Характеристики

Лицензия: Своя лицензия
Архитектура: segformer · dense
Базовая модель: mit-b5
Опубликована: 25 мар 2025

Ссылки

Hugging Face

segformer-b5-finetuned-IDD-L2_v2

izzako

Сегментация изображенийИзображение → mask3 профессии

Дообученная SegFormer B5 модель для пиксельной сегментации объектов и элементов дорожных сцен на изображениях IDD.

1 668скачиваний за 30 дней0лайковвеса обновлены8 апр 2025

Бенчмарки

Как модель справилась с задачами пользователей ailist и что о ней публикуют открытые лидерборды

Публичные оценки

Из открытых лидербордов. Сравнивайте результаты одного теста при одинаковых условиях — отсутствие оценки не значит, что модель хуже.

В подключённых источниках результатов для этой модели пока нет.

Что умеет модель

Описание составлено по карточке модели и метаданным Hugging Face

izzako/segformer-b5-finetuned-IDD-L2_v2 — модель семантической сегментации изображений, дообученная на наборе IDD 20K Semantic Segmentation Dataset на базе nvidia/mit-b5. Она размечает пиксели уличных сцен, включая дороги, парковки, тротуары, железнодорожные пути, людей, транспорт, здания, растительность и небо. Модель подходит для анализа дорожных сцен и восприятия городской инфраструктуры по изображениям.

В README приведены результаты на оценочной выборке: mean IoU 0,718, mean accuracy 0,822 и общая точность 0,908. Качество различается по классам: например, IoU дороги составляет 0,939, а ограждений — 0,457. Файлы модели представлены в формате safetensors. Лицензия указана как other, поэтому условия использования следует сверить с карточкой модели.

Сценарии применения

сегментация дорог и тротуаров на уличных снимках
распознавание автомобилей, мотоциклов, авторикш и грузовиков на дороге
выделение пешеходов и велосипедистов в кадре
разметка зданий, растительности и неба на изображениях городской среды

Ограничения

Качество зависит от класса: в оценке IoU ограждений составляет 0,457, а железнодорожных путей — 0,503. Результаты приведены для оценочной выборки IDD 20K и не гарантируют такое же качество на других данных.

Для каких профессий подходит3

Разделы каталога профессий, в задачах которых используется эта модель

Все профессии

II.10Разработка и данныеРазработка и эксплуатация AI/ML-системAI помогает создавать, оценивать, развёртывать и контролировать AI/ML-системы на всём жизненном цикле.30 задач49 000моделей II.8Разработка и данныеИнженерия данных и аналитикаAI помогает готовить надёжные данные и превращать их в понятную аналитику.22 задачи23 488моделей IV.8Отраслевые задачиГосударственное управление, правосудие и чрезвычайные ситуацииAI помогает улучшать государственные услуги, управление ресурсами, правосудие и реагирование на чрезвычайные ситуации.47 задач2 517моделей

Карточка модели

README от автора на Hugging Face, как он его опубликовал

Открыть на Hugging Face

segformer-b5-finetuned-IDD-L2_v2

This model is a fine-tuned version of nvidia/mit-b5 on the IDD 20K Semantic Segmentation Dataset dataset. It achieves the following results on the evaluation set:

Loss: 0.5563
Mean Iou: 0.7180
Mean Accuracy: 0.8224
Overall Accuracy: 0.9083
Accuracy Road: 0.9716
Accuracy Parking: 0.7949
Accuracy Sidewalk: 0.8240
Accuracy Rail track: 0.6408
Accuracy Person: 0.8057
Accuracy Rider: 0.8434
Accuracy Motorcycle: 0.8762
Accuracy Autorickshaw: 0.9451
Accuracy Truck: 0.9122
Accuracy Curb: 0.8112
Accuracy Fence: 0.5699
Accuracy Billboard: 0.7605
Accuracy Pole: 0.6010
Accuracy Building: 0.8678
Accuracy Vegetation: 0.9495
Accuracy Sky: 0.9841
Iou Road: 0.9391
Iou Parking: 0.6620
Iou Sidewalk: 0.6707
Iou Rail track: 0.5025
Iou Person: 0.6726
Iou Rider: 0.7228
Iou Motorcycle: 0.7637
Iou Autorickshaw: 0.8882
Iou Truck: 0.8506
Iou Curb: 0.6721
Iou Fence: 0.4571
Iou Billboard: 0.6238
Iou Pole: 0.4831
Iou Building: 0.7293
Iou Vegetation: 0.8792
Iou Sky: 0.9707

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: 0.0006
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Road	Accuracy Parking	Accuracy Sidewalk	Accuracy Rail track	Accuracy Person	Accuracy Rider	Accuracy Motorcycle	Accuracy Autorickshaw	Accuracy Truck	Accuracy Curb	Accuracy Fence	Accuracy Billboard	Accuracy Pole	Accuracy Building	Accuracy Vegetation	Accuracy Sky	Iou Road	Iou Parking	Iou Sidewalk	Iou Rail track	Iou Person	Iou Rider	Iou Motorcycle	Iou Autorickshaw	Iou Truck	Iou Curb	Iou Fence	Iou Billboard	Iou Pole	Iou Building	Iou Vegetation	Iou Sky
0.3096	1.0	202	0.3310	0.6476	0.7663	0.8841	0.9538	0.7998	0.7838	0.5755	0.6912	0.7355	0.8445	0.9518	0.8603	0.7173	0.3997	0.6938	0.4456	0.8868	0.9531	0.9690	0.9256	0.6317	0.5876	0.4378	0.5705	0.5911	0.6722	0.8070	0.7758	0.6255	0.3351	0.5507	0.3663	0.6693	0.8551	0.9600
0.2786	2.0	404	0.3369	0.6560	0.7917	0.8774	0.8968	0.8571	0.8009	0.5009	0.7221	0.8105	0.8267	0.9102	0.9216	0.8204	0.6046	0.6927	0.5244	0.8650	0.9354	0.9773	0.8834	0.5556	0.5941	0.4215	0.6054	0.6340	0.6926	0.8403	0.7757	0.6229	0.3671	0.5681	0.4130	0.6967	0.8610	0.9641
0.2541	3.0	606	0.3013	0.6796	0.7930	0.8958	0.9724	0.7579	0.8132	0.5424	0.7483	0.8020	0.8235	0.9296	0.9107	0.8067	0.5309	0.7415	0.5359	0.8420	0.9502	0.9806	0.9336	0.6384	0.6194	0.4623	0.6185	0.6505	0.7102	0.8534	0.8099	0.6446	0.3983	0.5839	0.4236	0.6975	0.8641	0.9660
0.2304	4.0	808	0.3055	0.6860	0.8016	0.8947	0.9493	0.8219	0.7667	0.5809	0.7948	0.7780	0.8393	0.9303	0.9033	0.7869	0.5968	0.7070	0.5882	0.8711	0.9259	0.9851	0.9250	0.6310	0.6435	0.4815	0.6264	0.6589	0.7188	0.8609	0.8095	0.6519	0.4044	0.5840	0.4376	0.7066	0.8697	0.9671
0.214	5.0	1010	0.3138	0.6845	0.7921	0.8967	0.9526	0.8481	0.8431	0.4857	0.7539	0.7570	0.8772	0.9463	0.8889	0.7298	0.4582	0.7475	0.5890	0.8661	0.9449	0.9850	0.9289	0.6438	0.6386	0.4351	0.6379	0.6594	0.7182	0.8578	0.8239	0.6402	0.3870	0.5965	0.4537	0.6955	0.8695	0.9666
0.2029	6.0	1212	0.3123	0.6914	0.8010	0.8988	0.9697	0.7612	0.8111	0.6453	0.7988	0.8621	0.8112	0.9377	0.9028	0.7567	0.5211	0.7047	0.5390	0.8494	0.9638	0.9822	0.9339	0.6378	0.6746	0.4754	0.6271	0.6660	0.7143	0.8682	0.8305	0.6344	0.4264	0.5947	0.4335	0.7168	0.8614	0.9679
0.1837	7.0	1414	0.3201	0.6963	0.8009	0.9008	0.9614	0.8304	0.7910	0.5696	0.7538	0.8212	0.8642	0.9357	0.9092	0.7900	0.5475	0.6914	0.5137	0.9004	0.9509	0.9841	0.9337	0.6534	0.6600	0.4621	0.6489	0.6881	0.7289	0.8715	0.8325	0.6717	0.4265	0.5842	0.4306	0.7074	0.8725	0.9683
0.1904	8.0	1616	0.3075	0.6946	0.8092	0.8997	0.9693	0.7558	0.7707	0.6996	0.7802	0.8379	0.8572	0.9428	0.8965	0.7714	0.5436	0.7467	0.6316	0.8216	0.9398	0.9827	0.9350	0.6371	0.6389	0.4958	0.6434	0.6665	0.7272	0.8646	0.8259	0.6533	0.4153	0.5983	0.4638	0.7058	0.8745	0.9687
0.166	9.0	1818	0.3127	0.7018	0.8110	0.9030	0.9725	0.7634	0.8258	0.6332	0.7766	0.8286	0.8336	0.9274	0.9101	0.8199	0.5639	0.7524	0.5788	0.8569	0.9506	0.9828	0.9342	0.6427	0.6429	0.4986	0.6563	0.6966	0.7359	0.8724	0.8262	0.6580	0.4284	0.6059	0.4654	0.7232	0.8738	0.9687
0.157	10.0	2020	0.3267	0.7024	0.8168	0.9020	0.9604	0.8061	0.7791	0.6743	0.8065	0.8319	0.8585	0.9401	0.9024	0.7924	0.5971	0.7363	0.5982	0.8637	0.9412	0.9811	0.9341	0.6525	0.6635	0.5021	0.6506	0.6902	0.7353	0.8638	0.8235	0.6649	0.4250	0.6045	0.4686	0.7166	0.8754	0.9686
0.1543	11.0	2222	0.3260	0.7009	0.8131	0.9029	0.9753	0.7483	0.8429	0.6313	0.7917	0.8380	0.8368	0.9411	0.8959	0.7904	0.5871	0.7533	0.5913	0.8523	0.9509	0.9824	0.9341	0.6377	0.6377	0.4966	0.6583	0.6951	0.7325	0.8603	0.8235	0.6620	0.4303	0.6088	0.4709	0.7213	0.8763	0.9689
0.1474	12.0	2424	0.3394	0.7054	0.8179	0.9028	0.9585	0.8439	0.8303	0.6732	0.7800	0.8231	0.8755	0.9453	0.8985	0.7896	0.5814	0.7545	0.5047	0.8994	0.9452	0.9842	0.9351	0.6593	0.6707	0.5076	0.6632	0.6995	0.7427	0.8737	0.8403	0.6595	0.4371	0.6068	0.4317	0.7151	0.8752	0.9693
0.1365	13.0	2626	0.3347	0.7121	0.8220	0.9063	0.9706	0.7858	0.8210	0.6734	0.8003	0.8490	0.8637	0.9461	0.9108	0.7885	0.6000	0.7537	0.5916	0.8648	0.9510	0.9815	0.9387	0.6604	0.6668	0.5029	0.6702	0.7067	0.7480	0.8793	0.8403	0.6713	0.4449	0.6162	0.4739	0.7280	0.8774	0.9689
0.1301	14.0	2828	0.3412	0.7148	0.8159	0.9073	0.9734	0.7725	0.8092	0.6343	0.7902	0.8236	0.8692	0.9424	0.9058	0.8016	0.5709	0.7403	0.6210	0.8633	0.9524	0.9838	0.9384	0.6549	0.6718	0.5043	0.6695	0.7121	0.7544	0.8832	0.8444	0.6744	0.4519	0.6165	0.4869	0.7252	0.8787	0.9696
0.1216	15.0	3030	0.3623	0.7137	0.8187	0.9071	0.9686	0.8045	0.8266	0.6700	0.8050	0.8177	0.8601	0.9466	0.9172	0.8057	0.5325	0.7487	0.5887	0.8770	0.9476	0.9830	0.9390	0.6654	0.6707	0.5055	0.6697	0.7121	0.7563	0.8813	0.8468	0.6702	0.4311	0.6210	0.4756	0.7260	0.8795	0.9699
0.1198	16.0	3232	0.3660	0.7154	0.8230	0.9073	0.9703	0.8029	0.8357	0.6354	0.8038	0.8289	0.8623	0.9484	0.9166	0.8085	0.6043	0.7602	0.5904	0.8719	0.9444	0.9839	0.9385	0.6613	0.6717	0.4995	0.6695	0.7130	0.7542	0.8825	0.8511	0.6757	0.4554	0.6205	0.4779	0.7279	0.8782	0.9699
0.1264	17.0	3434	0.3688	0.7093	0.8161	0.9037	0.9631	0.7818	0.7960	0.7013	0.8166	0.8203	0.8650	0.9403	0.9087	0.7833	0.5580	0.7223	0.5917	0.8796	0.9515	0.9789	0.9339	0.6401	0.6720	0.4989	0.6648	0.7075	0.7501	0.8823	0.8481	0.6613	0.4408	0.6131	0.4704	0.7195	0.8767	0.9687
0.137	18.0	3636	0.3583	0.7079	0.8131	0.9048	0.9699	0.7900	0.8168	0.6048	0.8107	0.8355	0.8609	0.9340	0.9144	0.7944	0.4969	0.7738	0.6219	0.8578	0.9431	0.9840	0.9366	0.6536	0.6658	0.5010	0.6672	0.7036	0.7478	0.8792	0.8352	0.6600	0.4237	0.6080	0.4789	0.7174	0.8782	0.9695
0.1243	19.0	3838	0.3671	0.7050	0.8140	0.9045	0.9728	0.7609	0.8216	0.6357	0.7975	0.8456	0.8801	0.9328	0.8930	0.7846	0.5696	0.7238	0.5972	0.8748	0.9539	0.9808	0.9370	0.6470	0.6367	0.4900	0.6577	0.7019	0.7449	0.8665	0.8285	0.6600	0.4545	0.6079	0.4741	0.7283	0.8765	0.9693
0.1139	20.0	4040	0.3773	0.7114	0.8169	0.9063	0.9745	0.7752	0.8003	0.6377	0.7977	0.8502	0.8648	0.9407	0.9017	0.8282	0.5676	0.7387	0.6023	0.8613	0.9465	0.9837	0.9388	0.6584	0.6593	0.5051	0.6649	0.7108	0.7508	0.8775	0.8388	0.6628	0.4501	0.6150	0.4735	0.7273	0.8790	0.9698
0.107	21.0	4242	0.3894	0.7156	0.8208	0.9075	0.9698	0.8161	0.8183	0.6471	0.7960	0.8217	0.8852	0.9457	0.9072	0.8157	0.5433	0.7691	0.6085	0.8623	0.9432	0.9830	0.9399	0.6685	0.6703	0.5096	0.6752	0.7144	0.7542	0.8816	0.8473	0.6670	0.4446	0.6210	0.4815	0.7263	0.8791	0.9701
0.1031	22.0	4444	0.4017	0.7151	0.8233	0.9073	0.9691	0.8062	0.8372	0.6409	0.8117	0.8351	0.8776	0.9459	0.9121	0.8086	0.5799	0.7538	0.6018	0.8625	0.9452	0.9853	0.9387	0.6634	0.6669	0.5031	0.6714	0.7145	0.7572	0.8839	0.8469	0.6665	0.4540	0.6171	0.4809	0.7264	0.8800	0.9702
0.0999	23.0	4646	0.4116	0.7157	0.8202	0.9077	0.9743	0.7844	0.8315	0.6443	0.7960	0.8279	0.8840	0.9517	0.8990	0.8163	0.5712	0.7390	0.6063	0.8684	0.9476	0.9821	0.9395	0.6627	0.6683	0.5113	0.6759	0.7177	0.7575	0.8790	0.8417	0.6645	0.4521	0.6198	0.4838	0.7280	0.8797	0.9703
0.0963	24.0	4848	0.4264	0.7153	0.8177	0.9074	0.9754	0.7781	0.8163	0.6371	0.8127	0.8306	0.8785	0.9477	0.9128	0.7785	0.5563	0.7605	0.5915	0.8720	0.9529	0.9827	0.9389	0.6555	0.6780	0.5042	0.6689	0.7179	0.7590	0.8842	0.8467	0.6665	0.4506	0.6248	0.4772	0.7244	0.8784	0.9703
0.096	25.0	5050	0.4291	0.7157	0.8208	0.9078	0.9727	0.7987	0.8335	0.6217	0.8103	0.8348	0.8773	0.9458	0.9083	0.8045	0.5717	0.7563	0.5977	0.8668	0.9508	0.9825	0.9392	0.6624	0.6659	0.5007	0.6680	0.7168	0.7586	0.8855	0.8497	0.6674	0.4571	0.6224	0.4799	0.7274	0.8794	0.9702
0.0926	26.0	5252	0.4360	0.7169	0.8230	0.9081	0.9721	0.7966	0.8304	0.6249	0.8104	0.8352	0.8746	0.9530	0.9080	0.8115	0.5851	0.7736	0.5962	0.8616	0.9495	0.9848	0.9396	0.6643	0.6679	0.5000	0.6711	0.7208	0.7626	0.8836	0.8481	0.6672	0.4583	0.6277	0.4808	0.7288	0.8794	0.9704
0.0903	27.0	5454	0.4453	0.7166	0.8226	0.9077	0.9726	0.7853	0.8258	0.6282	0.8004	0.8374	0.8824	0.9457	0.9191	0.8191	0.5907	0.7600	0.5891	0.8771	0.9470	0.9820	0.9392	0.6602	0.6705	0.5029	0.6762	0.7191	0.7619	0.8867	0.8472	0.6657	0.4559	0.6253	0.4776	0.7277	0.8797	0.9702
0.093	28.0	5656	0.4438	0.7152	0.8195	0.9072	0.9696	0.7980	0.8227	0.6421	0.7945	0.8561	0.8705	0.9487	0.9011	0.8025	0.5575	0.7460	0.5975	0.8694	0.9504	0.9847	0.9372	0.6581	0.6650	0.5061	0.6746	0.7209	0.7549	0.8822	0.8461	0.6640	0.4536	0.6201	0.4814	0.7290	0.8786	0.9706
0.0883	29.0	5858	0.4514	0.7173	0.8231	0.9079	0.9714	0.7987	0.8199	0.6475	0.8239	0.8308	0.8722	0.9480	0.9129	0.8073	0.5782	0.7659	0.5917	0.8685	0.9516	0.9807	0.9395	0.6652	0.6729	0.5059	0.6700	0.7222	0.7621	0.8844	0.8489	0.6674	0.4604	0.6219	0.4785	0.7282	0.8790	0.9700
0.0857	30.0	6060	0.4478	0.7175	0.8225	0.9081	0.9691	0.8046	0.8208	0.6461	0.8021	0.8401	0.8844	0.9456	0.9117	0.8035	0.5546	0.7730	0.6106	0.8596	0.9510	0.9836	0.9387	0.6648	0.6688	0.5043	0.6730	0.7209	0.7629	0.8874	0.8509	0.6693	0.4518	0.6229	0.4867	0.7293	0.8788	0.9703
0.0852	31.0	6262	0.4704	0.7165	0.8207	0.9078	0.9735	0.7878	0.8214	0.6576	0.8043	0.8423	0.8666	0.9484	0.9023	0.8110	0.5632	0.7524	0.5969	0.8748	0.9448	0.9844	0.9387	0.6608	0.6739	0.5054	0.6715	0.7218	0.7640	0.8829	0.8412	0.6700	0.4513	0.6232	0.4796	0.7300	0.8799	0.9706
0.0813	32.0	6464	0.4642	0.7173	0.8239	0.9080	0.9698	0.7989	0.8283	0.6422	0.8068	0.8390	0.8813	0.9479	0.9081	0.8147	0.5751	0.7772	0.5907	0.8678	0.9499	0.9842	0.9388	0.6639	0.6757	0.5069	0.6708	0.7210	0.7603	0.8852	0.8482	0.6683	0.4588	0.6207	0.4786	0.7296	0.8797	0.9708
0.0807	33.0	6666	0.4811	0.7177	0.8218	0.9079	0.9710	0.8003	0.8216	0.6509	0.7969	0.8468	0.8685	0.9483	0.9088	0.8122	0.5717	0.7543	0.5995	0.8662	0.9477	0.9843	0.9387	0.6627	0.6763	0.5087	0.6744	0.7237	0.7632	0.8847	0.8476	0.6656	0.4574	0.6202	0.4806	0.7295	0.8799	0.9707
0.0804	34.0	6868	0.4824	0.7176	0.8220	0.9079	0.9699	0.7993	0.8269	0.6392	0.8077	0.8321	0.8793	0.9479	0.9114	0.8028	0.5786	0.7535	0.5934	0.8784	0.9469	0.9849	0.9384	0.6621	0.6708	0.5042	0.6736	0.7208	0.7631	0.8879	0.8515	0.6696	0.4597	0.6227	0.4789	0.7280	0.8794	0.9706
0.0769	35.0	7070	0.4960	0.7178	0.8214	0.9081	0.9694	0.8053	0.8223	0.6416	0.7990	0.8403	0.8779	0.9476	0.9132	0.8100	0.5511	0.7574	0.6043	0.8715	0.9481	0.9829	0.9386	0.6632	0.6749	0.5031	0.6727	0.7218	0.7654	0.8878	0.8495	0.6725	0.4490	0.6229	0.4840	0.7285	0.8798	0.9706
0.0776	36.0	7272	0.4918	0.7168	0.8219	0.9077	0.9697	0.8043	0.8247	0.6536	0.8009	0.8346	0.8787	0.9517	0.9077	0.8046	0.5543	0.7730	0.5920	0.8680	0.9473	0.9847	0.9387	0.6631	0.6732	0.5020	0.6755	0.7227	0.7651	0.8844	0.8483	0.6669	0.4485	0.6235	0.4788	0.7277	0.8795	0.9704
0.0782	37.0	7474	0.4938	0.7171	0.8237	0.9081	0.9731	0.7905	0.8343	0.6444	0.8066	0.8508	0.8759	0.9471	0.9091	0.8169	0.5733	0.7636	0.5973	0.8642	0.9479	0.9843	0.9392	0.6621	0.6678	0.5046	0.6731	0.7221	0.7623	0.8866	0.8501	0.6702	0.4521	0.6232	0.4812	0.7300	0.8791	0.9706
0.0758	38.0	7676	0.5168	0.7151	0.8202	0.9069	0.9760	0.7477	0.8282	0.6616	0.8080	0.8463	0.8742	0.9456	0.9139	0.8158	0.5488	0.7629	0.5967	0.8654	0.9481	0.9845	0.9359	0.6395	0.6683	0.5066	0.6746	0.7230	0.7636	0.8883	0.8498	0.6677	0.4439	0.6222	0.4794	0.7285	0.8797	0.9706
0.0745	39.0	7878	0.5100	0.7173	0.8203	0.9081	0.9725	0.7948	0.8194	0.6488	0.8033	0.8417	0.8719	0.9457	0.9111	0.8140	0.5465	0.7558	0.5942	0.8723	0.9471	0.9850	0.9388	0.6622	0.6752	0.5043	0.6732	0.7239	0.7649	0.8869	0.8501	0.6683	0.4472	0.6225	0.4804	0.7295	0.8793	0.9708
0.0719	40.0	8080	0.5147	0.7181	0.8233	0.9083	0.9706	0.8003	0.8253	0.6494	0.8033	0.8491	0.8741	0.9453	0.9125	0.8069	0.5725	0.7652	0.5952	0.8685	0.9493	0.9847	0.9391	0.6636	0.6737	0.5047	0.6761	0.7237	0.7627	0.8859	0.8501	0.6725	0.4535	0.6223	0.4818	0.7291	0.8796	0.9706
0.0731	41.0	8282	0.5221	0.7173	0.8217	0.9083	0.9726	0.7939	0.8299	0.6356	0.8109	0.8419	0.8685	0.9448	0.9132	0.8095	0.5581	0.7649	0.6038	0.8674	0.9478	0.9846	0.9393	0.6625	0.6710	0.5008	0.6714	0.7237	0.7655	0.8859	0.8475	0.6690	0.4523	0.6234	0.4846	0.7296	0.8798	0.9708
0.0723	42.0	8484	0.5245	0.7176	0.8230	0.9084	0.9711	0.7990	0.8275	0.6417	0.8100	0.8396	0.8831	0.9465	0.9126	0.8083	0.5616	0.7649	0.6050	0.8640	0.9497	0.9839	0.9396	0.6645	0.6709	0.5026	0.6707	0.7224	0.7624	0.8864	0.8493	0.6709	0.4531	0.6237	0.4845	0.7304	0.8794	0.9706
0.0706	43.0	8686	0.5327	0.7176	0.8226	0.9081	0.9710	0.7981	0.8270	0.6415	0.8121	0.8409	0.8774	0.9466	0.9109	0.8068	0.5685	0.7562	0.6078	0.8638	0.9498	0.9833	0.9389	0.6618	0.6714	0.5029	0.6722	0.7222	0.7637	0.8858	0.8482	0.6700	0.4576	0.6225	0.4857	0.7297	0.8788	0.9706
0.0709	44.0	8888	0.5422	0.7179	0.8229	0.9082	0.9721	0.7964	0.8261	0.6365	0.8063	0.8440	0.8737	0.9464	0.9125	0.8077	0.5823	0.7622	0.5988	0.8687	0.9480	0.9843	0.9394	0.6630	0.6712	0.5026	0.6733	0.7231	0.7639	0.8876	0.8503	0.6704	0.4573	0.6245	0.4814	0.7285	0.8794	0.9707
0.0683	45.0	9090	0.5411	0.7182	0.8223	0.9083	0.9722	0.7947	0.8254	0.6432	0.8020	0.8436	0.8730	0.9457	0.9137	0.8080	0.5735	0.7592	0.6024	0.8688	0.9483	0.9836	0.9390	0.6613	0.6734	0.5018	0.6729	0.7237	0.7636	0.8871	0.8498	0.6718	0.4586	0.6245	0.4837	0.7292	0.8796	0.9707
0.0686	46.0	9292	0.5469	0.7182	0.8228	0.9083	0.9715	0.7973	0.8247	0.6366	0.8079	0.8475	0.8720	0.9456	0.9139	0.8086	0.5793	0.7554	0.6059	0.8661	0.9495	0.9834	0.9388	0.6614	0.6704	0.5012	0.6729	0.7241	0.7626	0.8875	0.8501	0.6724	0.4604	0.6232	0.4853	0.7301	0.8795	0.9707
0.0673	47.0	9494	0.5504	0.7178	0.8220	0.9082	0.9735	0.7857	0.8243	0.6351	0.8083	0.8416	0.8741	0.9473	0.9124	0.8112	0.5724	0.7629	0.6056	0.8660	0.9475	0.9847	0.9389	0.6585	0.6707	0.5021	0.6726	0.7233	0.7642	0.8870	0.8504	0.6718	0.4571	0.6235	0.4844	0.7298	0.8795	0.9707
0.0667	48.0	9696	0.5545	0.7179	0.8221	0.9083	0.9722	0.7925	0.8240	0.6389	0.8065	0.8420	0.8738	0.9461	0.9132	0.8114	0.5671	0.7628	0.6023	0.8671	0.9491	0.9843	0.9389	0.6613	0.6711	0.5024	0.6729	0.7229	0.7634	0.8876	0.8501	0.6722	0.4564	0.6232	0.4844	0.7297	0.8794	0.9707
0.0677	49.0	9898	0.5565	0.7179	0.8224	0.9083	0.9719	0.7950	0.8279	0.6404	0.8029	0.8418	0.8759	0.9477	0.9114	0.8106	0.5707	0.7605	0.6001	0.8690	0.9480	0.9844	0.9390	0.6622	0.6699	0.5027	0.6732	0.7231	0.7639	0.8872	0.8502	0.6717	0.4576	0.6238	0.4826	0.7293	0.8795	0.9707
0.0674	50.0	10100	0.5563	0.7180	0.8224	0.9083	0.9716	0.7949	0.8240	0.6408	0.8057	0.8434	0.8762	0.9451	0.9122	0.8112	0.5699	0.7605	0.6010	0.8678	0.9495	0.9841	0.9391	0.6620	0.6707	0.5025	0.6726	0.7228	0.7637	0.8882	0.8506	0.6721	0.4571	0.6238	0.4831	0.7293	0.8792	0.9707

Framework versions

Transformers 4.42.4
Pytorch 2.3.1+cu121
Datasets 2.19.1
Tokenizers 0.19.1

Как запустить

На своей GPU в ailist. Цена — по текущим предложениям провайдеров

GPU в ailist

Универсальный рантайм · image-segmentation

Можно запустить

Подбираем GPU…

Класс архитектуры: SegformerForSemanticSegmentation
Ревизия весов: 8059703