Modelpublic

CureLink/curelink-biomed-nli-synth1-260426

sourceHugging Faceupdated 5mo agoView on Hugging Face

0likes23downloads

Model Card

curelink-biomed-nli-synth1-260426

This model is a fine-tuned version of CureLink/curelink-biomed-nli-v2 on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.6260
—Accuracy: 0.7029
—F1 Macro: 0.6932

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: 2e-06
—trainbatchsize: 1
—evalbatchsize: 1
—seed: 42
—gradientaccumulationsteps: 8
—totaltrainbatch_size: 8
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 50
—num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
12.9686	0.0344	200	0.9536	0.5219	0.3271
9.3602	0.0688	400	0.8072	0.56	0.3567
8.6758	0.1032	600	0.7477	0.5771	0.3577
8.4485	0.1376	800	0.7196	0.5886	0.3821
8.5295	0.1720	1000	0.7203	0.5810	0.3432
8.2626	0.2064	1200	0.7251	0.5848	0.3440
8.2063	0.2408	1400	0.6911	0.6095	0.3964
7.9263	0.2753	1600	0.7404	0.5752	0.3074
8.1049	0.3097	1800	0.6888	0.5810	0.3480
7.7745	0.3441	2000	0.6601	0.6286	0.4058
7.6881	0.3785	2200	0.6480	0.6210	0.5892
7.6799	0.4129	2400	0.6500	0.6019	0.5489
7.3053	0.4473	2600	0.6763	0.6019	0.5190
7.5197	0.4817	2800	0.7302	0.5829	0.4473
7.6030	0.5161	3000	0.6579	0.6019	0.5311
7.3845	0.5505	3200	0.6350	0.6267	0.6050
7.5394	0.5849	3400	0.6261	0.6495	0.6325
7.3222	0.6193	3600	0.6190	0.6667	0.6531
7.1954	0.6537	3800	0.6204	0.6667	0.6561
7.1886	0.6881	4000	0.6265	0.6648	0.6336
7.3738	0.7225	4200	0.6129	0.6705	0.6536
6.9979	0.7570	4400	0.6067	0.6724	0.6608
7.2239	0.7914	4600	0.6041	0.6667	0.6587
6.9000	0.8258	4800	0.6094	0.6686	0.6588
7.2602	0.8602	5000	0.6229	0.6629	0.6408
6.8126	0.8946	5200	0.6287	0.6648	0.6410
6.6980	0.9290	5400	0.6249	0.6629	0.6468
7.1687	0.9634	5600	0.6183	0.6686	0.6531
6.5912	0.9978	5800	0.6295	0.6686	0.6473
6.9909	1.0322	6000	0.6277	0.6743	0.6620
6.7530	1.0666	6200	0.6201	0.6762	0.6679
6.6378	1.1010	6400	0.6383	0.6762	0.6510
7.0990	1.1354	6600	0.6138	0.6743	0.6644
6.8045	1.1698	6800	0.6194	0.6686	0.6513
6.7787	1.2042	7000	0.6129	0.6819	0.6735
6.8148	1.2386	7200	0.6166	0.6838	0.6765
6.5366	1.2730	7400	0.6232	0.6857	0.6770
6.6520	1.3074	7600	0.6201	0.6838	0.6772
6.8052	1.3418	7800	0.6188	0.6857	0.6786
6.1560	1.3762	8000	0.6222	0.6876	0.6808
6.4957	1.4106	8200	0.6224	0.6838	0.6749
6.4563	1.4451	8400	0.6314	0.6838	0.6679
6.8127	1.4795	8600	0.6154	0.6914	0.6843
6.5587	1.5139	8800	0.6178	0.68	0.6692
6.3086	1.5483	9000	0.6207	0.6876	0.6724
6.3119	1.5827	9200	0.6158	0.6895	0.6758
6.4399	1.6171	9400	0.6144	0.6914	0.6814
6.3186	1.6515	9600	0.6270	0.6914	0.6770
6.2869	1.6859	9800	0.6181	0.6914	0.6814
6.32	1.7203	10000	0.6217	0.6933	0.6813
6.3284	1.7547	10200	0.6199	0.6914	0.6796
6.3140	1.7891	10400	0.6297	0.6990	0.6822
6.4686	1.8235	10600	0.6166	0.6914	0.6875
6.1366	1.8579	10800	0.6267	0.6933	0.6793
6.7102	1.8923	11000	0.6266	0.6895	0.6736
6.1974	1.9268	11200	0.6190	0.7048	0.6945
6.1358	1.9612	11400	0.6323	0.7029	0.6919
6.4884	1.9956	11600	0.6256	0.6857	0.6810
6.4385	2.0299	11800	0.6394	0.6952	0.6787
6.6774	2.0643	12000	0.6239	0.7067	0.6995
6.4918	2.0987	12200	0.6256	0.7124	0.7016
6.3618	2.1332	12400	0.6250	0.6876	0.6836
6.3480	2.1676	12600	0.6260	0.6895	0.6857
6.2335	2.2020	12800	0.6235	0.6914	0.6863
5.9912	2.2364	13000	0.6219	0.7048	0.6954
6.2472	2.2708	13200	0.6236	0.6990	0.6921
6.4709	2.3052	13400	0.6239	0.6971	0.6923
6.3383	2.3396	13600	0.6250	0.7048	0.6932
6.2136	2.3740	13800	0.6277	0.7067	0.6944
6.4244	2.4084	14000	0.6215	0.6952	0.6882
6.2413	2.4428	14200	0.6215	0.7010	0.6906
5.9511	2.4772	14400	0.6271	0.7010	0.6882
6.1934	2.5116	14600	0.6253	0.7029	0.6924
6.0209	2.5460	14800	0.6239	0.6971	0.6899
6.1706	2.5804	15000	0.6243	0.6914	0.685
6.2248	2.6149	15200	0.6245	0.6914	0.6857
6.1845	2.6493	15400	0.6242	0.6952	0.6885
6.2007	2.6837	15600	0.6237	0.6952	0.6882
6.2353	2.7181	15800	0.6236	0.6933	0.6871
6.3275	2.7525	16000	0.6232	0.6990	0.6917
6.0985	2.7869	16200	0.6253	0.7048	0.6950
5.9443	2.8213	16400	0.6246	0.7029	0.6937
5.9273	2.8557	16600	0.6257	0.7048	0.6950
6.0522	2.8901	16800	0.6263	0.7029	0.6928
6.5546	2.9245	17000	0.6256	0.7048	0.6950
6.1825	2.9589	17200	0.6260	0.7048	0.6950
5.8006	2.9933	17400	0.6260	0.7029	0.6932
5.8006	3.0	17439	0.6260	0.7029	0.6932

Framework versions

—Transformers 5.6.2
—Pytorch 2.11.0
—Datasets 4.8.4
—Tokenizers 0.22.2