.0007 train/Accuracy=0.9787 train/patch_loss=1.047 train/patch_acc=0.6875 train/subpatch_loss=0.0768 train/subpatch_acc=1.0000 train/location_loss=0.8193 train/location_acc=0.6875 train/token_losses=11.96 throughput/total_tokens=9,601,024 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 16:05:51.237 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=587/600, eta=6 minutes] train/CrossEntropyLoss=0.0292 train/ZLoss=0.0002 train/Accuracy=0.9848 train/patch_loss=0.6220 train/patch_acc=0.7500 train/subpatch_loss=0.0600 train/subpatch_acc=1.0000 train/location_loss=0.5098 train/location_acc=0.8750 train/token_losses=7.255 throughput/total_tokens=9,617,408 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 16:06:15.769 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=588/600, eta=5 minutes] train/CrossEntropyLoss=0.0447 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.5397 train/patch_acc=0.7812 train/subpatch_loss=0.0422 train/subpatch_acc=1.0000 train/location_loss=0.6746 train/location_acc=0.8125 train/token_losses=7.185 throughput/total_tokens=9,633,792 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 16:06:40.303 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=589/600, eta=5 minutes] train/CrossEntropyLoss=0.0112 train/ZLoss=0.0002 train/Accuracy=0.9970 train/patch_loss=0.6568 train/patch_acc=0.7812 train/subpatch_loss=0.0598 train/subpatch_acc=1.0000 train/location_loss=0.9357 train/location_acc=0.7500 train/token_losses=9.236 throughput/total_tokens=9,650,176 throughput/device/tokens_per_second=667.7 throughput/device/batches_per_second=0.0408 2026-07-27 16:07:04.870 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=590/600, eta=5 minutes] train/CrossEntropyLoss=0.0628 train/ZLoss=0.0005 train/Accuracy=0.9787 train/patch_loss=0.8171 train/patch_acc=0.7188 train/subpatch_loss=0.0551 train/subpatch_acc=1.0000 train/location_loss=0.9442 train/location_acc=0.6250 train/token_losses=10.53 throughput/total_tokens=9,666,560 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 System/Peak GPU Memory (MB)=41,776 2026-07-27 16:07:29.408 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=591/600, eta=4 minutes] train/CrossEntropyLoss=0.0064 train/ZLoss=0.0001 train/Accuracy=1.0000 train/patch_loss=0.7360 train/patch_acc=0.7812 train/subpatch_loss=0.0522 train/subpatch_acc=1.0000 train/location_loss=0.8306 train/location_acc=0.6875 train/token_losses=9.419 throughput/total_tokens=9,682,944 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 16:07:53.936 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=592/600, eta=4 minutes] train/CrossEntropyLoss=0.0175 train/ZLoss=0.0001 train/Accuracy=0.9970 train/patch_loss=0.8587 train/patch_acc=0.7188 train/subpatch_loss=0.1135 train/subpatch_acc=1.0000 train/location_loss=0.9184 train/location_acc=0.7500 train/token_losses=11.00 throughput/total_tokens=9,699,328 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 16:08:18.455 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=593/600, eta=3 minutes] train/CrossEntropyLoss=0.0527 train/ZLoss=0.0001 train/Accuracy=0.9787 train/patch_loss=0.9136 train/patch_acc=0.7812 train/subpatch_loss=0.1482 train/subpatch_acc=0.9375 train/location_loss=0.7207 train/location_acc=0.7500 train/token_losses=10.78 throughput/total_tokens=9,715,712 throughput/device/tokens_per_second=667.6 throughput/device/batches_per_second=0.0407 2026-07-27 16:08:42.982 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=594/600, eta=3 minutes] train/CrossEntropyLoss=0.0495 train/ZLoss=0.0006 train/Accuracy=0.9848 train/patch_loss=0.5771 train/patch_acc=0.8750 train/subpatch_loss=0.0726 train/subpatch_acc=1.0000 train/location_loss=0.9124 train/location_acc=0.7500 train/token_losses=8.557 throughput/total_tokens=9,732,096 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 16:09:07.520 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=595/600, eta=3 minutes] train/CrossEntropyLoss=0.0206 train/ZLoss=0.0002 train/Accuracy=0.9939 train/patch_loss=0.5541 train/patch_acc=0.9062 train/subpatch_loss=0.0654 train/subpatch_acc=1.0000 train/location_loss=0.5786 train/location_acc=0.8125 train/token_losses=7.009 throughput/total_tokens=9,748,480 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 16:09:32.068 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=596/600, eta=2 minutes] train/CrossEntropyLoss=0.0385 train/ZLoss=0.0002 train/Accuracy=0.9878 train/patch_loss=0.7315 train/patch_acc=0.7812 train/subpatch_loss=0.0235 train/subpatch_acc=1.0000 train/location_loss=0.6193 train/location_acc=0.9375 train/token_losses=8.424 throughput/total_tokens=9,764,864 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 16:09:56.621 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=597/600, eta=2 minutes] train/CrossEntropyLoss=0.0530 train/ZLoss=0.0001 train/Accuracy=0.9909 train/patch_loss=0.6641 train/patch_acc=0.8438 train/subpatch_loss=0.0924 train/subpatch_acc=1.0000 train/location_loss=0.6889 train/location_acc=0.7500 train/token_losses=8.438 throughput/total_tokens=9,781,248 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 16:10:21.153 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=598/600, eta=1 minute] train/CrossEntropyLoss=0.2034 train/ZLoss=0.0007 train/Accuracy=0.9482 train/patch_loss=0.6522 train/patch_acc=0.7812 train/subpatch_loss=0.0989 train/subpatch_acc=1.0000 train/location_loss=0.9145 train/location_acc=0.8125 train/token_losses=9.271 throughput/total_tokens=9,797,632 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 16:10:45.690 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=599/600, eta=1 minute] train/CrossEntropyLoss=0.0442 train/ZLoss=0.0002 train/Accuracy=0.9817 train/patch_loss=0.5183 train/patch_acc=0.8438 train/subpatch_loss=0.0702 train/subpatch_acc=1.0000 train/location_loss=0.4058 train/location_acc=1.0000 train/token_losses=6.050 throughput/total_tokens=9,814,016 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 2026-07-27 16:11:10.206 2012bc4f83ae:0 olmo.train.trainer:1050 INFO [step=600/600, eta=1 minute] train/CrossEntropyLoss=0.0098 train/ZLoss=0.0001 train/Accuracy=0.9939 train/patch_loss=0.9509 train/patch_acc=0.7188 train/subpatch_loss=0.0828 train/subpatch_acc=1.0000 train/location_loss=0.8065 train/location_acc=0.7500 train/token_losses=11.16 throughput/total_tokens=9,830,400 throughput/device/tokens_per_second=667.8 throughput/device/batches_per_second=0.0408 System/Peak GPU Memory (MB)=41,776 2026-07-27 16:11:10.209 2012bc4f83ae:0 olmo.train.trainer:1415 INFO Saving checkpoint... 2026-07-27 16:11:11.445 2012bc4f83ae:0 olmo.train.trainer:1420 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step600 2026-07-27 16:11:13.064 2012bc4f83ae:0 olmo.train.trainer:1516 INFO Saving final checkpoint... 2026-07-27 16:11:13.796 2012bc4f83ae:0 olmo.train.trainer:1519 INFO Checkpoint saved to /runpod-volume/firewarning-pointing/checkpoints/molmopoint-strong-curated-ground-v1/step600 2026-07-27 16:11:13.797 2012bc4f83ae:0 train:412 INFO Training complete [rank0]:[W727 16:11:14.561122367 ProcessGroupNCCL.cpp:1538] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())