COPAS-mms1ball-Nov28

This model is a fine-tuned version of facebook/mms-1b-all on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.7926
• Wer: 0.9939

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: 3
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
41.8398	0.2	100	37.3207	1.0
35.3547	0.4	200	27.8741	1.0
24.7565	0.6	300	19.3606	1.0
17.5639	0.8	400	12.4668	1.0
11.4454	1.0	500	8.1410	1.0
7.2036	1.2	600	6.0696	1.0
5.8017	1.4	700	5.3140	1.0
5.1698	1.6	800	5.0560	1.0
5.0254	1.8	900	4.9179	1.0
4.905	2.0	1000	4.8243	1.0
4.7647	2.2	1100	4.7402	1.0
4.736	2.4	1200	4.6689	1.0
4.6916	2.6	1300	4.6021	1.0
4.5545	2.8	1400	4.5514	1.0
4.5882	3.0	1500	4.5018	1.0
4.5573	3.2	1600	4.4525	1.0
4.4508	3.4	1700	4.4071	1.0
4.369	3.6	1800	4.3695	1.0
4.3941	3.8	1900	4.3237	1.0
4.3409	4.0	2000	4.2806	1.0
4.3205	4.2	2100	4.2383	1.0
4.2341	4.4	2200	4.1957	1.0
4.225	4.6	2300	4.1566	1.0
4.2211	4.8	2400	4.1210	1.0
4.1923	5.0	2500	4.0818	1.0
4.1785	5.2	2600	4.0491	1.0
4.1258	5.4	2700	4.0162	1.0
4.058	5.6	2800	3.9873	1.0
4.0715	5.8	2900	3.9629	1.0
4.0675	6.0	3000	3.9299	1.0
4.0338	6.2	3100	3.9041	1.0
3.9778	6.4	3200	3.8747	1.0
3.9507	6.6	3300	3.8510	0.9998
3.9829	6.8	3400	3.8248	0.9998
3.9373	7.0	3500	3.8036	0.9998
3.9433	7.2	3600	3.7812	0.9998
3.8741	7.4	3700	3.7618	0.9996
3.9197	7.6	3800	3.7420	0.9996
3.8497	7.8	3900	3.7223	0.9996
3.8934	8.0	4000	3.7031	0.9996
3.8427	8.2	4100	3.6863	0.9998
3.7761	8.4	4200	3.6698	0.9992
3.8544	8.6	4300	3.6574	0.9992
3.7864	8.8	4400	3.6425	0.9987
3.7913	9.0	4500	3.6284	0.9985
3.8103	9.2	4600	3.6162	0.9983
3.813	9.4	4700	3.6125	0.9985
3.7552	9.6	4800	3.5955	0.9983
3.744	9.8	4900	3.5832	0.9983
3.7656	10.0	5000	3.5749	0.9983
3.7119	10.2	5100	3.5686	0.9979
3.7246	10.4	5200	3.5613	0.9979
3.6999	10.6	5300	3.5483	0.9979
3.6942	10.8	5400	3.5395	0.9979
3.7076	11.0	5500	3.5338	0.9979
3.6577	11.2	5600	3.5219	0.9973
3.6771	11.4	5700	3.5147	0.9968
3.6948	11.6	5800	3.5055	0.9966
3.6699	11.8	5900	3.4969	0.9970
3.6306	12.0	6000	3.4868	0.9968
3.6393	12.2	6100	3.4800	0.9966
3.6745	12.4	6200	3.4712	0.9964
3.6641	12.6	6300	3.4659	0.9962
3.6167	12.8	6400	3.4595	0.9962
3.5831	13.0	6500	3.4537	0.9968
3.614	13.2	6600	3.4478	0.9966
3.6363	13.4	6700	3.4400	0.9966
3.6048	13.6	6800	3.4337	0.9966
3.5488	13.8	6900	3.4302	0.9966
3.5874	14.0	7000	3.4221	0.9964
3.5673	14.2	7100	3.4161	0.9962
3.5918	14.4	7200	3.4074	0.9964
3.6221	14.6	7300	3.4017	0.9964
3.516	14.8	7400	3.3931	0.9962
3.5529	15.0	7500	3.3872	0.9960
3.5173	15.2	7600	3.3806	0.9962
3.5608	15.4	7700	3.3721	0.9962
3.6101	15.6	7800	3.3639	0.9960
3.546	15.8	7900	3.3600	0.9958
3.481	16.0	8000	3.3549	0.9956
3.5324	16.2	8100	3.3471	0.9958
3.48	16.4	8200	3.3426	0.9962
3.5563	16.6	8300	3.3362	0.9956
3.5228	16.8	8400	3.3274	0.9949
3.4865	17.0	8500	3.3205	0.9954
3.5468	17.2	8600	3.3145	0.9962
3.4576	17.4	8700	3.3095	0.9962
3.4646	17.6	8800	3.3037	0.9964
3.4869	17.8	8900	3.2983	0.9966
3.4992	18.0	9000	3.2931	0.9962
3.4827	18.2	9100	3.2877	0.9960
3.4456	18.4	9200	3.2828	0.9956
3.5042	18.6	9300	3.2753	0.9960
3.4347	18.8	9400	3.2682	0.9968
3.4161	19.0	9500	3.2632	0.9962
3.4209	19.2	9600	3.2591	0.9958
3.4458	19.4	9700	3.2498	0.9962
3.4085	19.6	9800	3.2460	0.9956
3.4897	19.8	9900	3.2420	0.9958
3.4025	20.0	10000	3.2348	0.9960
3.4297	20.2	10100	3.2282	0.9962
3.4365	20.4	10200	3.2221	0.9968
3.4129	20.6	10300	3.2183	0.9962
3.4254	20.8	10400	3.2141	0.9960
3.3604	21.0	10500	3.2090	0.9958
3.3915	21.2	10600	3.2024	0.9956
3.4077	21.4	10700	3.1985	0.9958
3.3831	21.6	10800	3.1936	0.9962
3.414	21.8	10900	3.1885	0.9960
3.3778	22.0	11000	3.1834	0.9958
3.3987	22.2	11100	3.1798	0.9954
3.4096	22.4	11200	3.1766	0.9956
3.3784	22.6	11300	3.1724	0.9960
3.4194	22.8	11400	3.1680	0.9958
3.3011	23.0	11500	3.1658	0.9958
3.3206	23.2	11600	3.1631	0.9956
3.3476	23.4	11700	3.1577	0.9956
3.3604	23.6	11800	3.1540	0.9949
3.4032	23.8	11900	3.1475	0.9949
3.3523	24.0	12000	3.1421	0.9947
3.3223	24.2	12100	3.1386	0.9949
3.3869	24.4	12200	3.1342	0.9941
3.3354	24.6	12300	3.1295	0.9943
3.288	24.8	12400	3.1268	0.9941
3.3012	25.0	12500	3.1214	0.9939
3.3247	25.2	12600	3.1181	0.9939
3.3291	25.4	12700	3.1167	0.9935
3.3392	25.6	12800	3.1127	0.9939
3.281	25.8	12900	3.1089	0.9943
3.3083	26.0	13000	3.1030	0.9951
3.3973	26.2	13100	3.0982	0.9945
3.2582	26.4	13200	3.0948	0.9956
3.2509	26.6	13300	3.0916	0.9947
3.3027	26.8	13400	3.0875	0.9954
3.295	27.0	13500	3.0833	0.9937
3.2916	27.2	13600	3.0805	0.9943
3.2945	27.4	13700	3.0774	0.9937
3.2584	27.6	13800	3.0747	0.9939
3.3343	27.8	13900	3.0699	0.9945
3.24	28.0	14000	3.0661	0.9949
3.2768	28.2	14100	3.0614	0.9941
3.2713	28.4	14200	3.0587	0.9935
3.1811	28.6	14300	3.0544	0.9935
3.3279	28.8	14400	3.0506	0.9945
3.3166	29.0	14500	3.0470	0.9943
3.2904	29.2	14600	3.0454	0.9945
3.1675	29.4	14700	3.0395	0.9941
3.2665	29.6	14800	3.0368	0.9939
3.2087	29.8	14900	3.0320	0.9943
3.3436	30.0	15000	3.0290	0.9945
3.2558	30.2	15100	3.0267	0.9941
3.2631	30.4	15200	3.0222	0.9941
3.3143	30.6	15300	3.0184	0.9941
3.1722	30.8	15400	3.0135	0.9943
3.1736	31.0	15500	3.0101	0.9937
3.2694	31.2	15600	3.0052	0.9941
3.2143	31.4	15700	3.0015	0.9937
3.2431	31.6	15800	2.9993	0.9939
3.194	31.8	15900	2.9961	0.9937
3.1784	32.0	16000	2.9906	0.9937
3.239	32.2	16100	2.9866	0.9930
3.1766	32.4	16200	2.9837	0.9945
3.2049	32.6	16300	2.9788	0.9945
3.2638	32.8	16400	2.9769	0.9943
3.1008	33.0	16500	2.9749	0.9941
3.1918	33.2	16600	2.9728	0.9947
3.2645	33.4	16700	2.9702	0.9949
3.1329	33.6	16800	2.9615	0.9949
3.2031	33.8	16900	2.9575	0.9947
3.1297	34.0	17000	2.9542	0.9947
3.115	34.2	17100	2.9521	0.9947
3.1786	34.4	17200	2.9503	0.9947
3.1434	34.6	17300	2.9452	0.9949
3.2159	34.8	17400	2.9415	0.9943
3.1425	35.0	17500	2.9366	0.9943
3.1596	35.2	17600	2.9328	0.9943
3.1411	35.4	17700	2.9308	0.9935
3.2655	35.6	17800	2.9263	0.9941
3.1058	35.8	17900	2.9235	0.9928
3.1415	36.0	18000	2.9210	0.9930
3.1031	36.2	18100	2.9178	0.9935
3.1074	36.4	18200	2.9148	0.9939
3.0887	36.6	18300	2.9107	0.9937
3.2359	36.8	18400	2.9078	0.9932
3.137	37.0	18500	2.9060	0.9935
3.1064	37.2	18600	2.9044	0.9935
3.0584	37.4	18700	2.9010	0.9947
3.1004	37.6	18800	2.8977	0.9943
3.1034	37.8	18900	2.8948	0.9945
3.2163	38.0	19000	2.8906	0.9945
3.0611	38.2	19100	2.8864	0.9949
3.0713	38.4	19200	2.8852	0.9947
3.1233	38.6	19300	2.8816	0.9947
3.1374	38.8	19400	2.8776	0.9943
3.157	39.0	19500	2.8758	0.9937
3.1202	39.2	19600	2.8747	0.9935
3.0945	39.4	19700	2.8713	0.9941
3.0415	39.6	19800	2.8680	0.9947
3.0462	39.8	19900	2.8626	0.9941
3.1603	40.0	20000	2.8618	0.9943
3.0741	40.2	20100	2.8571	0.9945
3.0228	40.4	20200	2.8556	0.9949
3.1765	40.6	20300	2.8511	0.9943
3.027	40.8	20400	2.8478	0.9949
3.0472	41.0	20500	2.8451	0.9949
3.0993	41.2	20600	2.8446	0.9937
3.0562	41.4	20700	2.8414	0.9943
3.1409	41.6	20800	2.8383	0.9945
3.004	41.8	20900	2.8355	0.9943
3.0377	42.0	21000	2.8352	0.9945
3.1136	42.2	21100	2.8304	0.9949
3.0709	42.4	21200	2.8272	0.9947
3.0435	42.6	21300	2.8227	0.9947
3.0247	42.8	21400	2.8226	0.9943
3.0393	43.0	21500	2.8220	0.9949
3.037	43.2	21600	2.8182	0.9954
3.0403	43.4	21700	2.8149	0.9951
3.1406	43.6	21800	2.8141	0.9947
2.9519	43.8	21900	2.8129	0.9943
2.9742	44.0	22000	2.8075	0.9949
3.0384	44.2	22100	2.8039	0.9947
3.0387	44.4	22200	2.8021	0.9951
3.0851	44.6	22300	2.8025	0.9945
3.0079	44.8	22400	2.7977	0.9945
2.9731	45.0	22500	2.7951	0.9937
2.9938	45.2	22600	2.7972	0.9939
2.9564	45.4	22700	2.7926	0.9939

Framework versions

• Transformers 4.43.4
• Pytorch 2.4.1
• Datasets 3.0.0
• Tokenizers 0.19.1