blt-text-reversal

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

• Loss: 2.2099
• Char Accuracy: 0.0073
• Sequence Accuracy: 0.0
• Edit Distance: 117.8885

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: 0.0001
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Char Accuracy	Sequence Accuracy	Edit Distance
3.9289	1.0	79	3.4203	0.0497	0.0	130.3595
3.2214	2.0	158	3.0127	0.0	0.0	123.831
2.9326	3.0	237	2.8196	0.0	0.0	121.3795
2.7944	4.0	316	2.7267	0.0	0.0	121.179
2.7204	5.0	395	2.6716	0.0001	0.0	120.889
2.6723	6.0	474	2.6335	0.0001	0.0	120.652
2.6379	7.0	553	2.6048	0.0002	0.0	120.4915
2.6126	8.0	632	2.5820	0.0004	0.0	120.372
2.5905	9.0	711	2.5628	0.0006	0.0	120.348
2.5713	10.0	790	2.5460	0.0014	0.0	120.2635
2.5554	11.0	869	2.5310	0.0018	0.0	120.198
2.5395	12.0	948	2.5166	0.0017	0.0	120.044
2.525	13.0	1027	2.5025	0.0025	0.0	119.9965
2.5104	14.0	1106	2.4892	0.0022	0.0	119.8795
2.4977	15.0	1185	2.4762	0.0029	0.0	119.825
2.4843	16.0	1264	2.4640	0.0031	0.0	119.642
2.4722	17.0	1343	2.4523	0.0030	0.0	119.6285
2.4602	18.0	1422	2.4406	0.0027	0.0	119.5435
2.4478	19.0	1501	2.4286	0.0034	0.0	119.5165
2.4364	20.0	1580	2.4174	0.0033	0.0	119.513
2.4242	21.0	1659	2.4065	0.0034	0.0	119.4225
2.4129	22.0	1738	2.3957	0.0037	0.0	119.3935
2.4021	23.0	1817	2.3860	0.0039	0.0	119.3675
2.3918	24.0	1896	2.3759	0.0039	0.0	119.343
2.3824	25.0	1975	2.3662	0.0042	0.0	119.2295
2.3729	26.0	2054	2.3574	0.0042	0.0	119.172
2.3641	27.0	2133	2.3487	0.0043	0.0	119.11
2.3556	28.0	2212	2.3405	0.0049	0.0	119.0445
2.3459	29.0	2291	2.3326	0.0052	0.0	118.9715
2.339	30.0	2370	2.3247	0.0057	0.0	118.9435
2.3305	31.0	2449	2.3168	0.0050	0.0	118.8695
2.3224	32.0	2528	2.3098	0.0059	0.0	118.85
2.3152	33.0	2607	2.3026	0.0059	0.0	118.759
2.3079	34.0	2686	2.2960	0.0061	0.0	118.7105
2.3012	35.0	2765	2.2893	0.0058	0.0	118.671
2.2949	36.0	2844	2.2825	0.0060	0.0	118.586
2.2885	37.0	2923	2.2768	0.0061	0.0	118.476
2.2806	38.0	3002	2.2708	0.0063	0.0	118.532
2.2749	39.0	3081	2.2652	0.0065	0.0	118.4015
2.2684	40.0	3160	2.2601	0.0065	0.0	118.3975
2.2635	41.0	3239	2.2543	0.007	0.0	118.2825
2.2587	42.0	3318	2.2488	0.0070	0.0	118.247
2.2532	43.0	3397	2.2442	0.0071	0.0	118.175
2.2472	44.0	3476	2.2393	0.0065	0.0	118.1065
2.2414	45.0	3555	2.2350	0.0070	0.0	118.0595
2.2369	46.0	3634	2.2302	0.0072	0.0	118.0635
2.2324	47.0	3713	2.2260	0.0070	0.0	118.0125
2.2281	48.0	3792	2.2220	0.0067	0.0	118.051
2.2234	49.0	3871	2.2178	0.0070	0.0	117.994
2.2192	50.0	3950	2.2133	0.0068	0.0	117.888
2.2145	51.0	4029	2.2099	0.0073	0.0	117.8885

Framework versions

• Transformers 4.55.4
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4