mt5-summarize-tr

pnr-svc/mt5-summarize-tr

mt5-summarize-tr is a machine learning model from pnr-svc. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as apache-2.0.

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

Downloads · 30 days

13% of all-time downloads

All-time downloads

Public

Parameters

300M

20.4 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors1.2 GB · 98%

At a glance

Library: transformers
License: apache-2.0
Model type: mt5
Access: Public
Created: Dec 29, 2023
Updated: Dec 29, 2023
SHA: d0c71ca8

Base models

google/mt5-small

Model card

From the Hugging Face model README

mt5-summarize-tr

This model is a fine-tuned version of google/mt5-small on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 2.0312
Rouge1: 0.3697
Rouge2: 0.2183
Rougel: 0.3248
Rougelsum: 0.3253

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: 0.0005
train_batch_size: 2
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 90
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
4.125	0.12	100	3.0694	0.2411	0.1271	0.2143	0.2131
3.5224	0.24	200	2.7433	0.3152	0.1669	0.2738	0.2743
3.2845	0.35	300	2.5655	0.3254	0.1738	0.2870	0.2863
3.1946	0.47	400	2.5012	0.3080	0.1659	0.2795	0.2798
3.2075	0.59	500	2.4685	0.3488	0.2036	0.3117	0.3120
3.0507	0.71	600	2.4249	0.3337	0.1947	0.2900	0.2898
3.0301	0.82	700	2.4027	0.3396	0.1949	0.2973	0.2987
3.032	0.94	800	2.3871	0.3165	0.1789	0.2777	0.2769
2.8716	1.06	900	2.3811	0.3446	0.2113	0.3104	0.3093
2.8563	1.18	1000	2.3448	0.3603	0.2248	0.3251	0.3247
2.8408	1.3	1100	2.3438	0.3387	0.1945	0.2960	0.2963
2.7755	1.41	1200	2.3275	0.3473	0.2038	0.2910	0.2920
2.8019	1.53	1300	2.2786	0.3438	0.2021	0.3029	0.3022
2.7457	1.65	1400	2.2380	0.3308	0.1887	0.2985	0.2985
2.7216	1.77	1500	2.2517	0.3515	0.2112	0.3079	0.3052
2.71	1.88	1600	2.2712	0.3397	0.1978	0.2955	0.2932
2.6489	2.0	1700	2.2678	0.3441	0.2059	0.3047	0.3049
2.6232	2.12	1800	2.2526	0.3242	0.1901	0.2874	0.2864
2.5419	2.24	1900	2.2526	0.2985	0.1636	0.2611	0.2610
2.6232	2.36	2000	2.2108	0.3454	0.2053	0.2995	0.2999
2.5973	2.47	2100	2.2166	0.3363	0.1986	0.2975	0.2968
2.5738	2.59	2200	2.1973	0.3263	0.1802	0.2803	0.2812
2.5279	2.71	2300	2.1978	0.3186	0.1778	0.2829	0.2857
2.5613	2.83	2400	2.1921	0.3539	0.2126	0.3125	0.3166
2.5856	2.94	2500	2.1830	0.3719	0.2293	0.3295	0.3307
2.5181	3.06	2600	2.1531	0.3581	0.2177	0.3245	0.3239
2.4885	3.18	2700	2.1702	0.3409	0.1980	0.3034	0.3051
2.4172	3.3	2800	2.1562	0.3387	0.1872	0.2927	0.2925
2.5102	3.41	2900	2.1398	0.3387	0.2004	0.2947	0.2944
2.5086	3.53	3000	2.1253	0.3453	0.1935	0.3031	0.3029
2.4235	3.65	3100	2.1530	0.3495	0.1990	0.2932	0.2936
2.364	3.77	3200	2.1215	0.3564	0.2057	0.3049	0.3072
2.4752	3.89	3300	2.1223	0.3250	0.1818	0.2764	0.2772
2.3531	4.0	3400	2.1100	0.3453	0.1945	0.2931	0.2943
2.3746	4.12	3500	2.1162	0.3494	0.1976	0.2991	0.2998
2.3519	4.24	3600	2.1069	0.3496	0.2118	0.3061	0.3052
2.3154	4.36	3700	2.1217	0.3544	0.2170	0.3257	0.3269
2.3163	4.47	3800	2.1092	0.3445	0.2007	0.3044	0.3045
2.3641	4.59	3900	2.1098	0.3484	0.2072	0.3127	0.3126
2.298	4.71	4000	2.1019	0.3571	0.2190	0.3211	0.3224
2.3414	4.83	4100	2.0955	0.3534	0.2002	0.2949	0.2957
2.3812	4.95	4200	2.1027	0.3531	0.2058	0.3009	0.3011
2.2614	5.06	4300	2.1107	0.3379	0.1902	0.2891	0.2886
2.2528	5.18	4400	2.0953	0.3789	0.2229	0.3263	0.3243
2.3287	5.3	4500	2.0977	0.3671	0.2120	0.3200	0.3180
2.2855	5.42	4600	2.1044	0.3567	0.2085	0.3030	0.3057
2.2969	5.53	4700	2.0925	0.3459	0.2103	0.3126	0.3119
2.2438	5.65	4800	2.0879	0.3577	0.2020	0.2999	0.2996
2.2088	5.77	4900	2.0862	0.3662	0.2212	0.3222	0.3259
2.2319	5.89	5000	2.0839	0.3530	0.2040	0.3108	0.3130
2.2879	6.01	5100	2.0761	0.3693	0.2236	0.3285	0.3288
2.2508	6.12	5200	2.0810	0.3924	0.2424	0.3440	0.3448
2.2634	6.24	5300	2.0760	0.3722	0.2170	0.3162	0.3165
2.1484	6.36	5400	2.0725	0.3624	0.2110	0.3097	0.3115
2.1863	6.48	5500	2.0663	0.3820	0.2332	0.3338	0.3348
2.1935	6.59	5600	2.0762	0.3906	0.2299	0.3421	0.3425
2.224	6.71	5700	2.0536	0.3788	0.2362	0.3337	0.3336
2.1938	6.83	5800	2.0663	0.3551	0.2017	0.3047	0.3044
2.2589	6.95	5900	2.0602	0.3782	0.2298	0.3365	0.3360
2.1185	7.07	6000	2.0632	0.3617	0.2170	0.3237	0.3239
2.1924	7.18	6100	2.0508	0.3771	0.2211	0.3284	0.3295
2.2012	7.3	6200	2.0602	0.3714	0.2184	0.3315	0.3315
2.1617	7.42	6300	2.0606	0.3311	0.1800	0.2883	0.2862
2.1557	7.54	6400	2.0444	0.3827	0.2318	0.3386	0.3383
2.1965	7.65	6500	2.0391	0.3777	0.2334	0.3401	0.3414
2.1599	7.77	6600	2.0393	0.3648	0.2121	0.3154	0.3152
2.1505	7.89	6700	2.0483	0.3905	0.2411	0.3485	0.3496
2.191	8.01	6800	2.0412	0.3815	0.2373	0.3327	0.3337
2.1513	8.12	6900	2.0461	0.3745	0.2252	0.3363	0.3374
2.1232	8.24	7000	2.0474	0.3672	0.2183	0.3264	0.3270
2.1249	8.36	7100	2.0422	0.3846	0.2270	0.3379	0.3388
2.1743	8.48	7200	2.0358	0.3707	0.2086	0.3192	0.3184
2.1684	8.6	7300	2.0391	0.3532	0.2061	0.3048	0.3035
2.1257	8.71	7400	2.0388	0.3773	0.2198	0.3218	0.3219
2.1186	8.83	7500	2.0356	0.3655	0.2135	0.3091	0.3077
2.054	8.95	7600	2.0308	0.3739	0.2276	0.3208	0.3208
2.1534	9.07	7700	2.0253	0.3649	0.2150	0.3154	0.3157
2.0906	9.18	7800	2.0327	0.3752	0.2233	0.3173	0.3174
2.0703	9.3	7900	2.0314	0.3610	0.2100	0.3153	0.3150
2.1088	9.42	8000	2.0289	0.3636	0.2106	0.3171	0.3174
2.0719	9.54	8100	2.0298	0.3682	0.2145	0.3225	0.3223
2.0873	9.66	8200	2.0294	0.3729	0.2185	0.3225	0.3228
2.0953	9.77	8300	2.0316	0.3729	0.2185	0.3225	0.3228
2.0825	9.89	8400	2.0312	0.3697	0.2183	0.3248	0.3253

Framework versions

Transformers 4.35.2
Pytorch 2.1.0+cu121
Datasets 2.16.0
Tokenizers 0.15.0

Library: transformers
Type: mt5
License: apache-2.0
Created: Dec 29, 2023
Updated: Dec 29, 2023