3d2smiles_finetune

This model is a fine-tuned version of weathon/3d2smiles_pretrain on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4533
• Accuracy: 0.3446

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0003
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1727	0.1887	10	1.8958	0.0225
0.6846	0.3774	20	1.0548	0.0674
0.7281	0.5660	30	0.8115	0.0899
0.4237	0.7547	40	0.6916	0.1199
0.382	0.9434	50	0.6418	0.1011
0.2166	1.1321	60	0.5568	0.1161
0.4252	1.3208	70	0.5942	0.1161
0.2349	1.5094	80	0.5469	0.1461
0.3193	1.6981	90	0.5241	0.1536
0.2882	1.8868	100	0.5122	0.1910
0.1594	2.0755	110	0.4991	0.2022
0.2488	2.2642	120	0.5354	0.2135
0.2995	2.4528	130	0.5288	0.1985
0.2753	2.6415	140	0.4949	0.2022
0.1867	2.8302	150	0.4521	0.2659
0.2008	3.0189	160	0.4660	0.2659
0.1039	3.2075	170	0.4611	0.2772
0.1859	3.3962	180	0.4913	0.2434
0.1246	3.5849	190	0.4688	0.2809
0.1892	3.7736	200	0.4414	0.2659
0.1974	3.9623	210	0.4506	0.2846
0.1104	4.1509	220	0.4447	0.2809
0.2212	4.3396	230	0.4230	0.2809
0.0833	4.5283	240	0.4459	0.2921
0.1657	4.7170	250	0.4493	0.3109
0.0871	4.9057	260	0.4442	0.3258
0.0645	5.0943	270	0.4282	0.3483
0.1196	5.2830	280	0.4619	0.3221
0.1213	5.4717	290	0.4466	0.3558
0.0972	5.6604	300	0.4357	0.3446
0.0697	5.8491	310	0.4464	0.3184
0.1113	6.0377	320	0.4428	0.3221
0.0646	6.2264	330	0.4488	0.3184
0.0747	6.4151	340	0.4388	0.3296
0.098	6.6038	350	0.4611	0.3109
0.0917	6.7925	360	0.4242	0.3483
0.0803	6.9811	370	0.4088	0.3521
0.0447	7.1698	380	0.4297	0.3371
0.1359	7.3585	390	0.4494	0.3371
0.0784	7.5472	400	0.4054	0.3633
0.076	7.7358	410	0.4434	0.3408
0.1167	7.9245	420	0.4104	0.3558
0.0633	8.1132	430	0.3991	0.3521
0.0881	8.3019	440	0.4055	0.3521
0.1245	8.4906	450	0.4198	0.3483
0.0657	8.6792	460	0.4668	0.2996
0.0989	8.8679	470	0.4557	0.3109
0.0658	9.0566	480	0.4495	0.3258
0.0504	9.2453	490	0.4407	0.3333
0.0931	9.4340	500	0.4467	0.3221
0.0628	9.6226	510	0.4268	0.3371
0.079	9.8113	520	0.4298	0.3184
0.0946	10.0	530	0.4588	0.3333
0.0851	10.1887	540	0.4317	0.3596
0.0583	10.3774	550	0.4501	0.3596
0.0513	10.5660	560	0.4348	0.3670
0.0367	10.7547	570	0.4124	0.3221
0.0733	10.9434	580	0.4516	0.3558
0.0782	11.1321	590	0.4370	0.3371
0.0258	11.3208	600	0.4887	0.3333
0.0958	11.5094	610	0.4156	0.3333
0.0874	11.6981	620	0.4492	0.3333
0.0496	11.8868	630	0.4383	0.3408
0.056	12.0755	640	0.4333	0.3296
0.0544	12.2642	650	0.4404	0.3408
0.0448	12.4528	660	0.4588	0.3333
0.0627	12.6415	670	0.4185	0.3333
0.0568	12.8302	680	0.4491	0.3558
0.0374	13.0189	690	0.4601	0.3446
0.0581	13.2075	700	0.4208	0.3558
0.0664	13.3962	710	0.4717	0.3446
0.0673	13.5849	720	0.4535	0.3296
0.0881	13.7736	730	0.4785	0.3184
0.0535	13.9623	740	0.4491	0.3446
0.1069	14.1509	750	0.4769	0.3408
0.0611	14.3396	760	0.4417	0.3109
0.0449	14.5283	770	0.4570	0.3296
0.0667	14.7170	780	0.4544	0.3446
0.1015	14.9057	790	0.4569	0.3408
0.0497	15.0943	800	0.4142	0.3558
0.0953	15.2830	810	0.4320	0.3596
0.0738	15.4717	820	0.4462	0.3558
0.0434	15.6604	830	0.4566	0.3708
0.0375	15.8491	840	0.4211	0.3745
0.043	16.0377	850	0.4177	0.3633
0.0799	16.2264	860	0.4191	0.3558
0.0474	16.4151	870	0.4398	0.3558
0.0786	16.6038	880	0.4505	0.3333
0.0426	16.7925	890	0.4500	0.3371
0.0526	16.9811	900	0.4420	0.3408
0.0206	17.1698	910	0.4380	0.3408
0.0345	17.3585	920	0.4423	0.3371
0.07	17.5472	930	0.4548	0.3446
0.0314	17.7358	940	0.4484	0.3408
0.0862	17.9245	950	0.4519	0.3371
0.0583	18.1132	960	0.4492	0.3333
0.082	18.3019	970	0.4653	0.3483
0.0285	18.4906	980	0.4626	0.3408
0.028	18.6792	990	0.4507	0.3371
0.03	18.8679	1000	0.4491	0.3333
0.0542	19.0566	1010	0.4508	0.3446
0.0253	19.2453	1020	0.4587	0.3483
0.039	19.4340	1030	0.4599	0.3521
0.0305	19.6226	1040	0.4548	0.3446
0.0495	19.8113	1050	0.4534	0.3446
0.0403	20.0	1060	0.4533	0.3446

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0