fineweb-edu-scorer-mdeberta-binary

This model is a fine-tuned version of microsoft/mdeberta-v3-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0644
• Precision: 0.8738
• Recall: 0.5228
• F1 Macro: 0.5213
• Accuracy: 0.9132

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: 5e-05
• train_batch_size: 128
• eval_batch_size: 256
• seed: 0
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1 Macro	Accuracy
No log	0	0	0.1176	0.4549	0.5	0.4764	0.9098
0.0765	0.3042	1000	0.0758	0.4549	0.5	0.4764	0.9098
0.0715	0.6085	2000	0.0723	0.4549	0.5	0.4764	0.9098
0.0703	0.9127	3000	0.0718	0.6216	0.5003	0.4771	0.9098
0.0696	1.2169	4000	0.0714	0.5799	0.5001	0.4766	0.9098
0.0703	1.5211	5000	0.0693	0.6633	0.5005	0.4776	0.9098
0.07	1.8254	6000	0.0688	0.7407	0.5004	0.4773	0.9099
0.0686	2.1296	7000	0.0694	0.8123	0.5023	0.4812	0.9101
0.0703	2.4338	8000	0.0733	0.8295	0.5172	0.5109	0.9120
0.0687	2.7381	9000	0.0681	0.8479	0.5013	0.4790	0.9100
0.0671	3.0423	10000	0.0678	0.9021	0.5049	0.4864	0.9106
0.0684	3.3465	11000	0.0674	0.8906	0.5055	0.4876	0.9107
0.069	3.6507	12000	0.0676	0.8979	0.5081	0.4928	0.9111
0.0681	3.9550	13000	0.0679	0.8975	0.5027	0.4819	0.9103
0.0713	4.2592	14000	0.0673	0.8769	0.5112	0.4990	0.9115
0.0695	4.5634	15000	0.0670	0.9100	0.5059	0.4883	0.9108
0.0684	4.8677	16000	0.0672	0.9100	0.5059	0.4883	0.9108
0.0657	5.1719	17000	0.0668	0.8835	0.5131	0.5027	0.9118
0.0672	5.4761	18000	0.0664	0.9108	0.5095	0.4956	0.9114
0.0681	5.7803	19000	0.0683	0.9026	0.5040	0.4846	0.9105
0.066	6.0846	20000	0.0667	0.8911	0.5164	0.5090	0.9124
0.0669	6.3888	21000	0.0664	0.9133	0.5076	0.4919	0.9111
0.0678	6.6930	22000	0.0662	0.8915	0.5150	0.5063	0.9122
0.0679	6.9973	23000	0.0664	0.8685	0.5201	0.5162	0.9128
0.0659	7.3015	24000	0.0658	0.8981	0.5125	0.5016	0.9118
0.067	7.6057	25000	0.0658	0.9018	0.5116	0.4997	0.9117
0.0654	7.9099	26000	0.0658	0.9058	0.5106	0.4977	0.9116
0.0651	8.2142	27000	0.0659	0.8758	0.5176	0.5114	0.9125
0.067	8.5184	28000	0.0664	0.9107	0.5083	0.4933	0.9112
0.0651	8.8226	29000	0.0661	0.8791	0.5216	0.5189	0.9131
0.0675	9.1269	30000	0.0654	0.8900	0.5153	0.5070	0.9122
0.0683	9.4311	31000	0.0665	0.9097	0.5093	0.4951	0.9114
0.0651	9.7353	32000	0.0658	0.9053	0.5104	0.4975	0.9115
0.0706	10.0395	33000	0.0655	0.9104	0.5117	0.5000	0.9118
0.0661	10.3438	34000	0.0654	0.8783	0.5219	0.5196	0.9131
0.0655	10.6480	35000	0.0667	0.9045	0.5093	0.4951	0.9113
0.0657	10.9522	36000	0.0655	0.9105	0.5129	0.5023	0.9120
0.0667	11.2565	37000	0.0651	0.9033	0.5149	0.5061	0.9122
0.0658	11.5607	38000	0.0651	0.8713	0.5225	0.5208	0.9131
0.0641	11.8649	39000	0.0670	0.9050	0.5083	0.4933	0.9112
0.0671	12.1692	40000	0.0650	0.8917	0.5158	0.5079	0.9123
0.0645	12.4734	41000	0.0650	0.8749	0.5226	0.5208	0.9132
0.0671	12.7776	42000	0.0655	0.9095	0.5137	0.5039	0.9121
0.0667	13.0818	43000	0.0660	0.8478	0.5371	0.5472	0.9148
0.0647	13.3861	44000	0.0657	0.8644	0.5314	0.5371	0.9143
0.0634	13.6903	45000	0.0651	0.9086	0.5156	0.5075	0.9124
0.0676	13.9945	46000	0.0648	0.8749	0.5226	0.5208	0.9132
0.0671	14.2988	47000	0.0652	0.8641	0.5293	0.5333	0.9140
0.0656	14.6030	48000	0.0647	0.8753	0.5204	0.5167	0.9129
0.0652	14.9072	49000	0.0646	0.8740	0.5200	0.5160	0.9128
0.0646	15.2114	50000	0.0646	0.8732	0.5198	0.5156	0.9128
0.0632	15.5157	51000	0.0646	0.8763	0.5225	0.5206	0.9132
0.0626	15.8199	52000	0.0650	0.9055	0.5156	0.5075	0.9124
0.0643	16.1241	53000	0.0650	0.9072	0.5162	0.5086	0.9125
0.0651	16.4284	54000	0.0646	0.8730	0.5248	0.5249	0.9135
0.0669	16.7326	55000	0.0645	0.8749	0.5220	0.5198	0.9131
0.0679	17.0368	56000	0.0647	0.8759	0.5276	0.5301	0.9139
0.0665	17.3410	57000	0.0663	0.9113	0.5120	0.5005	0.9118
0.066	17.6453	58000	0.0645	0.8785	0.5202	0.5163	0.9129
0.0652	17.9495	59000	0.0645	0.8750	0.5261	0.5273	0.9137
0.0659	18.2537	60000	0.0647	0.8839	0.5185	0.5132	0.9127
0.0641	18.5580	61000	0.0645	0.8729	0.5208	0.5176	0.9129
0.0635	18.8622	62000	0.0645	0.8729	0.5208	0.5176	0.9129
0.0623	19.1664	63000	0.0645	0.8731	0.5226	0.5208	0.9132
0.0633	19.4706	64000	0.0644	0.8756	0.5234	0.5224	0.9133
0.0678	19.7749	65000	0.0644	0.8738	0.5228	0.5213	0.9132

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.2
• Tokenizers 0.21.1