T5La-Large-WeightedLoss

hrezaei/T5La-Large-WeightedLoss

T5La-Large-WeightedLoss is a text generation model from hrezaei. Use it when you need the model to write or continue text. It is set up for transformers.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

20% of all-time downloads

All-time downloads

Public

Parameters

816M

42.4 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors3.3 GB · 100%

At a glance

Task: Text Generation
Library: transformers
Model type: t5la
Access: Public
Created: Sep 24, 2025
Updated: Jul 28, 2026
SHA: 1a4ad7e4

Try a prompt

Datasets

HuggingFaceFW/fineweb

Model card

From the Hugging Face model README

T5La-Large-WeightedLoss

This model is a fine-tuned version of on the HuggingFaceFW/fineweb sample-350BT dataset. It achieves the following results on the evaluation set:

Perplexity: 59.9708
Loss: 4.0939
Accuracy: 0.0396
Lookahead Perplexity: 744.4827
Lookahead Loss: 6.6127

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

Wandb Report

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 4
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
training_steps: 524288

Training results

Training Loss	Epoch	Step	Accuracy	Lookahead Loss	Lookahead Perplexity	Validation Loss	Perplexity
4.6301	0.0095	5000	0.0312	7.2366	1389.3693	4.5864	98.1387
4.4871	0.0191	10000	0.0318	7.1133	1228.2065	4.4621	86.6700
4.4417	0.0286	15000	0.0340	7.0459	1148.1578	4.3906	80.6868
4.3978	0.0381	20000	0.0352	6.9924	1088.3650	4.3427	76.9114
4.3419	0.0477	25000	0.0349	6.9566	1050.0462	4.3110	74.5165
4.316	0.0572	30000	0.0361	6.9482	1041.2903	4.3202	75.2024
4.3488	0.0668	35000	0.0346	6.9130	1005.3051	4.2879	72.8156
4.3213	0.0763	40000	0.0347	6.8934	985.7009	4.2729	71.7300
4.2945	0.0858	45000	0.0337	6.8687	961.7435	4.2490	70.0335
4.3156	0.0954	50000	0.0363	6.8545	948.1267	4.2447	69.7315
4.3335	0.1049	55000	0.0359	6.8432	937.4398	4.2384	69.2975
4.2959	0.1144	60000	0.0374	6.8479	941.9400	4.2551	70.4671
4.2732	0.1240	65000	0.0372	6.8335	928.3968	4.2418	69.5327
4.2924	0.1335	70000	0.0384	6.8230	918.7655	4.2447	69.7361
4.2923	0.1431	75000	0.0364	6.8392	933.7382	4.2519	70.2380
4.2252	0.1526	80000	0.0371	6.8151	911.5434	4.2382	69.2813
4.3282	0.1621	85000	0.0372	6.8016	899.2575	4.2262	68.4571
4.2749	0.1717	90000	0.0379	6.7829	882.5851	4.2115	67.4584
4.2947	0.1812	95000	0.0375	6.7889	887.9544	4.2235	68.2707
4.2664	0.1907	100000	0.0377	6.7943	892.7447	4.2326	68.8933
4.295	0.2003	105000	0.0370	6.7986	896.5748	4.2416	69.5213
4.2937	0.2098	110000	0.0380	6.7741	874.8497	4.2202	68.0471
4.2807	0.2193	115000	0.0391	6.7792	879.3644	4.2272	68.5222
4.3136	0.2289	120000	0.0384	6.7866	885.9135	4.2357	69.1117
4.3081	0.2384	125000	0.0381	6.7903	889.1959	4.2430	69.6157
4.3328	0.2480	130000	0.0396	6.7835	883.1839	4.2406	69.4502
4.3221	0.2575	135000	0.0400	6.7754	876.0138	4.2366	69.1692
4.3163	0.2670	140000	0.0389	6.7802	880.2415	4.2404	69.4384
4.2954	0.2766	145000	0.0381	6.7683	869.8558	4.2300	68.7163
4.3153	0.2861	150000	0.0398	6.7570	860.0185	4.2200	68.0362
4.2762	0.2956	155000	0.0391	6.7518	855.6387	4.2128	67.5476
4.2877	0.3052	160000	0.0382	6.7626	864.9080	4.2276	68.5495
4.3023	0.3147	165000	0.0395	6.7556	858.8960	4.2232	68.2494
4.337	0.3242	170000	0.0396	6.7766	877.0693	4.2479	69.9564
4.2891	0.3338	175000	0.0383	6.7645	866.5396	4.2347	69.0438
4.3681	0.3433	180000	0.0395	6.7677	869.3163	4.2391	69.3437
4.3192	0.3529	185000	0.0390	6.7627	864.9563	4.2324	68.8804
4.2911	0.3624	190000	0.0382	6.7589	861.7325	4.2307	68.7676
4.332	0.3719	195000	0.0394	6.7499	853.9386	4.2234	68.2637
4.311	0.3815	200000	0.0392	6.7693	870.7021	4.2438	69.6714
4.2935	0.3910	205000	0.0388	6.7643	866.3254	4.2368	69.1861
4.3381	0.4005	210000	0.0398	6.7418	847.1075	4.2172	67.8446
4.2862	0.4101	215000	0.0387	6.7396	845.2019	4.2118	67.4806
4.2843	0.4196	220000	0.0397	6.7369	842.9649	4.2114	67.4486
4.3212	0.4292	225000	0.0399	6.7340	840.4940	4.2066	67.1268
4.2976	0.4387	230000	0.0390	6.7264	834.1246	4.2024	66.8453
4.322	0.4482	235000	0.0370	6.7456	850.2819	4.2172	67.8427
4.2604	0.4578	240000	0.0389	6.7255	833.4224	4.2010	66.7518
4.2827	0.4673	245000	0.0374	6.7311	838.0625	4.2070	67.1551
4.2629	0.4768	250000	0.0398	6.7232	831.4937	4.2016	66.7902
4.2719	0.4864	255000	0.0387	6.7234	831.6196	4.2018	66.8038
4.2518	0.4959	260000	0.0392	6.7199	828.7639	4.1976	66.5268
4.2439	0.5054	265000	0.0381	6.7225	830.8789	4.1985	66.5882
4.2873	0.5150	270000	0.0383	6.7290	836.3176	4.2064	67.1118
4.2885	0.5245	275000	0.0394	6.7120	822.2442	4.1908	66.0741
4.2603	0.5341	280000	0.0392	6.7111	821.5043	4.1885	65.9229
4.2623	0.5436	285000	0.0399	6.7100	820.5507	4.1877	65.8688
4.2669	0.5531	290000	0.0385	6.7113	821.6451	4.1869	65.8204
4.3208	0.5627	295000	0.0400	6.6984	811.1413	4.1780	65.2359
4.249	0.5722	300000	0.0397	6.6998	812.2105	4.1786	65.2716
4.2383	0.5817	305000	0.0390	6.6950	808.3597	4.1740	64.9756
4.2473	0.5913	310000	0.0387	6.7067	817.9059	4.1827	65.5417
4.2612	0.6008	315000	0.0391	6.6928	806.6095	4.1715	64.8147
4.1997	0.6104	320000	0.0392	6.6874	802.2146	4.1658	64.4451
4.2437	0.6199	325000	0.0400	6.6893	803.7320	4.1681	64.5955
4.2243	0.6294	330000	0.0399	6.6818	797.7855	4.1615	64.1649
4.2524	0.6390	335000	0.0399	6.6764	793.4797	4.1580	63.9440
4.2496	0.6485	340000	0.0396	6.6827	798.4811	4.1633	64.2805
4.2444	0.6580	345000	0.0394	6.6818	797.7289	4.1606	64.1105
4.2127	0.6676	350000	0.0402	6.6686	787.2920	4.1505	63.4653
4.2095	0.6771	355000	0.0390	6.6778	794.5674	4.1568	63.8654
4.2228	0.6866	360000	0.0392	6.6670	786.0658	4.1472	63.2587
4.2044	0.6962	365000	0.0397	6.6652	784.5846	4.1455	63.1497
4.2489	0.7057	370000	0.0395	6.6617	781.8825	4.1434	63.0145
4.2101	0.7153	375000	0.0396	6.6568	778.0572	4.1384	62.7031
4.2117	0.7248	380000	0.0395	6.6607	781.1064	4.1406	62.8427
4.2015	0.7343	385000	0.0397	6.6533	775.3367	4.1356	62.5253
4.2242	0.7439	390000	0.0395	6.6529	775.0460	4.1344	62.4528
4.1964	0.7534	395000	0.0397	6.6506	773.2814	4.1314	62.2637
4.1939	0.7629	400000	0.0401	6.6492	772.1992	4.1299	62.1728
4.229	0.7725	405000	0.0397	6.6447	768.7131	4.1276	62.0275
4.1805	0.7820	410000	0.0399	6.6408	765.7219	4.1229	61.7394
4.2036	0.7915	415000	0.0394	6.6451	769.0443	4.1261	61.9348
4.169	0.8011	420000	0.0404	6.6364	762.3091	4.1192	61.5076
4.1815	0.8106	425000	0.0395	6.6399	764.9846	4.1210	61.6187
4.189	0.8202	430000	0.0397	6.6364	762.3756	4.1181	61.4406
4.1774	0.8297	435000	0.0391	6.6365	762.4051	4.1172	61.3861
4.1925	0.8392	440000	0.0398	6.6308	758.1226	4.1129	61.1245
4.1868	0.8488	445000	0.0391	6.6303	757.6731	4.1124	61.0916
4.1189	0.8583	450000	61.2059	4.1142	0.0394	759.8155	6.6331
4.2156	0.8678	455000	61.0326	4.1114	0.0394	757.6308	6.6302
4.1966	0.8774	460000	60.7623	4.1070	0.0398	754.0364	6.6254
4.158	0.8869	465000	60.7495	4.1068	0.0398	754.6634	6.6263
4.1639	0.8965	470000	60.5520	4.1035	0.0392	751.4165	6.6220
4.1936	0.9060	475000	60.4407	4.1017	0.0396	750.2459	6.6204
4.154	0.9155	480000	60.3311	4.0998	0.0393	748.5255	6.6181
4.1473	0.9251	485000	60.3326	4.0999	0.0397	748.9214	6.6186
4.1857	0.9346	490000	60.1977	4.0976	0.0397	746.6764	6.6156
4.1724	0.9441	495000	60.1996	4.0977	0.0397	747.4420	6.6167
4.166	0.9537	500000	60.1117	4.0962	0.0394	746.2841	6.6151
4.1581	1.0095	505000	60.1274	4.0965	0.0395	746.5216	6.6154
4.1443	1.0191	510000	60.1217	4.0964	0.0395	746.7120	6.6157
4.1641	1.0286	515000	60.0107	4.0945	0.0397	744.8850	6.6132
4.1673	1.0381	520000	59.9055	4.0928	0.0397	743.3574	6.6112

Framework versions

Transformers 4.57.0.dev0
Pytorch 2.8.0+cu128
Datasets 4.0.0
Tokenizers 0.22.1

Task: Text Generation
Library: transformers
Type: t5la
Created: Sep 24, 2025
Updated: Jul 28, 2026