learning_rate: 5e-05
train_batch_size: 1
eval_batch_size: 1
seed: 12345
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 8
total_eval_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Roc Auc	Best Threshold	Tpr At Best Threshold	Fpr At Best Threshold	Accuracy At Best Threshold	Avg Score True Positives	Avg Score True Negatives
No log	0	0	40.8364	0.6060	0.0100	0.4225	0.1364	0.5269	0.1403	0.0493
29.7371	0.0820	5	40.0075	0.6064	0.0059	0.4718	0.2045	0.5484	0.1425	0.0525
10.7812	0.1639	10	38.0285	0.6069	0.0167	0.4225	0.1364	0.5269	0.1625	0.0589
23.8144	0.2459	15	31.8214	0.6079	0.0430	0.4507	0.1818	0.5376	0.2301	0.0764
23.0412	0.3279	20	21.9043	0.6160	0.2644	0.4859	0.2045	0.5591	0.3888	0.1813
9.0435	0.4098	25	14.9583	0.6292	0.8880	0.4437	0.1364	0.5430	0.5743	0.4038
6.6739	0.4918	30	14.1920	0.6310	0.8973	0.4930	0.1591	0.5753	0.5885	0.4332
19.2456	0.5738	35	13.1857	0.6332	0.9252	0.4859	0.1591	0.5699	0.6149	0.4725
21.8909	0.6557	40	12.5382	0.6383	0.9526	0.4789	0.1136	0.5753	0.6317	0.5060
25.8242	0.7377	45	12.2496	0.6469	0.9381	0.4930	0.1136	0.5860	0.6328	0.4863
10.5411	0.8197	50	13.2440	0.6446	0.8529	0.4789	0.0909	0.5806	0.5731	0.3908
14.1703	0.9016	55	12.6475	0.6446	0.8836	0.4577	0.0909	0.5645	0.5706	0.3693
21.1759	0.9836	60	11.8701	0.6548	0.8728	0.4366	0.0909	0.5484	0.5727	0.3563
1.8075	1.0656	65	9.8060	0.6600	0.9641	0.3592	0.0455	0.5	0.6302	0.4392
19.251	1.1475	70	8.4607	0.6663	0.9615	0.4437	0.1364	0.5430	0.6952	0.5426
29.1023	1.2295	75	7.7036	0.6728	0.9859	0.3803	0.0682	0.5108	0.7256	0.5618
7.1992	1.3115	80	7.9317	0.6763	0.9700	0.3873	0.0682	0.5161	0.6891	0.4948
9.6169	1.3934	85	8.2385	0.6797	0.9325	0.4437	0.1136	0.5484	0.6490	0.4402
18.6351	1.4754	90	8.9583	0.6828	0.7126	0.5282	0.2045	0.5914	0.6042	0.3711
20.2742	1.5574	95	9.9653	0.6960	0.5215	0.5704	0.2273	0.6183	0.5538	0.2883
11.1821	1.6393	100	10.0245	0.6892	0.6002	0.5352	0.1818	0.6022	0.5534	0.2992
5.5346	1.7213	105	9.9380	0.6975	0.3849	0.6408	0.2727	0.6613	0.5694	0.3001
5.0771	1.8033	110	9.3142	0.7010	0.7013	0.5423	0.1818	0.6075	0.6015	0.3389
11.795	1.8852	115	9.3091	0.7047	0.6838	0.5634	0.1818	0.6237	0.6006	0.3244
4.2519	1.9672	120	10.0429	0.7001	0.5472	0.5775	0.2045	0.6290	0.5735	0.2945
7.8352	2.0492	125	10.6101	0.7025	0.5583	0.5352	0.1591	0.6075	0.5427	0.2592
2.0506	2.1311	130	10.7215	0.7052	0.5351	0.5493	0.1364	0.6237	0.5411	0.2463
6.9675	2.2131	135	10.6602	0.7080	0.6132	0.5423	0.1136	0.6237	0.5520	0.2479
20.5126	2.2951	140	10.1269	0.7188	0.6567	0.5563	0.1136	0.6344	0.5752	0.2590
4.883	2.3770	145	9.9011	0.7121	0.5098	0.5986	0.1591	0.6559	0.5858	0.2777
9.3277	2.4590	150	9.8027	0.7158	0.5837	0.5704	0.1364	0.6398	0.5922	0.2803
11.6811	2.5410	155	10.0075	0.7104	0.6202	0.5775	0.1136	0.6505	0.5812	0.2766
3.6915	2.6230	160	9.6717	0.7159	0.5641	0.5845	0.1364	0.6505	0.5902	0.2850
13.0371	2.7049	165	9.6799	0.7175	0.6243	0.5704	0.1364	0.6398	0.5965	0.2917
1.3146	2.7869	170	9.6060	0.7152	0.6160	0.5634	0.1364	0.6344	0.5959	0.2953
8.3405	2.8689	175	9.8191	0.7086	0.6540	0.5634	0.1364	0.6344	0.5888	0.2988
4.0103	2.9508	180	9.7316	0.7155	0.6030	0.5775	0.1364	0.6452	0.5932	0.2883

Framework versions

PEFT 0.18.1
Transformers 4.57.6
Pytorch 2.9.1+cu128
Datasets 4.5.0
Tokenizers 0.22.2

leobianco/bosch_RM_Qwen_S12345_LLM_false_STRUCT_true_epo3_lr5e-5_r8_2602040939

bosch_RM_Qwen_S12345_LLM_false_STRUCT_true_epo3_lr5e-5_r8_2602040939

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

leobianco/
bosch_RM_Qwen_S12345_LLM_false_STRUCT_true_epo3_lr5e-5_r8_2602040939