lr0.0001_bs16_0509_1312

This model is a fine-tuned version of nvidia/mit-b0 on the greenkwd/upwellingdetection_SST dataset. It achieves the following results on the evaluation set:

• Loss: 0.2346
• Mean Iou: 0.4702
• Mean Accuracy: 0.7945
• Overall Accuracy: 0.8491
• Accuracy Land: nan
• Accuracy Upwelling: 0.9236
• Accuracy Not Upwelling: 0.6653
• Iou Land: 0.0
• Iou Upwelling: 0.8435
• Iou Not Upwelling: 0.5672
• Dice Macro: 0.8549
• Dice Micro: 0.9006

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Land	Accuracy Upwelling	Accuracy Not Upwelling	Iou Land	Iou Upwelling	Iou Not Upwelling	Dice Macro	Dice Micro
1.0224	0.8	20	1.0716	0.2616	0.6381	0.5648	nan	0.4643	0.8118	0.0	0.4338	0.3512	0.4026	0.4031
0.8487	1.6	40	0.8774	0.3634	0.7053	0.7328	nan	0.7704	0.6401	0.0	0.6803	0.4100	0.7370	0.7857
0.6645	2.4	60	0.6208	0.3492	0.6305	0.7699	nan	0.9530	0.3081	0.0	0.7697	0.2779	0.7291	0.8378
0.5226	3.2	80	0.4942	0.3968	0.7026	0.7874	nan	0.9063	0.4989	0.0	0.7777	0.4126	0.7826	0.8537
0.4647	4.0	100	0.4295	0.4169	0.7317	0.8060	nan	0.9070	0.5563	0.0	0.7939	0.4569	0.7988	0.8628
0.4104	4.8	120	0.4055	0.4187	0.7283	0.8120	nan	0.9250	0.5317	0.0	0.8025	0.4536	0.7984	0.8648
0.4076	5.6	140	0.3841	0.4376	0.7729	0.8136	nan	0.8681	0.6777	0.0	0.7969	0.5158	0.8220	0.8712
0.3611	6.4	160	0.3652	0.4496	0.7873	0.8284	nan	0.8834	0.6912	0.0	0.8088	0.5399	0.8299	0.8778
0.3466	7.2	180	0.3495	0.4225	0.7280	0.8021	nan	0.9019	0.5541	0.0	0.8119	0.4555	0.8146	0.8766
0.316	8.0	200	0.3381	0.4442	0.7658	0.8281	nan	0.9124	0.6192	0.0	0.8165	0.5160	0.8267	0.8806
0.3123	8.8	220	0.3184	0.4373	0.7462	0.8308	nan	0.9466	0.5459	0.0	0.8237	0.4882	0.8241	0.8852
0.2886	9.6	240	0.3166	0.4363	0.7576	0.8086	nan	0.8777	0.6374	0.0	0.8098	0.4992	0.8289	0.8813
0.2959	10.4	260	0.3131	0.4434	0.7582	0.8353	nan	0.9365	0.5799	0.0	0.8254	0.5047	0.8258	0.8848
0.3001	11.2	280	0.3111	0.4340	0.7450	0.8179	nan	0.9193	0.5706	0.0	0.8163	0.4856	0.8245	0.8829
0.2919	12.0	300	0.3101	0.4393	0.7671	0.8110	nan	0.8704	0.6638	0.0	0.8048	0.5133	0.8314	0.8810
0.278	12.8	320	0.2937	0.4524	0.7805	0.8312	nan	0.9018	0.6592	0.0	0.8185	0.5388	0.8386	0.8867
0.3074	13.6	340	0.2927	0.4306	0.7307	0.8156	nan	0.9302	0.5312	0.0	0.8277	0.4642	0.8242	0.8882
0.3108	14.4	360	0.3043	0.4368	0.7757	0.8029	nan	0.8402	0.7113	0.0	0.7915	0.5188	0.8320	0.8779
0.27	15.2	380	0.2989	0.4213	0.7184	0.8130	nan	0.9439	0.4928	0.0	0.8205	0.4434	0.8139	0.8825
0.3288	16.0	400	0.3112	0.4528	0.7862	0.8387	nan	0.9099	0.6625	0.0	0.8140	0.5443	0.8258	0.8753
0.2526	16.8	420	0.2946	0.4308	0.7346	0.8193	nan	0.9347	0.5346	0.0	0.8235	0.4690	0.8221	0.8845
0.2689	17.6	440	0.2791	0.4499	0.7757	0.8242	nan	0.8900	0.6614	0.0	0.8202	0.5296	0.8395	0.8879
0.3047	18.4	460	0.2772	0.4528	0.7723	0.8334	nan	0.9157	0.6288	0.0	0.8292	0.5291	0.8380	0.8897
0.3546	19.2	480	0.3067	0.4285	0.7391	0.8287	nan	0.9508	0.5275	0.0	0.8128	0.4727	0.8096	0.8745
0.224	20.0	500	0.2791	0.4319	0.7308	0.8131	nan	0.9239	0.5377	0.0	0.8292	0.4665	0.8255	0.8882
0.2624	20.8	520	0.2736	0.4494	0.7785	0.8242	nan	0.8867	0.6702	0.0	0.8159	0.5322	0.8401	0.8882
0.2491	21.6	540	0.2672	0.4496	0.7653	0.8371	nan	0.9340	0.5966	0.0	0.8308	0.5180	0.8369	0.8916
0.2926	22.4	560	0.2667	0.4432	0.7516	0.8321	nan	0.9405	0.5628	0.0	0.8324	0.4970	0.8322	0.8912
0.2534	23.2	580	0.2677	0.4512	0.7708	0.8352	nan	0.9226	0.6190	0.0	0.8278	0.5258	0.8370	0.8898
0.2531	24.0	600	0.2684	0.4540	0.7836	0.8313	nan	0.8951	0.6720	0.0	0.8218	0.5401	0.8421	0.8899
0.2393	24.8	620	0.2700	0.4616	0.7955	0.8369	nan	0.8930	0.6979	0.0	0.8256	0.5590	0.8476	0.8923
0.2357	25.6	640	0.2564	0.4603	0.7816	0.8378	nan	0.9158	0.6474	0.0	0.8356	0.5453	0.8486	0.8960
0.2435	26.4	660	0.2563	0.4619	0.7909	0.8409	nan	0.9080	0.6738	0.0	0.8320	0.5537	0.8474	0.8940
0.2793	27.2	680	0.2757	0.4351	0.7415	0.8140	nan	0.9109	0.5722	0.0	0.8260	0.4792	0.8284	0.8879
0.2163	28.0	700	0.2602	0.4570	0.7851	0.8314	nan	0.8951	0.6751	0.0	0.8260	0.5450	0.8463	0.8925
0.2273	28.8	720	0.2529	0.4564	0.7740	0.8426	nan	0.9336	0.6143	0.0	0.8375	0.5318	0.8433	0.8960
0.2613	29.6	740	0.2543	0.4567	0.7808	0.8323	nan	0.9009	0.6607	0.0	0.8319	0.5382	0.8468	0.8953
0.2311	30.4	760	0.2539	0.4504	0.7604	0.8352	nan	0.9369	0.5840	0.0	0.8376	0.5135	0.8412	0.8966
0.2483	31.2	780	0.2849	0.4388	0.7510	0.8379	nan	0.9581	0.5439	0.0	0.8204	0.4960	0.8182	0.8802
0.2112	32.0	800	0.2497	0.4598	0.7804	0.8412	nan	0.9228	0.6379	0.0	0.8373	0.5422	0.8472	0.8970
0.257	32.8	820	0.2642	0.4447	0.7548	0.8339	nan	0.9422	0.5675	0.0	0.8280	0.5059	0.8314	0.8896
0.2268	33.6	840	0.2545	0.4558	0.7797	0.8354	nan	0.9116	0.6478	0.0	0.8287	0.5385	0.8440	0.8933
0.2334	34.4	860	0.2640	0.4396	0.7460	0.8321	nan	0.9509	0.5411	0.0	0.8302	0.4886	0.8290	0.8902
0.2394	35.2	880	0.2566	0.4641	0.8006	0.8399	nan	0.8936	0.7076	0.0	0.8250	0.5673	0.8486	0.8924
0.2852	36.0	900	0.2466	0.4586	0.7755	0.8400	nan	0.9282	0.6228	0.0	0.8392	0.5365	0.8470	0.8977
0.296	36.8	920	0.2572	0.4610	0.7861	0.8423	nan	0.9189	0.6533	0.0	0.8327	0.5502	0.8447	0.8933
0.2949	37.6	940	0.2540	0.4567	0.7721	0.8455	nan	0.9447	0.5996	0.0	0.8380	0.5322	0.8413	0.8956
0.2362	38.4	960	0.2623	0.4510	0.7679	0.8346	nan	0.9250	0.6108	0.0	0.8313	0.5218	0.8379	0.8906
0.2631	39.2	980	0.2531	0.4568	0.7765	0.8387	nan	0.9229	0.6301	0.0	0.8334	0.5370	0.8443	0.8951
0.229	40.0	1000	0.2603	0.4407	0.7461	0.8223	nan	0.9236	0.5686	0.0	0.8327	0.4893	0.8329	0.8913
0.2179	40.8	1020	0.2453	0.4590	0.7762	0.8446	nan	0.9368	0.6156	0.0	0.8390	0.5381	0.8449	0.8968
0.2275	41.6	1040	0.2491	0.4585	0.7844	0.8333	nan	0.9006	0.6682	0.0	0.8307	0.5447	0.8473	0.8937
0.2604	42.4	1060	0.2495	0.4487	0.7596	0.8425	nan	0.9507	0.5686	0.0	0.8363	0.5097	0.8351	0.8942
0.2399	43.2	1080	0.2494	0.4580	0.7756	0.8399	nan	0.9281	0.6232	0.0	0.8357	0.5385	0.8436	0.8948
0.2614	44.0	1100	0.2511	0.4538	0.7732	0.8272	nan	0.9018	0.6447	0.0	0.8327	0.5287	0.8449	0.8935
0.262	44.8	1120	0.2822	0.4439	0.7689	0.8344	nan	0.9233	0.6145	0.0	0.8129	0.5188	0.8258	0.8798
0.2294	45.6	1140	0.2567	0.4446	0.7557	0.8219	nan	0.9107	0.6008	0.0	0.8315	0.5023	0.8365	0.8908
0.246	46.4	1160	0.2547	0.4499	0.7619	0.8421	nan	0.9499	0.5740	0.0	0.8348	0.5149	0.8359	0.8932
0.2351	47.2	1180	0.2429	0.4587	0.7792	0.8312	nan	0.9019	0.6566	0.0	0.8371	0.5391	0.8494	0.8974
0.2752	48.0	1200	0.2516	0.4547	0.7749	0.8366	nan	0.9193	0.6304	0.0	0.8321	0.5319	0.8430	0.8943
0.228	48.8	1220	0.2569	0.4499	0.7788	0.8251	nan	0.8866	0.6710	0.0	0.8191	0.5307	0.8417	0.8904
0.2511	49.6	1240	0.2435	0.4663	0.7881	0.8477	nan	0.9296	0.6466	0.0	0.8407	0.5581	0.8507	0.8981
0.2357	50.4	1260	0.2345	0.4696	0.7917	0.8517	nan	0.9322	0.6513	0.0	0.8454	0.5632	0.8530	0.9011
0.2554	51.2	1280	0.2480	0.4593	0.7795	0.8454	nan	0.9341	0.6248	0.0	0.8351	0.5428	0.8418	0.8932
0.2418	52.0	1300	0.2593	0.4567	0.7817	0.8303	nan	0.8980	0.6655	0.0	0.8274	0.5426	0.8462	0.8926
0.2548	52.8	1320	0.2449	0.4596	0.7769	0.8423	nan	0.9301	0.6238	0.0	0.8399	0.5389	0.8465	0.8971
0.2386	53.6	1340	0.2531	0.4578	0.7723	0.8490	nan	0.9543	0.5903	0.0	0.8381	0.5353	0.8397	0.8937
0.2501	54.4	1360	0.2502	0.4671	0.7969	0.8453	nan	0.9119	0.6820	0.0	0.8354	0.5657	0.8494	0.8942
0.2368	55.2	1380	0.2409	0.4700	0.8035	0.8455	nan	0.9023	0.7047	0.0	0.8351	0.5749	0.8553	0.8982
0.234	56.0	1400	0.2360	0.4769	0.8078	0.8551	nan	0.9206	0.6950	0.0	0.8431	0.5874	0.8579	0.9010
0.2368	56.8	1420	0.2390	0.4770	0.8101	0.8563	nan	0.9187	0.7014	0.0	0.8419	0.5891	0.8557	0.8990
0.2126	57.6	1440	0.2444	0.4669	0.7930	0.8460	nan	0.9183	0.6677	0.0	0.8382	0.5625	0.8498	0.8964
0.2635	58.4	1460	0.2399	0.4671	0.7949	0.8442	nan	0.9110	0.6789	0.0	0.8389	0.5625	0.8527	0.8986
0.268	59.2	1480	0.2499	0.4682	0.7959	0.8515	nan	0.9264	0.6654	0.0	0.8377	0.5670	0.8467	0.8936
0.1881	60.0	1500	0.2337	0.4671	0.7881	0.8468	nan	0.9267	0.6494	0.0	0.8430	0.5583	0.8523	0.8996
0.215	60.8	1520	0.2359	0.4661	0.7926	0.8422	nan	0.9083	0.6768	0.0	0.8397	0.5586	0.8530	0.8989
0.1991	61.6	1540	0.2378	0.4683	0.7993	0.8452	nan	0.9067	0.6919	0.0	0.8370	0.5678	0.8531	0.8977
0.2195	62.4	1560	0.2426	0.4668	0.7893	0.8538	nan	0.9394	0.6392	0.0	0.8412	0.5591	0.8477	0.8973
0.2181	63.2	1580	0.2451	0.4664	0.7898	0.8506	nan	0.9316	0.6480	0.0	0.8400	0.5591	0.8469	0.8959
0.236	64.0	1600	0.2425	0.4622	0.7831	0.8475	nan	0.9342	0.6320	0.0	0.8384	0.5481	0.8454	0.8961
0.2326	64.8	1620	0.2393	0.4704	0.7997	0.8470	nan	0.9101	0.6893	0.0	0.8402	0.5709	0.8549	0.8992
0.2286	65.6	1640	0.2465	0.4559	0.7764	0.8326	nan	0.9072	0.6456	0.0	0.8339	0.5337	0.8442	0.8941
0.2152	66.4	1660	0.2473	0.4568	0.7765	0.8363	nan	0.9176	0.6354	0.0	0.8333	0.5373	0.8451	0.8949
0.2313	67.2	1680	0.2418	0.4650	0.7879	0.8484	nan	0.9310	0.6448	0.0	0.8387	0.5563	0.8473	0.8964
0.2197	68.0	1700	0.2468	0.4534	0.7648	0.8332	nan	0.9276	0.6021	0.0	0.8388	0.5215	0.8429	0.8952
0.2111	68.8	1720	0.2510	0.4509	0.7640	0.8298	nan	0.9207	0.6074	0.0	0.8324	0.5203	0.8402	0.8930
0.2473	69.6	1740	0.2370	0.4663	0.7898	0.8466	nan	0.9233	0.6564	0.0	0.8410	0.5578	0.8519	0.8994
0.2286	70.4	1760	0.2437	0.4635	0.7854	0.8456	nan	0.9269	0.6439	0.0	0.8382	0.5522	0.8476	0.8964
0.2636	71.2	1780	0.2367	0.4627	0.7860	0.8385	nan	0.9101	0.6619	0.0	0.8381	0.5499	0.8508	0.8979
0.2529	72.0	1800	0.2335	0.4719	0.8077	0.8456	nan	0.8964	0.7191	0.0	0.8377	0.5780	0.8585	0.9002
0.2449	72.8	1820	0.2304	0.4730	0.7986	0.8522	nan	0.9262	0.6710	0.0	0.8446	0.5745	0.8569	0.9019
0.2266	73.6	1840	0.2413	0.4644	0.7862	0.8500	nan	0.9361	0.6363	0.0	0.8401	0.5532	0.8463	0.8960
0.2214	74.4	1860	0.2405	0.4650	0.7841	0.8457	nan	0.9316	0.6367	0.0	0.8418	0.5532	0.8496	0.8981
0.2631	75.2	1880	0.2433	0.4625	0.7843	0.8444	nan	0.9257	0.6429	0.0	0.8362	0.5513	0.8476	0.8966
0.2177	76.0	1900	0.2318	0.4733	0.7991	0.8526	nan	0.9256	0.6726	0.0	0.8443	0.5758	0.8564	0.9010
0.2316	76.8	1920	0.2342	0.4724	0.7989	0.8518	nan	0.9240	0.6738	0.0	0.8437	0.5733	0.8557	0.9008
0.2241	77.6	1940	0.2343	0.4708	0.7970	0.8501	nan	0.9224	0.6715	0.0	0.8434	0.5690	0.8551	0.9006
0.247	78.4	1960	0.2449	0.4585	0.7772	0.8362	nan	0.9178	0.6366	0.0	0.8355	0.5399	0.8463	0.8955
0.2263	79.2	1980	0.2313	0.4776	0.8045	0.8571	nan	0.9286	0.6804	0.0	0.8490	0.5836	0.8588	0.9031
0.2188	80.0	2000	0.2353	0.4692	0.7913	0.8475	nan	0.9247	0.6579	0.0	0.8439	0.5636	0.8536	0.8999
0.2067	80.8	2020	0.2231	0.4807	0.8073	0.8589	nan	0.9278	0.6867	0.0	0.8546	0.5876	0.8627	0.9062
0.2392	81.6	2040	0.2293	0.4715	0.7941	0.8489	nan	0.9250	0.6633	0.0	0.8471	0.5673	0.8572	0.9022
0.2267	82.4	2060	0.2376	0.4692	0.7956	0.8497	nan	0.9225	0.6687	0.0	0.8408	0.5667	0.8518	0.8981
0.2264	83.2	2080	0.2352	0.4737	0.8011	0.8525	nan	0.9226	0.6796	0.0	0.8441	0.5769	0.8551	0.9002
0.2402	84.0	2100	0.2329	0.4650	0.7862	0.8442	nan	0.9213	0.6511	0.0	0.8429	0.5520	0.8521	0.9006
0.2247	84.8	2120	0.2410	0.4584	0.7806	0.8324	nan	0.9037	0.6576	0.0	0.8334	0.5418	0.8482	0.8955
0.2154	85.6	2140	0.2317	0.4718	0.7968	0.8464	nan	0.9148	0.6788	0.0	0.8449	0.5705	0.8587	0.9025
0.2332	86.4	2160	0.2354	0.4704	0.7971	0.8485	nan	0.9179	0.6763	0.0	0.8414	0.5698	0.8544	0.8996
0.2355	87.2	2180	0.2396	0.4580	0.7733	0.8377	nan	0.9252	0.6215	0.0	0.8392	0.5346	0.8461	0.8969
0.2708	88.0	2200	0.2349	0.4683	0.7961	0.8461	nan	0.9137	0.6784	0.0	0.8400	0.5647	0.8531	0.8985
0.2443	88.8	2220	0.2378	0.4681	0.7904	0.8478	nan	0.9270	0.6538	0.0	0.8414	0.5630	0.8533	0.9001
0.2212	89.6	2240	0.2330	0.4701	0.7936	0.8496	nan	0.9231	0.6642	0.0	0.8473	0.5629	0.8538	0.9006
0.2197	90.4	2260	0.2304	0.4750	0.8044	0.8526	nan	0.9173	0.6916	0.0	0.8445	0.5804	0.8582	0.9016
0.2235	91.2	2280	0.2308	0.4699	0.7920	0.8506	nan	0.9293	0.6547	0.0	0.8460	0.5638	0.8539	0.9011
0.2169	92.0	2300	0.2264	0.4761	0.8042	0.8556	nan	0.9243	0.6841	0.0	0.8479	0.5805	0.8594	0.9035
0.2196	92.8	2320	0.2350	0.4674	0.7891	0.8497	nan	0.9307	0.6475	0.0	0.8432	0.5590	0.8502	0.8986
0.2096	93.6	2340	0.2321	0.4695	0.7955	0.8476	nan	0.9179	0.6731	0.0	0.8420	0.5666	0.8551	0.9005
0.2226	94.4	2360	0.2262	0.4719	0.7947	0.8540	nan	0.9335	0.6558	0.0	0.8474	0.5682	0.8553	0.9027
0.2295	95.2	2380	0.2274	0.4739	0.8012	0.8523	nan	0.9221	0.6803	0.0	0.8446	0.5770	0.8578	0.9020
0.2568	96.0	2400	0.2471	0.4599	0.7805	0.8386	nan	0.9173	0.6437	0.0	0.8370	0.5426	0.8456	0.8947
0.2186	96.8	2420	0.2241	0.4748	0.7996	0.8549	nan	0.9298	0.6694	0.0	0.8477	0.5765	0.8590	0.9040
0.2368	97.6	2440	0.2396	0.4661	0.7877	0.8446	nan	0.9215	0.6540	0.0	0.8437	0.5546	0.8509	0.8980
0.2295	98.4	2460	0.2381	0.4644	0.7892	0.8445	nan	0.9187	0.6597	0.0	0.8375	0.5556	0.8486	0.8963
0.2212	99.2	2480	0.2342	0.4741	0.7993	0.8522	nan	0.9240	0.6746	0.0	0.8468	0.5756	0.8565	0.9015
0.2186	100.0	2500	0.2346	0.4702	0.7945	0.8491	nan	0.9236	0.6653	0.0	0.8435	0.5672	0.8549	0.9006

Framework versions

• Transformers 4.41.2
• Pytorch 2.6.0+cu124
• Datasets 3.2.0
• Tokenizers 0.19.1