bengali_distilbert-base-uncased_qa_model_AGGRO

This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1233
Exact Match: 96.0
F1 Score: 96.5803

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: 1e-05
train_batch_size: 4
eval_batch_size: 4
seed: 3407
gradient_accumulation_steps: 16
total_train_batch_size: 64
optimizer: Use adamw_torch 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
training_steps: 150

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	F1 Score
6.2695	0.0053	1	6.2555	0.0	1.8428
6.2856	0.0107	2	6.2495	0.0	1.8426
6.2687	0.0160	3	6.2374	0.0	2.3078
6.2391	0.0214	4	6.2193	0.0	2.6010
6.2282	0.0267	5	6.1950	0.0	4.1183
6.2043	0.0321	6	6.1642	0.0	5.3317
6.1733	0.0374	7	6.1262	0.0	7.8170
6.1645	0.0428	8	6.0803	0.0	11.2526
6.1105	0.0481	9	6.0251	0.0	16.0264
6.0441	0.0535	10	5.9587	0.0752	21.1011
5.9912	0.0588	11	5.8792	0.1504	26.4689
5.9173	0.0641	12	5.7841	0.2256	31.8109
5.8213	0.0695	13	5.6711	0.5263	39.3749
5.7171	0.0748	14	5.5387	1.6541	45.5642
5.578	0.0802	15	5.3878	5.0376	52.4798
5.4483	0.0855	16	5.2241	13.7594	60.2760
5.2736	0.0909	17	5.0689	26.3910	67.7741
5.1415	0.0962	18	4.9279	37.7444	73.4564
5.0219	0.1016	19	4.7951	45.0376	76.4874
4.9269	0.1069	20	4.6587	51.2030	78.0813
4.8435	0.1123	21	4.5119	56.0150	79.9834
4.6803	0.1176	22	4.3534	60.8271	81.9112
4.5054	0.1230	23	4.1882	63.9098	83.4606
4.3615	0.1283	24	4.0244	66.2406	84.6553
4.2467	0.1336	25	3.8643	67.8947	85.3130
4.1394	0.1390	26	3.7097	69.0226	85.9888
3.8872	0.1443	27	3.5543	69.7744	86.0222
3.8603	0.1497	28	3.3962	70.4511	86.1104
3.613	0.1550	29	3.2344	70.7519	85.9820
3.436	0.1604	30	3.0715	71.2782	85.9536
3.3932	0.1657	31	2.9065	71.8045	85.6757
3.0935	0.1711	32	2.7407	72.4060	85.7304
2.8048	0.1764	33	2.5738	72.2556	85.2360
2.7518	0.1818	34	2.4096	71.9549	84.5948
2.5618	0.1871	35	2.2503	72.4060	84.3275
2.5722	0.1924	36	2.0966	72.5564	83.6134
2.1802	0.1978	37	1.9526	72.2556	82.5643
2.3555	0.2031	38	1.8185	71.8045	81.5743
2.1654	0.2085	39	1.6953	71.6541	80.8691
2.1076	0.2138	40	1.5824	71.8797	80.8166
1.9513	0.2192	41	1.4814	72.1053	80.8478
1.7153	0.2245	42	1.3863	72.2556	80.7091
1.6886	0.2299	43	1.2987	73.0075	81.3141
1.4405	0.2352	44	1.2151	73.6842	81.5204
1.4156	0.2406	45	1.1379	74.3609	81.5054
1.4225	0.2459	46	1.0671	75.0376	81.6117
1.1923	0.2513	47	1.0020	75.6391	82.0752
1.261	0.2566	48	0.9410	77.1429	83.4101
1.0952	0.2619	49	0.8834	79.0977	84.9435
1.0926	0.2673	50	0.8300	80.3759	85.9522
1.0268	0.2726	51	0.7823	81.6541	87.1255
1.0769	0.2780	52	0.7388	83.2331	88.6903
0.9573	0.2833	53	0.6993	84.2105	89.7419
0.7588	0.2887	54	0.6617	85.2632	90.1684
1.0753	0.2940	55	0.6278	86.1654	90.6748
0.7576	0.2994	56	0.5962	86.8421	91.1746
0.8011	0.3047	57	0.5677	87.6692	91.6342
0.8378	0.3101	58	0.5439	88.1955	91.9146
0.5775	0.3154	59	0.5218	88.4211	91.9815
0.6629	0.3207	60	0.4998	88.7218	92.2042
0.5714	0.3261	61	0.4760	89.0226	92.3364
0.6091	0.3314	62	0.4565	89.6241	92.6017
0.6007	0.3368	63	0.4408	90.0752	92.8030
0.327	0.3421	64	0.4272	90.6015	93.1921
0.5532	0.3475	65	0.4154	90.9774	93.3778
0.5815	0.3528	66	0.4029	91.1278	93.3848
0.5712	0.3582	67	0.3881	91.2030	93.4882
0.514	0.3635	68	0.3727	91.5789	93.5535
0.3137	0.3689	69	0.3578	91.8045	93.5384
0.9101	0.3742	70	0.3436	92.0301	93.6937
0.5761	0.3796	71	0.3305	92.2556	93.8291
0.3942	0.3849	72	0.3184	92.4060	93.8592
0.439	0.3902	73	0.3070	92.5564	93.8866
0.2963	0.3956	74	0.2963	92.7068	93.9768
0.3094	0.4009	75	0.2858	93.0827	94.2465
0.2631	0.4063	76	0.2771	93.3083	94.3374
0.4139	0.4116	77	0.2686	93.3083	94.3374
0.43	0.4170	78	0.2605	93.3835	94.3382
0.3083	0.4223	79	0.2530	93.5338	94.4773
0.4137	0.4277	80	0.2453	93.7594	94.6420
0.2547	0.4330	81	0.2385	93.8346	94.6921
0.5111	0.4384	82	0.2328	93.8346	94.6929
0.1655	0.4437	83	0.2296	93.9098	94.7673
0.1278	0.4490	84	0.2270	93.9098	94.8197
0.3699	0.4544	85	0.2245	94.0602	94.9686
0.3444	0.4597	86	0.2206	94.0602	94.9686
0.2372	0.4651	87	0.2170	94.0602	94.9686
0.253	0.4704	88	0.2123	94.1353	95.0438
0.289	0.4758	89	0.2080	94.0602	94.9784
0.282	0.4811	90	0.2036	94.1353	95.0531
0.2492	0.4865	91	0.2001	94.2105	95.0525
0.2682	0.4918	92	0.1969	94.2105	95.0525
0.2086	0.4972	93	0.1942	94.2857	95.1277
0.2035	0.5025	94	0.1920	94.3609	95.2021
0.2133	0.5079	95	0.1904	94.2857	95.1269
0.203	0.5132	96	0.1891	94.3609	95.2014
0.2208	0.5185	97	0.1879	94.3609	95.2014
0.2481	0.5239	98	0.1867	94.4361	95.2766
0.1908	0.5292	99	0.1848	94.3609	95.2021
0.2138	0.5346	100	0.1830	94.5865	95.4264
0.2623	0.5399	101	0.1807	94.5865	95.4264
0.2578	0.5453	102	0.1788	94.5865	95.4271
0.1946	0.5506	103	0.1774	94.5865	95.4286
0.1798	0.5560	104	0.1760	94.5865	95.4286
0.1335	0.5613	105	0.1746	94.5865	95.4286
0.2609	0.5667	106	0.1737	94.6617	95.5014
0.2711	0.5720	107	0.1730	94.7368	95.5766
0.1588	0.5773	108	0.1721	94.7368	95.5766
0.3639	0.5827	109	0.1714	94.8120	95.6518
0.2097	0.5880	110	0.1708	94.8120	95.6518
0.1608	0.5934	111	0.1702	94.8120	95.6518
0.33	0.5987	112	0.1699	94.8872	95.7270
0.1842	0.6041	113	0.1695	94.8872	95.7270
0.2727	0.6094	114	0.1688	94.8872	95.7270
0.1655	0.6148	115	0.1682	94.8872	95.7270
0.1447	0.6201	116	0.1674	94.8872	95.7270
0.2265	0.6255	117	0.1668	94.8872	95.7270
0.1062	0.6308	118	0.1661	94.8120	95.6518
0.2557	0.6362	119	0.1655	94.7368	95.5766
0.2841	0.6415	120	0.1649	94.7368	95.5766
0.1628	0.6468	121	0.1643	94.7368	95.5766
0.2288	0.6522	122	0.1637	94.8120	95.5954
0.1822	0.6575	123	0.1631	94.8120	95.5954
0.3044	0.6629	124	0.1625	94.8872	95.6706
0.2134	0.6682	125	0.1620	94.8120	95.5954
0.1703	0.6736	126	0.1615	94.8120	95.5954
0.1305	0.6789	127	0.1611	94.8120	95.5954
0.2363	0.6843	128	0.1608	94.8120	95.5954
0.2053	0.6896	129	0.1604	94.8120	95.5954
0.1662	0.6950	130	0.1600	94.8120	95.5954
0.1044	0.7003	131	0.1598	94.8120	95.5954
0.2024	0.7056	132	0.1596	94.8120	95.5954
0.2001	0.7110	133	0.1595	94.8120	95.5954
0.1649	0.7163	134	0.1593	94.8120	95.5954
0.2752	0.7217	135	0.1591	94.8120	95.5954
0.2345	0.7270	136	0.1590	94.8120	95.5954
0.2979	0.7324	137	0.1588	94.8120	95.5954
0.2875	0.7377	138	0.1587	94.8120	95.5954
0.2685	0.7431	139	0.1586	94.8120	95.5954
0.2041	0.7484	140	0.1585	94.8120	95.5954
0.1408	0.7538	141	0.1584	94.8120	95.5954
0.2367	0.7591	142	0.1583	94.8120	95.5954
0.1896	0.7645	143	0.1583	94.8120	95.5954
0.2586	0.7698	144	0.1582	94.8120	95.5954
0.2275	0.7751	145	0.1582	94.8120	95.5954
0.2487	0.7805	146	0.1582	94.8120	95.5954
0.2777	0.7858	147	0.1582	94.8120	95.5954
0.2179	0.7912	148	0.1582	94.8120	95.5954
0.1737	0.7965	149	0.1582	94.8120	95.5954
0.1885	0.8019	150	0.1582	94.8120	95.5954

Framework versions

Transformers 4.46.3
Pytorch 2.4.0
Datasets 3.1.0
Tokenizers 0.20.3

Mediocre-Judge/bengali_distilbert-base-uncased_qa_model_AGGRO

bengali_distilbert-base-uncased_qa_model_AGGRO

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Seed this model