wav2vec2-large-xlsr-53-english-ser-cosine

This model is a fine-tuned version of jonatasgrosman/wav2vec2-large-xlsr-53-english on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4677
• Accuracy: 0.8677

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: 0.0001076429938136877
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_steps: 18
• num_epochs: 3.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.7885	0.01	10	1.7963	0.1791
1.7907	0.02	20	1.7973	0.2638
1.8288	0.02	30	1.7546	0.2465
1.7803	0.03	40	1.7500	0.2087
1.7941	0.04	50	1.6953	0.2950
1.7934	0.05	60	1.6342	0.3714
1.6559	0.06	70	1.6199	0.2892
1.6214	0.07	80	1.5400	0.4117
1.5226	0.07	90	1.3802	0.4519
1.4954	0.08	100	1.3506	0.4717
1.4062	0.09	110	1.3328	0.4766
1.4507	0.1	120	1.1985	0.5464
1.2812	0.11	130	1.2826	0.4922
1.1494	0.12	140	1.0960	0.6187
1.1035	0.12	150	1.1925	0.5645
1.2784	0.13	160	1.0955	0.6015
1.0302	0.14	170	1.0418	0.6072
1.0068	0.15	180	0.9261	0.6804
1.112	0.16	190	1.1529	0.5867
1.0308	0.16	200	0.8637	0.7058
1.0464	0.17	210	0.9205	0.6426
0.9531	0.18	220	0.9363	0.6886
1.0228	0.19	230	0.9637	0.6615
1.1446	0.2	240	1.3015	0.5489
1.1146	0.21	250	0.9328	0.6483
0.849	0.21	260	0.8504	0.6746
0.7977	0.22	270	0.9533	0.6697
0.9397	0.23	280	0.9300	0.7083
0.8625	0.24	290	1.1020	0.6401
1.333	0.25	300	0.9816	0.6442
1.0022	0.25	310	0.8472	0.7067
0.8002	0.26	320	0.7866	0.7149
0.8987	0.27	330	0.7979	0.6976
0.9473	0.28	340	0.8600	0.6878
0.9001	0.29	350	0.8141	0.7034
0.9793	0.3	360	0.9872	0.6450
0.9189	0.3	370	0.8561	0.6845
0.9281	0.31	380	0.9055	0.6919
0.7118	0.32	390	0.7937	0.6984
1.0565	0.33	400	0.7339	0.7313
0.8467	0.34	410	0.8262	0.6836
0.9601	0.35	420	0.7464	0.7346
0.8911	0.35	430	0.7229	0.7338
0.9033	0.36	440	0.7393	0.7223
0.8961	0.37	450	0.7272	0.7428
0.7216	0.38	460	0.7183	0.7436
0.6935	0.39	470	0.8003	0.7083
0.7588	0.39	480	0.8471	0.7116
0.8766	0.4	490	0.6976	0.7379
0.6866	0.41	500	0.6806	0.7584
0.6822	0.42	510	0.7669	0.7256
0.7067	0.43	520	0.6885	0.7461
0.6159	0.44	530	0.7020	0.7535
0.8814	0.44	540	0.7478	0.7256
0.7786	0.45	550	0.6302	0.7691
0.6363	0.46	560	0.6745	0.7691
0.8518	0.47	570	0.6242	0.7666
0.8194	0.48	580	0.7154	0.7379
0.6755	0.49	590	0.7056	0.7543
0.7743	0.49	600	0.6823	0.7486
0.6457	0.5	610	0.7160	0.7502
0.4976	0.51	620	0.8222	0.7149
0.929	0.52	630	0.7318	0.7371
0.7981	0.53	640	0.7417	0.7461
0.7243	0.53	650	0.6831	0.7461
0.7332	0.54	660	0.6273	0.7592
0.7827	0.55	670	0.6158	0.7724
0.7733	0.56	680	0.7515	0.7371
0.8527	0.57	690	0.7200	0.7412
0.8355	0.58	700	0.7738	0.7436
0.5383	0.58	710	0.9081	0.7132
1.0851	0.59	720	0.6135	0.7831
0.7345	0.6	730	0.7032	0.7642
0.6648	0.61	740	0.6146	0.7781
0.612	0.62	750	0.6338	0.7732
0.6101	0.62	760	0.6772	0.7740
0.6498	0.63	770	0.7153	0.7601
0.6258	0.64	780	0.7871	0.7329
0.7943	0.65	790	0.6975	0.7691
0.8176	0.66	800	0.7692	0.7313
0.6682	0.67	810	0.5766	0.8012
0.4808	0.67	820	0.5882	0.7847
0.6331	0.68	830	0.5855	0.7896
0.874	0.69	840	0.7082	0.7568
0.8984	0.7	850	0.6078	0.7732
0.5861	0.71	860	0.6469	0.7814
0.6896	0.72	870	0.6997	0.7560
0.8237	0.72	880	0.6279	0.7650
0.5818	0.73	890	0.6763	0.7691
0.4781	0.74	900	0.6867	0.7592
0.6851	0.75	910	0.6142	0.7724
0.455	0.76	920	0.9159	0.7141
0.808	0.76	930	0.7518	0.7617
1.0634	0.77	940	0.6015	0.7839
0.6956	0.78	950	0.5895	0.7872
0.5169	0.79	960	0.6394	0.7773
0.6213	0.8	970	0.6890	0.7699
0.5506	0.81	980	0.7471	0.7560
0.6233	0.81	990	0.6525	0.7872
0.7666	0.82	1000	0.8002	0.7403
0.5644	0.83	1010	0.7067	0.7387
0.6038	0.84	1020	0.6091	0.7823
0.6211	0.85	1030	0.6749	0.7707
0.6758	0.86	1040	0.7102	0.7502
0.7353	0.86	1050	0.6959	0.7560
0.5687	0.87	1060	0.6831	0.7675
0.5606	0.88	1070	0.5945	0.7847
0.7309	0.89	1080	0.6737	0.7412
0.5951	0.9	1090	0.6574	0.7675
0.6062	0.9	1100	0.6740	0.7502
0.9606	0.91	1110	0.5730	0.7839
0.6625	0.92	1120	0.5922	0.7749
0.7908	0.93	1130	0.5652	0.7823
0.6387	0.94	1140	0.5268	0.8118
0.7141	0.95	1150	0.5628	0.7896
0.5587	0.95	1160	0.6479	0.7609
0.4817	0.96	1170	0.5410	0.8044
0.4444	0.97	1180	0.5950	0.8044
0.6776	0.98	1190	0.5993	0.8012
0.5989	0.99	1200	0.5745	0.7987
0.6334	1.0	1210	0.6220	0.7913
0.5216	1.0	1220	0.5936	0.7938
0.5127	1.01	1230	0.6741	0.7839
0.5632	1.02	1240	0.6501	0.7954
0.5335	1.03	1250	0.5721	0.8061
0.511	1.04	1260	0.5630	0.8102
0.5424	1.04	1270	0.5396	0.8135
0.771	1.05	1280	0.5580	0.8012
0.435	1.06	1290	0.5764	0.8036
0.5203	1.07	1300	0.6032	0.7913
0.4689	1.08	1310	0.6431	0.7872
0.481	1.09	1320	0.6019	0.7987
0.5938	1.09	1330	0.6198	0.7938
0.3972	1.1	1340	0.5842	0.8061
0.368	1.11	1350	0.5066	0.8127
0.4644	1.12	1360	0.6058	0.8012
0.6914	1.13	1370	0.5384	0.8217
0.3341	1.13	1380	0.5535	0.8143
0.5301	1.14	1390	0.5916	0.8020
0.5294	1.15	1400	0.6297	0.7938
0.7029	1.16	1410	0.5581	0.8102
0.322	1.17	1420	0.6066	0.7831
0.6871	1.18	1430	0.5141	0.8151
0.4026	1.18	1440	0.6888	0.7716
0.4484	1.19	1450	0.5499	0.8077
0.3767	1.2	1460	0.4825	0.8225
0.4274	1.21	1470	0.4932	0.8274
0.4584	1.22	1480	0.5168	0.8299
0.5741	1.23	1490	0.6384	0.7798
0.3877	1.23	1500	0.5789	0.8044
0.3734	1.24	1510	0.6415	0.7855
0.7986	1.25	1520	0.5575	0.8077
0.5634	1.26	1530	0.5684	0.8143
0.5136	1.27	1540	0.5393	0.8143
0.5331	1.27	1550	0.5203	0.8176
0.2918	1.28	1560	0.5510	0.8151
0.4425	1.29	1570	0.5783	0.8094
0.4245	1.3	1580	0.5433	0.8209
0.3317	1.31	1590	0.5845	0.8085
0.4583	1.32	1600	0.6147	0.7954
0.3298	1.32	1610	0.6249	0.8053
0.5248	1.33	1620	0.5722	0.8094
0.665	1.34	1630	0.5446	0.8217
0.3917	1.35	1640	0.5316	0.8258
0.4321	1.36	1650	0.5598	0.8217
0.3005	1.37	1660	0.6190	0.8151
0.4992	1.37	1670	0.5546	0.8184
0.586	1.38	1680	0.6416	0.7913
0.6481	1.39	1690	0.5324	0.8135
0.4008	1.4	1700	0.5786	0.8012
0.3463	1.41	1710	0.5145	0.8209
0.4994	1.41	1720	0.5650	0.8192
0.4093	1.42	1730	0.5191	0.8365
0.6375	1.43	1740	0.5734	0.8135
0.2303	1.44	1750	0.5447	0.8102
0.4824	1.45	1760	0.5139	0.8250
0.5439	1.46	1770	0.4979	0.8258
0.4751	1.46	1780	0.4896	0.8340
0.534	1.47	1790	0.4656	0.8348
0.4526	1.48	1800	0.5322	0.8316
0.4618	1.49	1810	0.5216	0.8233
0.3825	1.5	1820	0.4792	0.8225
0.4557	1.5	1830	0.5071	0.8118
0.5725	1.51	1840	0.5152	0.8102
0.7004	1.52	1850	0.5080	0.8217
0.4367	1.53	1860	0.4920	0.8357
0.3682	1.54	1870	0.5253	0.8299
0.4411	1.55	1880	0.6186	0.8069
0.5391	1.55	1890	0.5074	0.8283
0.4673	1.56	1900	0.4858	0.8398
0.3542	1.57	1910	0.4767	0.8381
0.6483	1.58	1920	0.4694	0.8373
0.3837	1.59	1930	0.4678	0.8472
0.363	1.6	1940	0.4684	0.8463
0.6446	1.6	1950	0.4696	0.8365
0.5627	1.61	1960	0.4651	0.8472
0.3733	1.62	1970	0.5138	0.8291
0.5972	1.63	1980	0.5244	0.8250
0.2388	1.64	1990	0.5020	0.8266
0.6279	1.64	2000	0.5865	0.8118
0.5827	1.65	2010	0.5717	0.8176
0.4598	1.66	2020	0.4691	0.8439
0.3817	1.67	2030	0.5084	0.8340
0.2973	1.68	2040	0.4568	0.8447
0.4039	1.69	2050	0.4681	0.8505
0.4572	1.69	2060	0.4718	0.8389
0.3481	1.7	2070	0.4849	0.8283
0.4553	1.71	2080	0.4574	0.8414
0.4055	1.72	2090	0.4640	0.8463
0.4384	1.73	2100	0.5049	0.8431
0.5593	1.74	2110	0.5192	0.8513
0.3486	1.74	2120	0.4764	0.8480
0.4698	1.75	2130	0.4858	0.8447
0.211	1.76	2140	0.4976	0.8398
0.5209	1.77	2150	0.4934	0.8472
0.4281	1.78	2160	0.4714	0.8578
0.3902	1.78	2170	0.4863	0.8463
0.3083	1.79	2180	0.4807	0.8431
0.4642	1.8	2190	0.4712	0.8472
0.2382	1.81	2200	0.4641	0.8513
0.4154	1.82	2210	0.4900	0.8447
0.3637	1.83	2220	0.4790	0.8488
0.4864	1.83	2230	0.4742	0.8513
0.5024	1.84	2240	0.4803	0.8529
0.4139	1.85	2250	0.4672	0.8521
0.4131	1.86	2260	0.4895	0.8431
0.4851	1.87	2270	0.4432	0.8529
0.3846	1.88	2280	0.4417	0.8422
0.3778	1.88	2290	0.4477	0.8439
0.4128	1.89	2300	0.4336	0.8513
0.3755	1.9	2310	0.4678	0.8439
0.4672	1.91	2320	0.4740	0.8373
0.5216	1.92	2330	0.4343	0.8472
0.3469	1.92	2340	0.4542	0.8316
0.3283	1.93	2350	0.4587	0.8447
0.3495	1.94	2360	0.5050	0.8348
0.4518	1.95	2370	0.5309	0.8266
0.3023	1.96	2380	0.5113	0.8332
0.4014	1.97	2390	0.4989	0.8332
0.4963	1.97	2400	0.4539	0.8505
0.3421	1.98	2410	0.4889	0.8455
0.4126	1.99	2420	0.4696	0.8463
0.479	2.0	2430	0.4498	0.8513
0.3319	2.01	2440	0.4686	0.8488
0.2787	2.01	2450	0.4650	0.8447
0.2105	2.02	2460	0.4665	0.8505
0.4944	2.03	2470	0.4667	0.8488
0.2236	2.04	2480	0.4678	0.8463
0.3076	2.05	2490	0.4621	0.8513
0.2813	2.06	2500	0.4451	0.8562
0.2207	2.06	2510	0.4559	0.8562
0.3693	2.07	2520	0.4634	0.8513
0.3682	2.08	2530	0.4390	0.8562
0.2618	2.09	2540	0.4417	0.8529
0.3139	2.1	2550	0.4618	0.8529
0.1739	2.11	2560	0.4938	0.8488
0.4258	2.11	2570	0.4574	0.8496
0.2136	2.12	2580	0.4495	0.8529
0.2625	2.13	2590	0.4555	0.8570
0.3161	2.14	2600	0.4696	0.8537
0.2515	2.15	2610	0.4649	0.8661
0.3097	2.15	2620	0.4474	0.8685
0.3544	2.16	2630	0.4458	0.8603
0.2967	2.17	2640	0.4555	0.8669
0.4015	2.18	2650	0.4486	0.8652
0.079	2.19	2660	0.4624	0.8620
0.1754	2.2	2670	0.4805	0.8587
0.1854	2.2	2680	0.4803	0.8628
0.3181	2.21	2690	0.4792	0.8595
0.0808	2.22	2700	0.4740	0.8628
0.2027	2.23	2710	0.4846	0.8587
0.3211	2.24	2720	0.5074	0.8505
0.2448	2.25	2730	0.5276	0.8414
0.3618	2.25	2740	0.5133	0.8488
0.1822	2.26	2750	0.5002	0.8578
0.3095	2.27	2760	0.4827	0.8603
0.0762	2.28	2770	0.4792	0.8644
0.187	2.29	2780	0.4897	0.8644
0.5779	2.29	2790	0.4901	0.8652
0.292	2.3	2800	0.4764	0.8603
0.1865	2.31	2810	0.4798	0.8644
0.3594	2.32	2820	0.4837	0.8620
0.421	2.33	2830	0.4812	0.8562
0.1173	2.34	2840	0.4708	0.8603
0.278	2.34	2850	0.4693	0.8685
0.2294	2.35	2860	0.4724	0.8628
0.243	2.36	2870	0.4749	0.8620
0.3979	2.37	2880	0.4633	0.8628
0.4518	2.38	2890	0.4603	0.8669
0.2739	2.38	2900	0.4625	0.8685
0.1782	2.39	2910	0.4652	0.8677
0.3536	2.4	2920	0.4613	0.8644
0.0904	2.41	2930	0.4642	0.8611
0.2315	2.42	2940	0.4613	0.8661
0.1236	2.43	2950	0.4628	0.8652
0.1842	2.43	2960	0.4706	0.8620
0.2414	2.44	2970	0.4683	0.8652
0.3419	2.45	2980	0.4645	0.8677
0.2877	2.46	2990	0.4657	0.8636
0.2524	2.47	3000	0.4701	0.8652
0.1731	2.48	3010	0.4733	0.8644
0.1731	2.48	3020	0.4830	0.8595
0.0921	2.49	3030	0.4904	0.8603
0.1593	2.5	3040	0.4836	0.8595
0.467	2.51	3050	0.4706	0.8628
0.4225	2.52	3060	0.4598	0.8644
0.1251	2.52	3070	0.4511	0.8694
0.2181	2.53	3080	0.4487	0.8735
0.2247	2.54	3090	0.4452	0.8767
0.3722	2.55	3100	0.4469	0.8759
0.1069	2.56	3110	0.4536	0.8735
0.2174	2.57	3120	0.4571	0.8710
0.2586	2.57	3130	0.4626	0.8685
0.2803	2.58	3140	0.4665	0.8677
0.4484	2.59	3150	0.4581	0.8694
0.3104	2.6	3160	0.4539	0.8735
0.2411	2.61	3170	0.4531	0.8726
0.2157	2.62	3180	0.4565	0.8694
0.2342	2.62	3190	0.4549	0.8694
0.2921	2.63	3200	0.4570	0.8677
0.1988	2.64	3210	0.4590	0.8677
0.2142	2.65	3220	0.4601	0.8661
0.1666	2.66	3230	0.4652	0.8661
0.2296	2.66	3240	0.4709	0.8611
0.3847	2.67	3250	0.4676	0.8636
0.4149	2.68	3260	0.4654	0.8636
0.2602	2.69	3270	0.4614	0.8661
0.3786	2.7	3280	0.4605	0.8661
0.3509	2.71	3290	0.4590	0.8661
0.2254	2.71	3300	0.4564	0.8677
0.1775	2.72	3310	0.4553	0.8694
0.2269	2.73	3320	0.4546	0.8669
0.1792	2.74	3330	0.4549	0.8644
0.1107	2.75	3340	0.4580	0.8661
0.2062	2.75	3350	0.4598	0.8636
0.1641	2.76	3360	0.4621	0.8652
0.18	2.77	3370	0.4651	0.8652
0.0959	2.78	3380	0.4673	0.8661
0.217	2.79	3390	0.4672	0.8652
0.3293	2.8	3400	0.4673	0.8644
0.2691	2.8	3410	0.4669	0.8644
0.1945	2.81	3420	0.4659	0.8652
0.2712	2.82	3430	0.4660	0.8677
0.2287	2.83	3440	0.4663	0.8677
0.2103	2.84	3450	0.4661	0.8669
0.2713	2.85	3460	0.4663	0.8669
0.3182	2.85	3470	0.4665	0.8677
0.1698	2.86	3480	0.4668	0.8669
0.2663	2.87	3490	0.4669	0.8677
0.2091	2.88	3500	0.4670	0.8685
0.1406	2.89	3510	0.4677	0.8669
0.16	2.89	3520	0.4682	0.8661
0.1413	2.9	3530	0.4686	0.8661
0.3499	2.91	3540	0.4690	0.8661
0.205	2.92	3550	0.4688	0.8661
0.3849	2.93	3560	0.4684	0.8661
0.209	2.94	3570	0.4680	0.8669
0.1985	2.94	3580	0.4678	0.8677
0.1989	2.95	3590	0.4678	0.8677
0.2031	2.96	3600	0.4677	0.8677
0.2401	2.97	3610	0.4677	0.8677
0.2717	2.98	3620	0.4678	0.8677
0.2821	2.99	3630	0.4678	0.8677
0.1735	2.99	3640	0.4677	0.8677

Framework versions

• Transformers 4.40.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.1.dev0
• Tokenizers 0.15.2