ssc-lke-model

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.3630
• Cer: 0.4180
• Wer: 0.9952

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 100
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
4.9592	0.1618	100	3.0080	0.9936	0.9998
3.0455	0.3236	200	2.9443	0.9936	0.9998
2.9891	0.4854	300	2.9568	0.9936	0.9998
2.9251	0.6472	400	3.2078	0.9936	0.9998
2.8269	0.8091	500	2.8238	0.9873	1.0
2.6627	0.9709	600	2.5854	0.9705	1.0
2.552	1.1327	700	2.3775	0.9037	1.0
2.3909	1.2945	800	2.1959	0.7265	0.9991
2.247	1.4563	900	2.0131	0.6744	0.9992
2.1447	1.6181	1000	1.9431	0.6586	0.9991
2.0731	1.7799	1100	1.9004	0.5292	1.0149
2.0267	1.9417	1200	1.8489	0.5873	0.9983
1.9727	2.1036	1300	1.7685	0.5495	1.0009
1.9237	2.2654	1400	1.7706	0.5568	0.9969
1.884	2.4272	1500	1.7281	0.5260	0.9959
1.8408	2.5890	1600	1.6832	0.4931	0.9952
1.8211	2.7508	1700	1.6495	0.4929	0.9956
1.7979	2.9126	1800	1.7423	0.4922	0.9960
1.7821	3.0744	1900	1.6515	0.5092	0.9968
1.7059	3.2362	2000	1.6636	0.4875	1.0029
1.6935	3.3981	2100	1.6712	0.5069	0.9946
1.7209	3.5599	2200	1.6008	0.4946	0.9971
1.7215	3.7217	2300	1.7042	0.4724	0.9879
1.6823	3.8835	2400	1.6018	0.4617	0.9873
1.6631	4.0453	2500	1.6561	0.4745	0.9862
1.5967	4.2071	2600	1.7273	0.4651	0.9900
1.5947	4.3689	2700	1.7163	0.4651	0.9802
1.6023	4.5307	2800	1.5813	0.4501	1.0058
1.624	4.6926	2900	1.5864	0.4612	0.9861
1.5702	4.8544	3000	1.6096	0.4481	0.9922
1.569	5.0162	3100	1.5260	0.4337	0.9839
1.4795	5.1780	3200	1.5816	0.4335	0.9908
1.5097	5.3398	3300	1.5302	0.4342	0.9793
1.5537	5.5016	3400	1.5777	0.4289	1.0049
1.5089	5.6634	3500	1.5101	0.4291	0.9982
1.5073	5.8252	3600	1.4878	0.4603	0.9882
1.5034	5.9871	3700	1.5517	0.4322	1.0283
1.4335	6.1489	3800	1.6155	0.4351	1.1114
1.4086	6.3107	3900	1.7017	0.4517	0.9995
1.4468	6.4725	4000	1.5499	0.4429	0.9808
1.488	6.6343	4100	1.5266	0.4556	0.9844
1.4398	6.7961	4200	1.5643	0.4307	0.9841
1.4592	6.9579	4300	1.5186	0.4439	1.0002
1.3771	7.1197	4400	1.5960	0.4348	1.0572
1.3793	7.2816	4500	1.5730	0.4295	1.0026
1.4057	7.4434	4600	1.5218	0.4423	0.9784
1.3717	7.6052	4700	1.5154	0.4257	0.9692
1.3178	7.7670	4800	1.4786	0.4232	0.9674
1.4251	7.9288	4900	1.5020	0.4438	0.9776
1.3453	8.0906	5000	1.4952	0.4141	0.9720
1.2894	8.2524	5100	1.4786	0.4382	0.9741
1.3209	8.4142	5200	1.5410	0.4202	1.0041
1.3171	8.5761	5300	1.4860	0.4326	0.9743
1.3072	8.7379	5400	1.4815	0.4295	0.9747
1.3176	8.8997	5500	1.4677	0.4215	0.9666
1.3303	9.0615	5600	1.5632	0.4338	1.0389
1.272	9.2233	5700	1.5588	0.4469	0.9706
1.2647	9.3851	5800	1.5846	0.4203	1.0481
1.2736	9.5469	5900	1.5289	0.4203	0.9812
1.2404	9.7087	6000	1.5516	0.4221	0.9761
1.2323	9.8706	6100	1.5851	0.4385	0.9885
1.2793	10.0324	6200	1.5811	0.4264	0.9868
1.1518	10.1942	6300	1.5056	0.4134	0.9651
1.2106	10.3560	6400	1.5828	0.4237	1.0458
1.1944	10.5178	6500	1.5465	0.4357	0.9713
1.2183	10.6796	6600	1.6196	0.4275	1.0924
1.2096	10.8414	6700	1.5578	0.4177	0.9893
1.2175	11.0032	6800	1.5542	0.4349	0.9810
1.1563	11.1650	6900	1.5996	0.4296	0.9877
1.1515	11.3269	7000	1.6314	0.4289	0.9880
1.1464	11.4887	7100	1.5662	0.4236	0.9861
1.1394	11.6505	7200	1.5364	0.4182	0.9975
1.1288	11.8123	7300	1.5461	0.4150	0.9919
1.1558	11.9741	7400	1.4826	0.4110	0.9660
1.1019	12.1359	7500	1.5732	0.4265	0.9798
1.0752	12.2977	7600	1.5158	0.4137	0.9962
1.0767	12.4595	7700	1.5236	0.4098	0.9773
1.0777	12.6214	7800	1.5703	0.4047	1.0107
1.0735	12.7832	7900	1.6005	0.4129	0.9664
1.1235	12.9450	8000	1.5316	0.4117	0.9707
1.0589	13.1068	8100	1.5899	0.4126	0.9897
1.0143	13.2686	8200	1.5983	0.4082	0.9962
1.0443	13.4304	8300	1.5717	0.4148	1.0041
1.0472	13.5922	8400	1.5485	0.4036	0.9910
1.0385	13.7540	8500	1.5523	0.4070	0.9743
1.0417	13.9159	8600	1.4707	0.3993	0.9667
1.0014	14.0777	8700	1.5785	0.4167	0.9643
0.9603	14.2395	8800	1.6276	0.4256	0.9922
0.953	14.4013	8900	1.5862	0.4238	0.9687
0.9985	14.5631	9000	1.6115	0.4162	0.9664
0.9912	14.7249	9100	1.6869	0.4114	1.0162
1.0031	14.8867	9200	1.5644	0.4260	0.9899
0.9788	15.0485	9300	1.5787	0.4065	0.9795
0.9158	15.2104	9400	1.6169	0.4081	0.9677
0.9261	15.3722	9500	1.6508	0.4088	1.0009
0.9498	15.5340	9600	1.6283	0.4073	1.0046
0.9695	15.6958	9700	1.5756	0.4008	0.9891
0.9147	15.8576	9800	1.5783	0.4028	0.9701
0.9437	16.0194	9900	1.5883	0.4109	0.9672
0.8972	16.1812	10000	1.5767	0.4116	0.9723
0.9125	16.3430	10100	1.6018	0.4159	0.9690
0.884	16.5049	10200	1.5448	0.4179	0.9749
0.8832	16.6667	10300	1.5594	0.4041	0.9577
0.8791	16.8285	10400	1.5256	0.4151	0.9664
0.9031	16.9903	10500	1.6793	0.4100	0.9606
0.8321	17.1521	10600	1.6991	0.4157	0.9801
0.8423	17.3139	10700	1.6772	0.4155	0.9877
0.8189	17.4757	10800	1.6960	0.4113	0.9913
0.8337	17.6375	10900	1.7397	0.4074	0.9585
0.8267	17.7994	11000	1.6701	0.4116	0.9720
0.8573	17.9612	11100	1.6441	0.3954	0.9655
0.7768	18.1230	11200	1.7098	0.4040	0.9692
0.7796	18.2848	11300	1.7118	0.4136	0.9724
0.7918	18.4466	11400	1.6257	0.4082	0.9899
0.7744	18.6084	11500	1.7557	0.4156	0.9977
0.7872	18.7702	11600	1.7399	0.4144	1.0127
0.8028	18.9320	11700	1.8050	0.4079	0.9893
0.8	19.0939	11800	1.6490	0.4134	0.9859
0.7615	19.2557	11900	1.6924	0.4112	0.9717
0.7623	19.4175	12000	1.6945	0.4058	0.9603
0.7567	19.5793	12100	1.6489	0.4096	0.9595
0.737	19.7411	12200	1.6861	0.4091	0.9539
0.7412	19.9029	12300	1.6375	0.4114	0.9588
0.6968	20.0647	12400	1.9391	0.4056	0.9733
0.6733	20.2265	12500	1.8758	0.4111	0.9583
0.7108	20.3883	12600	1.7187	0.4257	0.9700
0.7107	20.5502	12700	1.8774	0.4213	0.9746
0.6913	20.7120	12800	1.8077	0.4134	0.9713
0.7081	20.8738	12900	1.7353	0.4075	0.9828
0.722	21.0356	13000	1.8029	0.4293	0.9740
0.6476	21.1974	13100	1.7747	0.4181	0.9767
0.6598	21.3592	13200	1.7614	0.4213	0.9982
0.6468	21.5210	13300	1.7454	0.4238	0.9807
0.6909	21.6828	13400	1.7486	0.4198	0.9922
0.6555	21.8447	13500	1.7060	0.4218	0.9655
0.6672	22.0065	13600	1.7403	0.4070	0.9706
0.6124	22.1683	13700	1.8231	0.4120	0.9712
0.6207	22.3301	13800	1.8872	0.4164	0.9787
0.6365	22.4919	13900	1.8193	0.4159	0.9773
0.6139	22.6537	14000	1.8839	0.4244	0.9747
0.6225	22.8155	14100	1.8512	0.4077	0.9746
0.6026	22.9773	14200	1.8275	0.4116	0.9750
0.5744	23.1392	14300	1.8838	0.4163	1.0041
0.5736	23.3010	14400	1.8719	0.4157	0.9769
0.5774	23.4628	14500	1.9813	0.4199	0.9876
0.5854	23.6246	14600	1.9412	0.4160	1.0040
0.5891	23.7864	14700	1.8991	0.4095	0.9727
0.5914	23.9482	14800	1.9801	0.4208	1.0025
0.5638	24.1100	14900	1.9532	0.4208	1.0012
0.5477	24.2718	15000	1.9387	0.4198	0.9923
0.56	24.4337	15100	1.9341	0.4220	0.9730
0.5448	24.5955	15200	1.9103	0.4132	0.9634
0.5617	24.7573	15300	1.9910	0.4251	1.0075
0.5551	24.9191	15400	1.8953	0.4208	0.9870
0.5669	25.0809	15500	2.0860	0.4184	0.9821
0.5121	25.2427	15600	2.1946	0.4166	1.0046
0.5239	25.4045	15700	2.0052	0.4257	0.9813
0.5186	25.5663	15800	2.0627	0.4284	0.9972
0.5192	25.7282	15900	2.1015	0.4222	0.9900
0.5257	25.8900	16000	2.0199	0.4181	0.9713
0.505	26.0518	16100	2.1748	0.4182	0.9982
0.481	26.2136	16200	2.2234	0.4201	0.9821
0.5021	26.3754	16300	2.0915	0.4130	0.9640
0.4931	26.5372	16400	2.2288	0.4213	0.9861
0.4889	26.6990	16500	2.1756	0.4172	0.9983
0.4924	26.8608	16600	2.1217	0.4163	0.9905
0.4778	27.0227	16700	2.2178	0.4180	0.9882
0.4872	27.1845	16800	2.1860	0.4195	0.9839
0.4633	27.3463	16900	2.2207	0.4209	0.9880
0.4718	27.5081	17000	2.2511	0.4193	0.9779
0.4779	27.6699	17100	2.2883	0.4168	0.9749
0.4878	27.8317	17200	2.2165	0.4162	0.9808
0.5061	27.9935	17300	2.2279	0.4198	0.9957
0.4456	28.1553	17400	2.3034	0.4196	0.9781
0.4464	28.3172	17500	2.2417	0.4128	0.9675
0.4453	28.4790	17600	2.3004	0.4147	0.9746
0.4721	28.6408	17700	2.2544	0.4180	1.0
0.4555	28.8026	17800	2.3031	0.4166	0.9854
0.4267	28.9644	17900	2.3252	0.4164	0.9854
0.4154	29.1262	18000	2.3207	0.4155	0.9931
0.468	29.2880	18100	2.3125	0.4149	0.9816
0.4565	29.4498	18200	2.3437	0.4167	0.9917
0.4339	29.6117	18300	2.3618	0.4185	0.9943
0.4242	29.7735	18400	2.3641	0.4172	0.9928
0.4495	29.9353	18500	2.3630	0.4180	0.9952

Framework versions

• Transformers 4.57.2
• Pytorch 2.9.1+cu128
• Datasets 3.6.0
• Tokenizers 0.22.0