learning_rate: 0.0001
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
28.8799	0.0820	5	39.5380	0.6111	0.0135	0.4296	0.1591	0.5269	0.1494	0.0526
8.1868	0.1639	10	34.1399	0.6094	0.0349	0.4296	0.1591	0.5269	0.1959	0.0654
21.8405	0.2459	15	19.6368	0.6202	0.5679	0.4366	0.1364	0.5376	0.4352	0.2222
10.9955	0.3279	20	12.4430	0.6336	0.9809	0.4437	0.1136	0.5484	0.6701	0.5668
15.0978	0.4098	25	11.4182	0.6445	0.9910	0.4648	0.0909	0.5699	0.7150	0.6401
6.6422	0.4918	30	13.3098	0.6514	0.8822	0.4718	0.0909	0.5753	0.5789	0.3841
19.4386	0.5738	35	15.0003	0.6522	0.6469	0.4930	0.1136	0.5860	0.5106	0.2622
17.6689	0.6557	40	13.9013	0.6540	0.6723	0.5141	0.1364	0.5968	0.5332	0.2836
21.243	0.7377	45	10.0193	0.6752	0.9315	0.5282	0.0909	0.6183	0.6684	0.4994
7.8117	0.8197	50	8.8910	0.6970	0.9089	0.5634	0.1364	0.6344	0.6905	0.4834
11.5014	0.9016	55	8.9064	0.7093	0.5176	0.6620	0.2727	0.6774	0.6400	0.3746
14.224	0.9836	60	9.3995	0.7031	0.8026	0.4296	0.0682	0.5484	0.5746	0.2947
0.2347	1.0656	65	6.6955	0.6989	0.9777	0.4014	0.0455	0.5323	0.7379	0.5410
13.506	1.1475	70	6.5436	0.7067	0.9790	0.4085	0.0455	0.5376	0.7365	0.5248
25.7709	1.2295	75	7.1836	0.7146	0.9105	0.5352	0.1591	0.6075	0.7123	0.4743
5.9195	1.3115	80	8.5983	0.7134	0.7387	0.5493	0.1818	0.6129	0.6288	0.3633
5.4344	1.3934	85	8.5560	0.7218	0.8522	0.4859	0.1136	0.5806	0.6389	0.3663
12.6959	1.4754	90	7.4516	0.7375	0.9124	0.5282	0.0909	0.6183	0.7145	0.4514
19.2404	1.5574	95	8.2965	0.7345	0.8213	0.5211	0.0909	0.6129	0.6525	0.3539
7.596	1.6393	100	10.2487	0.7310	0.5273	0.5493	0.1136	0.6290	0.5461	0.2208
4.8651	1.7213	105	11.6116	0.7326	0.3576	0.5493	0.1136	0.6290	0.4818	0.1625
5.0442	1.8033	110	9.0429	0.7333	0.6487	0.5563	0.1136	0.6344	0.5985	0.2790
7.5438	1.8852	115	7.8511	0.7363	0.8470	0.5423	0.0909	0.6290	0.6791	0.3813
1.0791	1.9672	120	7.9837	0.7367	0.7957	0.5493	0.0909	0.6344	0.6633	0.3560
5.2178	2.0492	125	9.3734	0.7362	0.5725	0.5634	0.0909	0.6452	0.5789	0.2535
1.3167	2.1311	130	9.4735	0.7422	0.5429	0.5915	0.0909	0.6667	0.5722	0.2318
8.1169	2.2131	135	9.3090	0.7420	0.5603	0.5915	0.1364	0.6559	0.5791	0.2387
13.6119	2.2951	140	8.7486	0.7481	0.6851	0.5634	0.0909	0.6452	0.6135	0.2628
2.2068	2.3770	145	8.0376	0.7407	0.7918	0.5634	0.0909	0.6452	0.6579	0.3291
5.528	2.4590	150	7.8566	0.7511	0.7241	0.5915	0.0909	0.6667	0.6677	0.3277
7.1032	2.5410	155	7.9564	0.7498	0.7670	0.5775	0.1136	0.6505	0.6614	0.3201
2.4178	2.6230	160	7.8716	0.7484	0.7114	0.6056	0.1364	0.6667	0.6616	0.3272
6.183	2.7049	165	7.8269	0.7458	0.7663	0.5775	0.0909	0.6559	0.6651	0.3330
1.5395	2.7869	170	7.8085	0.7481	0.7732	0.5915	0.0909	0.6667	0.6730	0.3444
4.6333	2.8689	175	7.6921	0.7540	0.7001	0.6127	0.1136	0.6774	0.6677	0.3320
4.145	2.9508	180	7.8318	0.7492	0.7291	0.5986	0.1136	0.6667	0.6659	0.3271

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_lr1e-4_r8_2602041310

bosch_RM_Qwen_S12345_LLM_false_STRUCT_true_epo3_lr1e-4_r8_2602041310

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_lr1e-4_r8_2602041310