alakxender/t5-dhivehi-title-generation-xs

T5 Dhivehi Title Generator

This model is a fine-tuned version of t5-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.1178
• Rouge1: 6.7126
• Rouge2: 1.3302
• Rougel: 6.7126
• Rougelsum: 6.6582
• Gen Len: 14.5499

Model description

This project trains a T5 model to generate Dhivehi article titles from content. The model is trained on the Dhivehi news dataset.

Usage

# Load model and tokenizer
MODEL_DIR = "alakxender/t5-dhivehi-title-generation-xs"
tokenizer = AutoTokenizer.from_pretrained(MODEL_DIR)
model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_DIR)

prefix = "2title: "
max_input_length = 512
max_target_length = 64

def generate_title(content):
    # Prepend prefix as in training
    input_text = prefix + content.strip()
    # Tokenize input
    inputs = tokenizer(
        input_text,
        max_length=max_input_length,
        truncation=True,
        return_tensors="pt"
    )
    # Move to GPU if available
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    model.to(device)
    inputs = {k: v.to(device) for k, v in inputs.items()}
    # Generate title
    with torch.no_grad():
        outputs = model.generate(
            input_ids=inputs["input_ids"],
            attention_mask=inputs["attention_mask"],
            max_length=max_target_length,
            num_beams=4,
            early_stopping=True,
            no_repeat_ngram_size=2,
        )
    # Decode output
    title = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return title

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 4e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
6.8792	0.0153	100	6.2230	1.5222	0.3415	1.5379	1.5177	10.388
6.2112	0.0305	200	5.9101	3.603	0.6567	3.5977	3.5814	12.1938
6.2513	0.0458	300	5.7717	5.6171	1.1147	5.6297	5.5875	16.258
5.96	0.0611	400	5.6119	5.2758	1.0352	5.2966	5.2529	15.5432
5.9496	0.0763	500	5.5386	5.9752	1.2664	5.963	5.9389	16.4105
5.8846	0.0916	600	5.4535	5.6049	1.0475	5.6204	5.5958	15.9775
5.6616	0.1069	700	5.4477	5.6402	1.1102	5.6451	5.6403	16.3292
5.7413	0.1221	800	5.3318	5.468	0.9484	5.4792	5.4518	15.3366
5.5988	0.1374	900	5.3000	5.3153	1.0083	5.3384	5.307	15.436
5.6162	0.1527	1000	5.2692	5.7038	1.0883	5.7273	5.6831	16.0608
5.5271	0.1679	1100	5.2247	5.5247	1.0133	5.5113	5.4821	15.9117
5.3815	0.1832	1200	5.2153	5.6452	1.0128	5.6405	5.6201	15.477
5.5763	0.1985	1300	5.1487	5.7312	1.0413	5.7345	5.6886	15.9016
5.5032	0.2137	1400	5.1644	5.7586	1.0508	5.7547	5.7252	16.3225
5.3544	0.2290	1500	5.1333	5.7043	1.0385	5.7149	5.6785	15.9197
5.4919	0.2443	1600	5.0498	5.8567	1.1326	5.8477	5.8144	15.7467
5.4061	0.2595	1700	5.0564	6.0958	1.1992	6.1058	6.0762	15.8132
5.4369	0.2748	1800	5.0176	5.4905	1.1141	5.5136	5.4786	15.2778
5.38	0.2901	1900	5.0210	5.5841	1.0581	5.5969	5.5559	15.3349
5.3269	0.3053	2000	5.0104	5.9337	1.1589	5.9582	5.9032	15.7377
5.437	0.3206	2100	4.9810	5.7737	1.3044	5.786	5.7351	15.5445
5.3303	0.3359	2200	4.9908	6.0089	1.1639	6.0111	5.9588	15.914
5.1874	0.3511	2300	4.9635	5.9637	1.1611	5.9818	5.9229	15.524
5.328	0.3664	2400	4.9479	5.8018	1.0592	5.8053	5.7504	15.8001
5.2337	0.3817	2500	4.9101	5.7631	1.1091	5.744	5.7028	15.4098
5.2394	0.3969	2600	4.9294	5.8914	1.2143	5.8931	5.846	15.7548
5.1938	0.4122	2700	4.9052	5.8297	1.2149	5.8493	5.7838	15.6987
5.3153	0.4275	2800	4.9112	5.9811	1.2093	5.9879	5.9273	15.8273
5.1655	0.4427	2900	4.8779	5.8308	1.2429	5.8331	5.7771	15.5677
5.2144	0.4580	3000	4.8754	5.84	1.2675	5.8366	5.7964	15.6208
5.0254	0.4733	3100	4.8659	5.6481	1.1158	5.636	5.6003	15.2963
5.1746	0.4885	3200	4.8736	5.4809	1.0609	5.4865	5.4412	14.9822
5.0947	0.5038	3300	4.8381	5.7212	1.2541	5.7284	5.6803	15.7235
5.1984	0.5191	3400	4.8377	5.8196	1.2115	5.8237	5.7856	15.6745
5.086	0.5344	3500	4.8386	6.0767	1.1947	6.0795	6.0219	15.5573
5.1677	0.5496	3600	4.8163	5.7754	1.1309	5.7586	5.7239	15.2026
5.1565	0.5649	3700	4.8198	5.7452	1.1141	5.7558	5.7071	15.3302
5.1513	0.5802	3800	4.8325	6.0296	1.1589	6.0027	5.9635	15.646
5.1446	0.5954	3900	4.7922	5.7831	1.2535	5.7801	5.7105	15.5603
5.1283	0.6107	4000	4.7761	5.6494	1.1197	5.6298	5.5878	15.2966
5.1421	0.6260	4100	4.7721	5.6959	1.2289	5.6987	5.6466	15.5411
4.9729	0.6412	4200	4.7814	5.8167	1.1107	5.8075	5.764	15.3611
5.1059	0.6565	4300	4.7585	5.7731	1.0469	5.7552	5.7267	14.8764
5.2112	0.6718	4400	4.7571	5.968	1.2216	5.9656	5.9329	15.6198
5.1182	0.6870	4500	4.7341	5.7995	1.1477	5.7877	5.7552	15.1485
5.1151	0.7023	4600	4.7535	5.903	1.2261	5.8946	5.851	15.1374
4.9995	0.7176	4700	4.7391	6.1057	1.2261	6.1091	6.0822	15.1488
4.9861	0.7328	4800	4.6981	6.1931	1.3078	6.1941	6.1601	15.3426
5.1464	0.7481	4900	4.7197	6.2321	1.2367	6.2086	6.163	15.5875
5.0899	0.7634	5000	4.7098	5.88	1.1958	5.8642	5.8184	15.1031
5.0756	0.7786	5100	4.6941	5.8947	1.1846	5.8942	5.8634	15.563
4.9396	0.7939	5200	4.6887	5.87	1.1779	5.8552	5.8198	15.648
5.0037	0.8092	5300	4.6995	6.1205	1.235	6.1121	6.0719	15.8052
4.8821	0.8244	5400	4.7036	6.0943	1.1986	6.0915	6.0568	15.477
4.9831	0.8397	5500	4.6668	6.081	1.2165	6.0771	6.0197	15.3184
5.0832	0.8550	5600	4.6789	5.8459	1.1001	5.8303	5.7894	14.9768
4.9712	0.8702	5700	4.6673	6.0228	1.1757	6.001	5.9557	15.5284
4.9886	0.8855	5800	4.6487	6.0868	1.2126	6.0487	6.0095	15.5364
4.9188	0.9008	5900	4.6578	5.9387	1.1981	5.9221	5.8878	15.3336
4.8827	0.9160	6000	4.6554	6.1011	1.2988	6.1009	6.0389	15.6984
5.0167	0.9313	6100	4.6413	6.063	1.2529	6.0292	5.9976	15.4814
4.8871	0.9466	6200	4.6498	6.15	1.2574	6.1426	6.0855	15.614
5.0042	0.9618	6300	4.6219	6.0292	1.2888	5.9981	5.9505	15.4887
5.0407	0.9771	6400	4.6249	6.3075	1.3649	6.2851	6.2358	15.6127
4.8514	0.9924	6500	4.6060	6.0714	1.2988	6.0511	6.0248	15.2882
4.8981	1.0076	6600	4.6131	6.1224	1.2977	6.0942	6.0546	15.3299
4.8712	1.0229	6700	4.6342	6.2696	1.2809	6.2687	6.2233	15.692
4.835	1.0382	6800	4.6174	6.1119	1.2082	6.0963	6.0567	15.354
4.8393	1.0534	6900	4.6030	6.1998	1.2468	6.1806	6.1253	15.5882
4.8345	1.0687	7000	4.5936	6.1975	1.2597	6.1727	6.1259	15.2553
4.9471	1.0840	7100	4.5985	5.9431	1.2205	5.9236	5.8934	15.4803
4.7736	1.0992	7200	4.6057	5.9228	1.169	5.8869	5.8684	15.5623
4.8208	1.1145	7300	4.5962	6.1235	1.151	6.099	6.062	15.1189
4.8049	1.1298	7400	4.5840	6.1604	1.1869	6.1383	6.0918	15.4699
5.0068	1.1450	7500	4.5652	6.3519	1.2507	6.3271	6.2789	15.657
4.8249	1.1603	7600	4.5739	6.1874	1.2608	6.1861	6.1324	15.6174
4.6873	1.1756	7700	4.5904	6.2498	1.3212	6.2449	6.1884	15.3255
4.902	1.1908	7800	4.5651	5.9632	1.2765	5.9437	5.9003	15.0269
4.8465	1.2061	7900	4.5635	6.1311	1.207	6.1209	6.0586	14.9506
4.8141	1.2214	8000	4.5691	6.3752	1.2653	6.3502	6.3052	15.4236
4.8471	1.2366	8100	4.5660	5.9979	1.1421	5.9568	5.9275	15.3134
4.6808	1.2519	8200	4.5571	6.1718	1.2121	6.1272	6.0931	15.6735
4.8067	1.2672	8300	4.5270	5.768	1.2798	5.7439	5.7203	14.8989
4.6854	1.2824	8400	4.5613	6.2097	1.2457	6.1692	6.1195	15.4132
4.8532	1.2977	8500	4.5363	6.2456	1.2457	6.2427	6.1689	15.0635
4.8005	1.3130	8600	4.5305	6.1416	1.3072	6.1155	6.0648	14.9039
4.771	1.3282	8700	4.5426	6.0616	1.1813	6.0572	6.0088	14.9731
4.7723	1.3435	8800	4.5336	6.1104	1.2429	6.0727	6.047	15.1814
4.7678	1.3588	8900	4.5253	6.2145	1.2932	6.2137	6.1509	15.0826
4.7618	1.3740	9000	4.5094	6.3622	1.3184	6.3504	6.3025	15.264
4.8751	1.3893	9100	4.5094	6.4225	1.3257	6.4038	6.364	15.4968
4.8887	1.4046	9200	4.5098	6.2097	1.2641	6.1888	6.1606	15.3077
4.8423	1.4198	9300	4.5017	6.3233	1.2585	6.2944	6.2522	15.128
4.7879	1.4351	9400	4.5019	6.3288	1.2809	6.3189	6.2621	15.043
4.7339	1.4504	9500	4.4938	6.2308	1.2249	6.2133	6.1738	15.1512
4.8811	1.4656	9600	4.5066	6.3685	1.2944	6.3431	6.2949	15.5237
4.7279	1.4809	9700	4.4954	6.5019	1.2977	6.49	6.4403	15.4182
4.7096	1.4962	9800	4.4883	6.3861	1.3442	6.3833	6.336	15.2684
4.6617	1.5115	9900	4.5191	6.7083	1.3727	6.6988	6.6599	15.8404
4.8223	1.5267	10000	4.4808	6.539	1.3879	6.5501	6.5074	15.3221
4.7611	1.5420	10100	4.4934	6.5553	1.3352	6.5457	6.5029	15.5566
4.6046	1.5573	10200	4.4904	6.5773	1.3789	6.5645	6.5398	15.4273
4.8756	1.5725	10300	4.4531	6.3349	1.2893	6.3325	6.2908	14.8586
4.6439	1.5878	10400	4.4774	6.5523	1.3722	6.5421	6.5079	15.5059
4.8026	1.6031	10500	4.4421	6.093	1.333	6.0677	6.046	14.8586
4.8174	1.6183	10600	4.4513	6.2914	1.3268	6.267	6.2451	14.9083
4.7608	1.6336	10700	4.4491	6.183	1.2988	6.1667	6.1335	14.9258
4.5969	1.6489	10800	4.4501	6.3943	1.3268	6.3758	6.3571	15.342
4.597	1.6641	10900	4.4526	6.3685	1.3145	6.3616	6.3184	15.0433
4.7094	1.6794	11000	4.4451	6.5296	1.3476	6.5385	6.501	15.2264
4.8282	1.6947	11100	4.4425	6.5981	1.2927	6.5781	6.54	15.3366
4.7397	1.7099	11200	4.4419	6.6182	1.3252	6.5956	6.5573	15.3305
4.6132	1.7252	11300	4.4412	6.2149	1.3201	6.2154	6.1678	14.7978
4.606	1.7405	11400	4.4582	6.4116	1.2569	6.4039	6.37	15.2237
4.733	1.7557	11500	4.4458	6.7438	1.3599	6.7167	6.6863	15.484
4.6921	1.7710	11600	4.4251	6.6062	1.3666	6.6096	6.5637	15.1545
4.6288	1.7863	11700	4.4362	6.6562	1.2921	6.6158	6.5855	15.5304
4.6309	1.8015	11800	4.4303	6.5864	1.3487	6.5712	6.5226	15.1676
4.6784	1.8168	11900	4.4263	6.5033	1.3548	6.4797	6.4476	15.2731
4.691	1.8321	12000	4.4242	6.3681	1.31	6.3831	6.3303	15.1555
4.6812	1.8473	12100	4.4420	6.6432	1.3705	6.6152	6.5589	15.7424
4.7354	1.8626	12200	4.4091	6.641	1.3134	6.5928	6.5603	15.5012
4.7835	1.8779	12300	4.4131	6.5701	1.2876	6.5344	6.4984	15.2865
4.6506	1.8931	12400	4.4276	6.6217	1.3089	6.6032	6.5719	15.4363
4.7024	1.9084	12500	4.4053	6.7856	1.4203	6.7505	6.7131	15.2879
4.8208	1.9237	12600	4.3992	6.5757	1.366	6.5383	6.5043	15.4269
4.6342	1.9389	12700	4.4129	6.3757	1.3459	6.3472	6.3218	15.3557
4.5243	1.9542	12800	4.4207	6.7337	1.4612	6.7134	6.685	15.7531
4.5997	1.9695	12900	4.3912	6.3883	1.3134	6.3927	6.3396	14.7172
4.5639	1.9847	13000	4.3915	6.5787	1.3588	6.5673	6.5458	15.1226
4.5789	2.0	13100	4.4070	6.4433	1.2636	6.4038	6.4	15.1219
4.5501	2.0153	13200	4.4154	6.6731	1.2697	6.6372	6.6105	15.4733
4.507	2.0305	13300	4.3930	6.5478	1.2731	6.5277	6.5058	15.0776
4.5843	2.0458	13400	4.3939	6.6208	1.2417	6.6051	6.5737	14.8982
4.6278	2.0611	13500	4.3859	6.905	1.3369	6.8699	6.8469	15.0118
4.5767	2.0763	13600	4.3740	6.746	1.2977	6.7001	6.665	15.4498
4.5453	2.0916	13700	4.3710	6.7645	1.2977	6.7207	6.7091	15.3275
4.5142	2.1069	13800	4.3808	6.8612	1.3201	6.8265	6.7862	15.1065
4.5548	2.1221	13900	4.3748	6.6404	1.2373	6.623	6.6017	15.0487
4.5441	2.1374	14000	4.3750	6.7119	1.2876	6.6838	6.6686	15.6406
4.7168	2.1527	14100	4.3751	6.8014	1.2429	6.7697	6.752	15.1371
4.592	2.1679	14200	4.3609	6.7528	1.2765	6.7221	6.6816	15.1082
4.5659	2.1832	14300	4.3562	6.6408	1.2765	6.6227	6.5832	15.3386
4.4912	2.1985	14400	4.3563	6.504	1.272	6.463	6.4333	14.9573
4.5699	2.2137	14500	4.3641	6.6389	1.2541	6.6246	6.5645	15.1592
4.522	2.2290	14600	4.3617	6.5548	1.3078	6.5333	6.504	15.3282
4.5608	2.2443	14700	4.3437	6.4075	1.2361	6.3814	6.3495	15.0581
4.6078	2.2595	14800	4.3447	6.6507	1.3431	6.6228	6.5793	14.997
4.6447	2.2748	14900	4.3500	6.4309	1.2305	6.4301	6.4068	15.2744
4.5038	2.2901	15000	4.3559	6.5807	1.2854	6.565	6.5233	15.3325
4.5696	2.3053	15100	4.3388	6.5562	1.2082	6.5324	6.4952	15.3013
4.553	2.3206	15200	4.3411	6.5835	1.2076	6.5568	6.5169	15.3191
4.4805	2.3359	15300	4.3404	6.4523	1.2916	6.4523	6.4045	14.6725
4.5343	2.3511	15400	4.3430	6.5648	1.1981	6.5345	6.4957	15.2536
4.4583	2.3664	15500	4.3379	6.7576	1.2641	6.7462	6.707	15.1861
4.5561	2.3817	15600	4.3170	6.5999	1.2569	6.5793	6.5329	15.2916
4.4905	2.3969	15700	4.3196	6.7578	1.2759	6.745	6.7134	15.2019
4.5696	2.4122	15800	4.3089	6.4109	1.2524	6.3958	6.3468	15.1347
4.3328	2.4275	15900	4.3565	6.877	1.2921	6.863	6.8373	15.8095
4.4544	2.4427	16000	4.3196	6.5505	1.2977	6.5455	6.4926	15.0558
4.5818	2.4580	16100	4.2998	6.6742	1.3201	6.6554	6.6078	14.9049
4.4936	2.4733	16200	4.3199	6.735	1.3336	6.7248	6.6883	15.4276
4.694	2.4885	16300	4.3144	6.4066	1.2205	6.3941	6.3397	15.1962
4.6011	2.5038	16400	4.2934	6.5654	1.2367	6.5441	6.5052	15.0907
4.5221	2.5191	16500	4.3122	6.5998	1.3005	6.6008	6.5677	15.0144
4.4675	2.5344	16600	4.3183	6.3583	1.1858	6.364	6.302	14.9254
4.4892	2.5496	16700	4.3069	6.6041	1.2681	6.5904	6.5453	15.1229
4.5587	2.5649	16800	4.3056	6.5436	1.1981	6.5439	6.4774	15.084
4.4603	2.5802	16900	4.2991	6.5008	1.2249	6.4801	6.4391	15.1884
4.4386	2.5954	17000	4.3124	6.7682	1.3257	6.751	6.7107	15.5902
4.4939	2.6107	17100	4.2928	6.7126	1.2535	6.7007	6.6619	14.995
4.5578	2.6260	17200	4.3008	6.7853	1.291	6.7572	6.7163	15.1085
4.548	2.6412	17300	4.2955	6.7035	1.3033	6.6839	6.6216	15.0729
4.5687	2.6565	17400	4.2840	6.7376	1.342	6.7083	6.6674	15.174
4.5676	2.6718	17500	4.2866	6.8067	1.342	6.7846	6.7284	15.2358
4.4962	2.6870	17600	4.2774	6.813	1.347	6.7891	6.7585	15.2241
4.4102	2.7023	17700	4.2897	6.896	1.3772	6.885	6.8506	15.2288
4.5514	2.7176	17800	4.2729	6.5213	1.3408	6.5006	6.4644	14.9678
4.5487	2.7328	17900	4.2841	6.8151	1.3537	6.7789	6.7645	15.1014
4.4603	2.7481	18000	4.2675	6.7447	1.3291	6.7285	6.6885	15.2304
4.4844	2.7634	18100	4.2974	6.9162	1.3974	6.8955	6.873	15.5049
4.4814	2.7786	18200	4.2635	6.6742	1.2753	6.66	6.6206	15.2267
4.4804	2.7939	18300	4.2828	6.5329	1.2529	6.5296	6.4855	15.3006
4.4999	2.8092	18400	4.2564	6.4838	1.2888	6.4725	6.4304	15.1619
4.439	2.8244	18500	4.2534	6.4984	1.3089	6.4757	6.4482	14.9792
4.3627	2.8397	18600	4.2706	6.5341	1.2927	6.5299	6.4734	15.0564
4.4925	2.8550	18700	4.2668	6.7546	1.2406	6.7416	6.6998	15.1505
4.4381	2.8702	18800	4.2654	6.8178	1.3257	6.8061	6.749	15.26
4.5014	2.8855	18900	4.2452	6.764	1.2854	6.7451	6.7065	15.1092
4.4463	2.9008	19000	4.2363	6.6859	1.2619	6.6847	6.6463	14.7299
4.486	2.9160	19100	4.2524	6.6067	1.1802	6.5808	6.5477	14.735
4.6934	2.9313	19200	4.2481	6.8828	1.2899	6.8595	6.8266	15.1602
4.4932	2.9466	19300	4.2391	6.7007	1.2026	6.6931	6.6468	14.8922
4.4375	2.9618	19400	4.2417	6.7496	1.2921	6.7383	6.6979	15.0752
4.4737	2.9771	19500	4.2271	6.6732	1.2227	6.6617	6.623	14.8959
4.4438	2.9924	19600	4.2314	6.8693	1.3044	6.8388	6.781	14.9372
4.5884	3.0076	19700	4.2391	6.5369	1.2524	6.5486	6.4994	14.7951
4.4385	3.0229	19800	4.2494	6.5711	1.2613	6.6066	6.5469	14.9829
4.3336	3.0382	19900	4.2422	6.8137	1.2529	6.8106	6.7544	15.43
4.3949	3.0534	20000	4.2314	6.5257	1.1919	6.5169	6.4449	14.8011
4.4021	3.0687	20100	4.2273	6.8886	1.3201	6.8936	6.8498	14.997
4.3833	3.0840	20200	4.2230	6.6811	1.3257	6.6993	6.6463	14.8942
4.5106	3.0992	20300	4.2517	6.735	1.3705	6.72	6.6882	15.1972
4.419	3.1145	20400	4.2412	6.7548	1.3537	6.7456	6.7027	15.2849
4.3442	3.1298	20500	4.2246	6.7109	1.3699	6.6963	6.6537	15.1686
4.3127	3.1450	20600	4.2300	6.8341	1.3369	6.8137	6.7662	14.9486
4.4789	3.1603	20700	4.2224	7.0101	1.3604	6.9831	6.9543	15.474
4.3206	3.1756	20800	4.2099	6.7742	1.342	6.747	6.7271	15.1589
4.3617	3.1908	20900	4.2143	6.7806	1.3246	6.7572	6.7425	15.0621
4.3868	3.2061	21000	4.2233	6.7806	1.342	6.7546	6.7534	14.8109
4.4374	3.2214	21100	4.2072	7.0038	1.4136	6.9894	6.9823	15.1334
4.4511	3.2366	21200	4.2118	6.7893	1.3593	6.796	6.764	14.8294
4.4317	3.2519	21300	4.2146	6.7398	1.2865	6.7282	6.7036	14.7659
4.3243	3.2672	21400	4.2152	6.7654	1.2966	6.7402	6.7022	14.7511
4.3431	3.2824	21500	4.2206	6.6872	1.3263	6.6611	6.6354	15.1965
4.3489	3.2977	21600	4.2020	6.6902	1.2294	6.675	6.6414	15.2019
4.4375	3.3130	21700	4.1962	6.7358	1.2697	6.7306	6.6902	15.0537
4.3343	3.3282	21800	4.2023	6.9761	1.3033	6.9549	6.9393	15.1444
4.3673	3.3435	21900	4.1960	6.7354	1.2367	6.7326	6.6823	14.8018
4.3379	3.3588	22000	4.2064	6.6404	1.2026	6.6335	6.5868	14.8418
4.4552	3.3740	22100	4.1917	6.6723	1.3156	6.6613	6.6148	14.9476
4.3742	3.3893	22200	4.2008	7.0879	1.3649	7.0845	7.0532	15.1058
4.2247	3.4046	22300	4.1945	6.9836	1.2932	6.9591	6.9265	15.1156
4.2983	3.4198	22400	4.1911	6.8481	1.3436	6.8345	6.801	14.9795
4.2524	3.4351	22500	4.1933	6.8429	1.3705	6.8301	6.7974	15.1334
4.3799	3.4504	22600	4.2044	6.7754	1.3201	6.7594	6.7238	15.2449
4.3351	3.4656	22700	4.1996	6.7478	1.3369	6.7354	6.6902	15.3856
4.3088	3.4809	22800	4.1830	6.6006	1.394	6.607	6.5574	14.9352
4.2826	3.4962	22900	4.1913	6.579	1.2865	6.5658	6.5318	15.1495
4.3173	3.5115	23000	4.1856	6.7014	1.3168	6.6822	6.6462	14.8069
4.2652	3.5267	23100	4.1815	6.8242	1.2753	6.8156	6.7638	15.0363
4.3326	3.5420	23200	4.1660	6.6178	1.31	6.6059	6.5704	14.8236
4.3819	3.5573	23300	4.1744	6.69	1.2417	6.6844	6.6238	15.0833
4.296	3.5725	23400	4.1765	6.7804	1.2417	6.7692	6.7302	15.0104
4.2839	3.5878	23500	4.1891	6.7177	1.2082	6.7068	6.6592	15.1189
4.3936	3.6031	23600	4.1627	6.6438	1.2462	6.6342	6.6022	14.8831
4.3749	3.6183	23700	4.1610	6.579	1.2026	6.5678	6.5286	14.7269
4.4921	3.6336	23800	4.1618	6.8785	1.3268	6.8869	6.8561	14.8542
4.2417	3.6489	23900	4.1715	6.6919	1.2149	6.6952	6.6428	15.1119
4.3279	3.6641	24000	4.1588	6.6078	1.2126	6.579	6.5398	14.5821
4.377	3.6794	24100	4.1621	6.5914	1.2832	6.5934	6.5618	14.5831
4.4372	3.6947	24200	4.1546	6.5958	1.3806	6.5734	6.5538	14.9976
4.2575	3.7099	24300	4.1522	6.669	1.4041	6.6574	6.6118	14.6829
4.4199	3.7252	24400	4.1481	6.4938	1.258	6.4797	6.4344	14.6234
4.3966	3.7405	24500	4.1480	6.759	1.3369	6.7278	6.7134	14.8781
4.4057	3.7557	24600	4.1590	6.7482	1.3912	6.708	6.6753	14.9526
4.3974	3.7710	24700	4.1685	6.8093	1.3257	6.7897	6.7613	15.0071
4.2648	3.7863	24800	4.1507	6.5587	1.2703	6.5346	6.4984	14.7471
4.3415	3.8015	24900	4.1499	6.9392	1.347	6.8959	6.888	15.2076
4.3974	3.8168	25000	4.1466	6.7832	1.2753	6.7694	6.734	14.957
4.2825	3.8321	25100	4.1301	6.5844	1.2496	6.5736	6.5191	14.7497
4.3653	3.8473	25200	4.1188	6.6626	1.2697	6.6756	6.6348	14.6167
4.3473	3.8626	25300	4.1329	6.601	1.2692	6.5773	6.5527	15.0312
4.2438	3.8779	25400	4.1512	6.6885	1.3403	6.6672	6.622	15.2207
4.4199	3.8931	25500	4.1339	6.5441	1.2865	6.535	6.5034	15.0265
4.3052	3.9084	25600	4.1405	6.7418	1.3196	6.7046	6.6598	15.3453
4.3431	3.9237	25700	4.1259	6.6682	1.3369	6.679	6.6447	14.8085
4.3	3.9389	25800	4.1163	6.6611	1.1566	6.6454	6.6028	14.4555
4.3897	3.9542	25900	4.1316	6.8313	1.3246	6.8121	6.7806	14.9375
4.3109	3.9695	26000	4.1415	6.8972	1.3677	6.8749	6.8356	14.9295
4.2875	3.9847	26100	4.1282	6.6918	1.3201	6.6868	6.6412	14.8344
4.4415	4.0	26200	4.1308	6.7908	1.3895	6.7688	6.7422	14.9446
4.2061	4.0153	26300	4.1286	6.7071	1.4041	6.7128	6.668	14.9362
4.1933	4.0305	26400	4.1110	6.5609	1.3537	6.5741	6.526	14.696
4.3083	4.0458	26500	4.1159	6.5279	1.338	6.5391	6.5033	14.9617
4.3101	4.0611	26600	4.1153	6.6397	1.3033	6.6272	6.5793	15.0558
4.2431	4.0763	26700	4.1102	6.5919	1.3537	6.6092	6.5825	14.8546
4.313	4.0916	26800	4.1142	6.7006	1.31	6.7078	6.6674	14.9751
4.221	4.1069	26900	4.1102	6.6186	1.2932	6.6282	6.5846	14.7403
4.1943	4.1221	27000	4.1122	6.6794	1.2865	6.6734	6.6218	14.8724
4.2413	4.1374	27100	4.1206	6.7126	1.3201	6.6918	6.661	14.5193
4.2038	4.1527	27200	4.1178	6.7126	1.3302	6.7126	6.6582	14.5499

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1