all-MiniLM-L6-v2-twitter-sentiment

This model is a fine-tuned version of sentence-transformers/all-MiniLM-L6-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7018
• Accuracy: 0.6891
• F1: 0.6879
• Precision: 0.6940
• Recall: 0.6891

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-05
• train_batch_size: 16
• eval_batch_size: 64
• seed: 42
• 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: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.0307	0.0702	200	1.0141	0.435	0.2643	0.5984	0.435
0.8189	0.1403	400	0.7692	0.6685	0.6557	0.6946	0.6685
0.746	0.2105	600	0.7140	0.6845	0.6814	0.6812	0.6845
0.7113	0.2806	800	0.6845	0.701	0.7003	0.7108	0.701
0.6897	0.3508	1000	0.6715	0.703	0.7023	0.7058	0.703
0.6705	0.4209	1200	0.6571	0.7	0.7017	0.7046	0.7
0.6948	0.4911	1400	0.6541	0.7	0.7024	0.7090	0.7
0.6926	0.5612	1600	0.6452	0.7115	0.7121	0.7138	0.7115
0.6763	0.6314	1800	0.6394	0.719	0.7200	0.7216	0.719
0.6997	0.7015	2000	0.6698	0.6935	0.6858	0.7050	0.6935
0.6465	0.7717	2200	0.6327	0.716	0.7167	0.7181	0.716
0.6478	0.8418	2400	0.6339	0.7265	0.7277	0.7301	0.7265
0.6534	0.9120	2600	0.6447	0.7175	0.7191	0.7271	0.7175
0.6408	0.9821	2800	0.6184	0.725	0.7245	0.7241	0.725
0.5421	1.0523	3000	0.6390	0.7145	0.7166	0.7216	0.7145
0.5701	1.1224	3200	0.6339	0.725	0.7242	0.7242	0.725
0.5974	1.1926	3400	0.6308	0.725	0.7238	0.7237	0.725
0.5872	1.2627	3600	0.6252	0.726	0.7272	0.7299	0.726
0.5847	1.3329	3800	0.6265	0.7315	0.7329	0.7359	0.7315
0.5906	1.4030	4000	0.6354	0.7205	0.7198	0.7244	0.7205
0.5541	1.4732	4200	0.6206	0.7305	0.7309	0.7314	0.7305
0.5922	1.5433	4400	0.6214	0.7305	0.7314	0.7334	0.7305
0.5816	1.6135	4600	0.6329	0.7275	0.7290	0.7358	0.7275
0.5312	1.6836	4800	0.6438	0.719	0.7170	0.7223	0.719
0.5477	1.7538	5000	0.6297	0.727	0.7269	0.7292	0.727
0.5488	1.8239	5200	0.6314	0.7295	0.7290	0.7292	0.7295
0.5541	1.8941	5400	0.6186	0.734	0.7340	0.7342	0.734
0.587	1.9642	5600	0.6103	0.734	0.7338	0.7340	0.734
0.5011	2.0344	5800	0.6428	0.729	0.7303	0.7362	0.729
0.5241	2.1045	6000	0.6288	0.7315	0.7322	0.7331	0.7315
0.5462	2.1747	6200	0.6369	0.728	0.7251	0.7263	0.728
0.4792	2.2448	6400	0.6365	0.729	0.7306	0.7339	0.729
0.5222	2.3150	6600	0.6320	0.73	0.7295	0.7292	0.73
0.4967	2.3851	6800	0.6326	0.729	0.7288	0.7296	0.729
0.5261	2.4553	7000	0.6316	0.726	0.7250	0.7246	0.726
0.519	2.5254	7200	0.6329	0.7285	0.7293	0.7307	0.7285
0.5168	2.5956	7400	0.6279	0.7315	0.7324	0.7343	0.7315
0.4944	2.6657	7600	0.6308	0.7325	0.7322	0.7321	0.7325
0.5048	2.7359	7800	0.6424	0.73	0.7306	0.7322	0.73
0.5201	2.8060	8000	0.6342	0.734	0.7343	0.7349	0.734
0.5391	2.8762	8200	0.6314	0.7305	0.7307	0.7311	0.7305
0.5361	2.9463	8400	0.6353	0.733	0.7335	0.7342	0.733

Framework versions

• Transformers 4.55.4
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4