lr0.0001_bs16_0509_1514

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: 0.2220
• Mean Iou: 0.4787
• Mean Accuracy: 0.8026
• Overall Accuracy: 0.8579
• Accuracy Land: nan
• Accuracy Upwelling: 0.9308
• Accuracy Not Upwelling: 0.6744
• Iou Land: 0.0
• Iou Upwelling: 0.8552
• Iou Not Upwelling: 0.5811
• Dice Macro: 0.8605
• Dice Micro: 0.9054

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
1.0733	0.8	20	1.0873	0.2354	0.4782	0.5349	nan	0.6093	0.3470	0.0	0.4941	0.2120	0.2696	0.3006
0.8574	1.6	40	0.8671	0.3476	0.6306	0.7635	nan	0.9393	0.3220	0.0	0.7597	0.2832	0.7181	0.8199
0.6299	2.4	60	0.5997	0.3465	0.6203	0.7681	nan	0.9582	0.2825	0.0	0.7780	0.2616	0.7261	0.8418
0.4894	3.2	80	0.4757	0.4100	0.7202	0.7987	nan	0.9030	0.5374	0.0	0.7893	0.4408	0.7917	0.8583
0.4301	4.0	100	0.4114	0.4331	0.7610	0.8133	nan	0.8815	0.6404	0.0	0.7987	0.5004	0.8132	0.8678
0.4131	4.8	120	0.3786	0.4301	0.7441	0.8223	nan	0.9243	0.5639	0.0	0.8125	0.4777	0.8054	0.8680
0.3352	5.6	140	0.3519	0.4457	0.7782	0.8253	nan	0.8863	0.6701	0.0	0.8120	0.5250	0.8263	0.8764
0.343	6.4	160	0.3306	0.4480	0.7673	0.8350	nan	0.9237	0.6110	0.0	0.8265	0.5176	0.8297	0.8851
0.3601	7.2	180	0.3195	0.4367	0.7458	0.8196	nan	0.9185	0.5731	0.0	0.8244	0.4856	0.8248	0.8833
0.3037	8.0	200	0.3118	0.4512	0.7755	0.8329	nan	0.9080	0.6429	0.0	0.8252	0.5284	0.8351	0.8867
0.2972	8.8	220	0.2943	0.4598	0.7795	0.8459	nan	0.9345	0.6245	0.0	0.8371	0.5421	0.8412	0.8926
0.2874	9.6	240	0.2955	0.4526	0.7802	0.8337	nan	0.9014	0.6590	0.0	0.8265	0.5312	0.8386	0.8895
0.3222	10.4	260	0.2955	0.4423	0.7517	0.8368	nan	0.9469	0.5564	0.0	0.8317	0.4951	0.8268	0.8884
0.3449	11.2	280	0.2936	0.4408	0.7662	0.8123	nan	0.8730	0.6594	0.0	0.8133	0.5092	0.8328	0.8831
0.2903	12.0	300	0.3006	0.4328	0.7378	0.8065	nan	0.8971	0.5784	0.0	0.8237	0.4748	0.8259	0.8843
0.2584	12.8	320	0.2797	0.4578	0.7803	0.8411	nan	0.9233	0.6374	0.0	0.8316	0.5419	0.8409	0.8909
0.276	13.6	340	0.2827	0.4274	0.7235	0.8120	nan	0.9267	0.5203	0.0	0.8312	0.4511	0.8206	0.8874
0.3025	14.4	360	0.2700	0.4592	0.7862	0.8360	nan	0.9015	0.6708	0.0	0.8320	0.5454	0.8459	0.8937
0.2609	15.2	380	0.2854	0.4315	0.7306	0.8138	nan	0.9262	0.5350	0.0	0.8296	0.4649	0.8233	0.8860
0.2781	16.0	400	0.2847	0.4478	0.7606	0.8405	nan	0.9451	0.5760	0.0	0.8326	0.5109	0.8275	0.8862
0.2491	16.8	420	0.2719	0.4496	0.7628	0.8308	nan	0.9210	0.6046	0.0	0.8353	0.5137	0.8382	0.8916
0.3232	17.6	440	0.2723	0.4573	0.7904	0.8318	nan	0.8858	0.6951	0.0	0.8242	0.5476	0.8441	0.8902
0.3251	18.4	460	0.2679	0.4458	0.7516	0.8385	nan	0.9519	0.5513	0.0	0.8393	0.4981	0.8313	0.8922
0.2918	19.2	480	0.2640	0.4545	0.7721	0.8435	nan	0.9364	0.6078	0.0	0.8344	0.5290	0.8372	0.8920
0.2652	20.0	500	0.2609	0.4518	0.7648	0.8289	nan	0.9134	0.6162	0.0	0.8362	0.5191	0.8415	0.8943
0.2659	20.8	520	0.2569	0.4535	0.7696	0.8301	nan	0.9091	0.6301	0.0	0.8380	0.5225	0.8436	0.8953
0.2421	21.6	540	0.2576	0.4687	0.7912	0.8547	nan	0.9403	0.6421	0.0	0.8433	0.5629	0.8482	0.8966
0.2691	22.4	560	0.2521	0.4511	0.7589	0.8414	nan	0.9494	0.5683	0.0	0.8426	0.5108	0.8381	0.8962
0.3014	23.2	580	0.2573	0.4667	0.7992	0.8455	nan	0.9064	0.6920	0.0	0.8350	0.5652	0.8487	0.8941
0.2588	24.0	600	0.2715	0.4476	0.7841	0.8162	nan	0.8578	0.7104	0.0	0.8119	0.5309	0.8405	0.8866
0.2843	24.8	620	0.2503	0.4570	0.7746	0.8333	nan	0.9110	0.6381	0.0	0.8385	0.5324	0.8467	0.8968
0.2384	25.6	640	0.2481	0.4585	0.7707	0.8428	nan	0.9397	0.6017	0.0	0.8443	0.5311	0.8460	0.8985
0.3404	26.4	660	0.2410	0.4675	0.7866	0.8542	nan	0.9403	0.6330	0.0	0.8484	0.5540	0.8484	0.8993
0.2827	27.2	680	0.2642	0.4417	0.7506	0.8188	nan	0.9071	0.5941	0.0	0.8335	0.4916	0.8339	0.8910
0.2257	28.0	700	0.2485	0.4591	0.7811	0.8340	nan	0.9052	0.6570	0.0	0.8360	0.5412	0.8475	0.8954
0.2477	28.8	720	0.2415	0.4631	0.7798	0.8458	nan	0.9323	0.6272	0.0	0.8466	0.5428	0.8492	0.8999
0.23	29.6	740	0.2624	0.4420	0.7568	0.8123	nan	0.8838	0.6299	0.0	0.8271	0.4990	0.8356	0.8890
0.2403	30.4	760	0.2487	0.4783	0.8259	0.8538	nan	0.8901	0.7618	0.0	0.8358	0.5991	0.8575	0.8970
0.2379	31.2	780	0.2596	0.4578	0.7748	0.8498	nan	0.9507	0.5989	0.0	0.8358	0.5375	0.8357	0.8902
0.2061	32.0	800	0.2384	0.4727	0.7980	0.8533	nan	0.9245	0.6716	0.0	0.8482	0.5699	0.8555	0.9019
0.223	32.8	820	0.2487	0.4649	0.7874	0.8474	nan	0.9268	0.6481	0.0	0.8404	0.5543	0.8467	0.8957
0.2432	33.6	840	0.2589	0.4650	0.7944	0.8496	nan	0.9218	0.6670	0.0	0.8341	0.5608	0.8439	0.8920
0.2624	34.4	860	0.2431	0.4593	0.7732	0.8405	nan	0.9300	0.6163	0.0	0.8452	0.5327	0.8469	0.8988
0.2419	35.2	880	0.2491	0.4714	0.8095	0.8498	nan	0.9021	0.7168	0.0	0.8349	0.5792	0.8516	0.8952
0.2937	36.0	900	0.2356	0.4677	0.7846	0.8487	nan	0.9328	0.6363	0.0	0.8499	0.5531	0.8529	0.9021
0.2544	36.8	920	0.2501	0.4526	0.7625	0.8409	nan	0.9432	0.5819	0.0	0.8418	0.5160	0.8385	0.8954
0.2796	37.6	940	0.2372	0.4747	0.8096	0.8529	nan	0.9095	0.7097	0.0	0.8425	0.5815	0.8568	0.9004
0.2408	38.4	960	0.2523	0.4534	0.7642	0.8415	nan	0.9451	0.5832	0.0	0.8405	0.5197	0.8383	0.8937
0.2616	39.2	980	0.2378	0.4667	0.7887	0.8476	nan	0.9249	0.6524	0.0	0.8440	0.5560	0.8519	0.9001
0.2252	40.0	1000	0.2464	0.4543	0.7697	0.8343	nan	0.9198	0.6197	0.0	0.8384	0.5246	0.8427	0.8950
0.2313	40.8	1020	0.2404	0.4570	0.7653	0.8427	nan	0.9461	0.5845	0.0	0.8479	0.5233	0.8455	0.9002
0.2319	41.6	1040	0.2420	0.4620	0.7779	0.8406	nan	0.9238	0.6320	0.0	0.8467	0.5393	0.8492	0.8989
0.2368	42.4	1060	0.2381	0.4693	0.7890	0.8581	nan	0.9456	0.6325	0.0	0.8490	0.5589	0.8490	0.9004
0.2641	43.2	1080	0.2402	0.4624	0.7769	0.8472	nan	0.9402	0.6136	0.0	0.8458	0.5413	0.8454	0.8979
0.2444	44.0	1100	0.2420	0.4609	0.7886	0.8351	nan	0.8958	0.6813	0.0	0.8361	0.5466	0.8497	0.8960
0.263	44.8	1120	0.2461	0.4648	0.7866	0.8520	nan	0.9392	0.6340	0.0	0.8400	0.5543	0.8449	0.8953
0.229	45.6	1140	0.2551	0.4443	0.7475	0.8295	nan	0.9397	0.5554	0.0	0.8390	0.4938	0.8351	0.8938
0.2396	46.4	1160	0.2354	0.4733	0.8039	0.8515	nan	0.9146	0.6933	0.0	0.8427	0.5772	0.8555	0.8998
0.2252	47.2	1180	0.2297	0.4758	0.8018	0.8548	nan	0.9241	0.6795	0.0	0.8503	0.5771	0.8587	0.9041
0.2897	48.0	1200	0.2383	0.4639	0.7877	0.8423	nan	0.9147	0.6607	0.0	0.8398	0.5518	0.8505	0.8982
0.2215	48.8	1220	0.2449	0.4617	0.7989	0.8376	nan	0.8872	0.7105	0.0	0.8281	0.5569	0.8493	0.8944
0.2529	49.6	1240	0.2372	0.4707	0.7967	0.8505	nan	0.9229	0.6705	0.0	0.8429	0.5691	0.8539	0.8992
0.2418	50.4	1260	0.2323	0.4739	0.8026	0.8564	nan	0.9251	0.6802	0.0	0.8470	0.5746	0.8540	0.9007
0.2532	51.2	1280	0.2388	0.4590	0.7746	0.8395	nan	0.9252	0.6240	0.0	0.8427	0.5342	0.8462	0.8975
0.238	52.0	1300	0.2477	0.4572	0.7817	0.8296	nan	0.8931	0.6704	0.0	0.8335	0.5382	0.8470	0.8940
0.2484	52.8	1320	0.2397	0.4683	0.7921	0.8505	nan	0.9257	0.6586	0.0	0.8449	0.5601	0.8520	0.8995
0.2433	53.6	1340	0.2387	0.4692	0.7880	0.8544	nan	0.9432	0.6329	0.0	0.8478	0.5596	0.8507	0.8997
0.2378	54.4	1360	0.2480	0.4596	0.7841	0.8405	nan	0.9153	0.6530	0.0	0.8347	0.5442	0.8441	0.8931
0.2319	55.2	1380	0.2304	0.4685	0.7923	0.8461	nan	0.9173	0.6673	0.0	0.8440	0.5616	0.8545	0.9008
0.2339	56.0	1400	0.2347	0.4715	0.8022	0.8519	nan	0.9163	0.6881	0.0	0.8418	0.5726	0.8534	0.8991
0.2436	56.8	1420	0.2414	0.4611	0.7885	0.8436	nan	0.9157	0.6613	0.0	0.8329	0.5503	0.8458	0.8946
0.2571	57.6	1440	0.2436	0.4644	0.7854	0.8486	nan	0.9322	0.6387	0.0	0.8420	0.5514	0.8465	0.8965
0.2254	58.4	1460	0.2321	0.4727	0.7982	0.8537	nan	0.9251	0.6713	0.0	0.8488	0.5694	0.8554	0.9023
0.243	59.2	1480	0.2386	0.4688	0.7889	0.8530	nan	0.9380	0.6399	0.0	0.8470	0.5595	0.8495	0.8984
0.1834	60.0	1500	0.2361	0.4591	0.7818	0.8369	nan	0.9098	0.6538	0.0	0.8353	0.5419	0.8480	0.8964
0.2446	60.8	1520	0.2334	0.4666	0.7897	0.8443	nan	0.9156	0.6637	0.0	0.8458	0.5539	0.8537	0.9011
0.1942	61.6	1540	0.2330	0.4623	0.7825	0.8440	nan	0.9240	0.6409	0.0	0.8427	0.5442	0.8487	0.8987
0.2226	62.4	1560	0.2320	0.4683	0.7842	0.8570	nan	0.9521	0.6162	0.0	0.8493	0.5556	0.8492	0.9008
0.2527	63.2	1580	0.2423	0.4651	0.7853	0.8505	nan	0.9356	0.6350	0.0	0.8439	0.5513	0.8465	0.8973
0.2486	64.0	1600	0.2402	0.4628	0.7877	0.8479	nan	0.9265	0.6490	0.0	0.8369	0.5515	0.8453	0.8952
0.2855	64.8	1620	0.2313	0.4734	0.7985	0.8524	nan	0.9238	0.6732	0.0	0.8482	0.5721	0.8566	0.9020
0.2238	65.6	1640	0.2439	0.4538	0.7725	0.8346	nan	0.9134	0.6316	0.0	0.8372	0.5242	0.8421	0.8947
0.216	66.4	1660	0.2494	0.4483	0.7665	0.8265	nan	0.9051	0.6279	0.0	0.8284	0.5165	0.8394	0.8917
0.2399	67.2	1680	0.2389	0.4718	0.7949	0.8565	nan	0.9380	0.6518	0.0	0.8466	0.5687	0.8504	0.8990
0.2318	68.0	1700	0.2408	0.4565	0.7725	0.8351	nan	0.9188	0.6262	0.0	0.8404	0.5292	0.8452	0.8959
0.221	68.8	1720	0.2467	0.4538	0.7724	0.8301	nan	0.9076	0.6372	0.0	0.8325	0.5290	0.8427	0.8930
0.2251	69.6	1740	0.2325	0.4674	0.7894	0.8487	nan	0.9267	0.6522	0.0	0.8456	0.5566	0.8534	0.9015
0.228	70.4	1760	0.2369	0.4616	0.7808	0.8458	nan	0.9305	0.6312	0.0	0.8425	0.5422	0.8465	0.8975
0.2677	71.2	1780	0.2336	0.4602	0.7803	0.8377	nan	0.9140	0.6465	0.0	0.8408	0.5398	0.8489	0.8981
0.2552	72.0	1800	0.2351	0.4712	0.8046	0.8463	nan	0.9016	0.7076	0.0	0.8397	0.5739	0.8579	0.9007
0.2282	72.8	1820	0.2261	0.4772	0.8103	0.8537	nan	0.9111	0.7094	0.0	0.8456	0.5860	0.8601	0.9028
0.2132	73.6	1840	0.2329	0.4706	0.7991	0.8525	nan	0.9243	0.6739	0.0	0.8397	0.5719	0.8511	0.8970
0.2187	74.4	1860	0.2415	0.4589	0.7818	0.8392	nan	0.9151	0.6484	0.0	0.8349	0.5419	0.8456	0.8951
0.2647	75.2	1880	0.2318	0.4697	0.7947	0.8505	nan	0.9229	0.6665	0.0	0.8448	0.5643	0.8532	0.9004
0.2243	76.0	1900	0.2296	0.4729	0.8023	0.8531	nan	0.9196	0.6850	0.0	0.8447	0.5741	0.8557	0.9006
0.231	76.8	1920	0.2368	0.4682	0.7951	0.8493	nan	0.9199	0.6703	0.0	0.8419	0.5628	0.8526	0.8995
0.2156	77.6	1940	0.2368	0.4676	0.7992	0.8455	nan	0.9083	0.6901	0.0	0.8349	0.5678	0.8530	0.8974
0.2476	78.4	1960	0.2386	0.4545	0.7727	0.8341	nan	0.9153	0.6300	0.0	0.8350	0.5284	0.8430	0.8946
0.2317	79.2	1980	0.2338	0.4702	0.8045	0.8485	nan	0.9066	0.7025	0.0	0.8376	0.5731	0.8544	0.8986
0.2164	80.0	2000	0.2319	0.4724	0.8007	0.8498	nan	0.9154	0.6861	0.0	0.8443	0.5729	0.8557	0.9002
0.1998	80.8	2020	0.2271	0.4733	0.8054	0.8499	nan	0.9085	0.7022	0.0	0.8436	0.5762	0.8585	0.9020
0.2466	81.6	2040	0.2276	0.4795	0.8077	0.8587	nan	0.9263	0.6891	0.0	0.8521	0.5864	0.8603	0.9037
0.2358	82.4	2060	0.2326	0.4692	0.7899	0.8524	nan	0.9339	0.6459	0.0	0.8471	0.5604	0.8514	0.9001
0.2125	83.2	2080	0.2408	0.4610	0.7837	0.8405	nan	0.9160	0.6514	0.0	0.8370	0.5459	0.8477	0.8969
0.2593	84.0	2100	0.2293	0.4703	0.7979	0.8478	nan	0.9131	0.6826	0.0	0.8451	0.5659	0.8557	0.9016
0.2266	84.8	2120	0.2400	0.4603	0.7844	0.8369	nan	0.9053	0.6634	0.0	0.8367	0.5443	0.8491	0.8970
0.2231	85.6	2140	0.2332	0.4658	0.7915	0.8427	nan	0.9100	0.6730	0.0	0.8406	0.5567	0.8537	0.9000
0.2282	86.4	2160	0.2301	0.4669	0.7885	0.8474	nan	0.9255	0.6515	0.0	0.8438	0.5568	0.8515	0.8995
0.2496	87.2	2180	0.2344	0.4666	0.7905	0.8449	nan	0.9147	0.6663	0.0	0.8446	0.5551	0.8522	0.8996
0.2511	88.0	2200	0.2370	0.4660	0.7931	0.8437	nan	0.9111	0.6750	0.0	0.8404	0.5577	0.8520	0.8983
0.2392	88.8	2220	0.2297	0.4721	0.7964	0.8523	nan	0.9268	0.6660	0.0	0.8470	0.5694	0.8557	0.9020
0.2278	89.6	2240	0.2257	0.4724	0.7965	0.8528	nan	0.9253	0.6677	0.0	0.8497	0.5674	0.8553	0.9020
0.2523	90.4	2260	0.2312	0.4706	0.8017	0.8487	nan	0.9091	0.6942	0.0	0.8418	0.5700	0.8557	0.9003
0.2214	91.2	2280	0.2259	0.4745	0.7994	0.8541	nan	0.9253	0.6735	0.0	0.8499	0.5735	0.8569	0.9027
0.2594	92.0	2300	0.2215	0.4732	0.7987	0.8545	nan	0.9282	0.6693	0.0	0.8480	0.5714	0.8570	0.9029
0.235	92.8	2320	0.2336	0.4678	0.7924	0.8506	nan	0.9266	0.6582	0.0	0.8435	0.5600	0.8505	0.8986
0.2179	93.6	2340	0.2260	0.4745	0.8009	0.8536	nan	0.9231	0.6788	0.0	0.8484	0.5750	0.8581	0.9030
0.2195	94.4	2360	0.2176	0.4785	0.8020	0.8595	nan	0.9356	0.6683	0.0	0.8543	0.5813	0.8606	0.9062
0.2355	95.2	2380	0.2253	0.4782	0.8082	0.8571	nan	0.9209	0.6955	0.0	0.8493	0.5853	0.8602	0.9039
0.2549	96.0	2400	0.2486	0.4555	0.7784	0.8346	nan	0.9087	0.6482	0.0	0.8318	0.5348	0.8423	0.8923
0.2084	96.8	2420	0.2264	0.4720	0.7972	0.8540	nan	0.9286	0.6657	0.0	0.8475	0.5686	0.8564	0.9029
0.2325	97.6	2440	0.2403	0.4625	0.7851	0.8418	nan	0.9172	0.6530	0.0	0.8400	0.5474	0.8477	0.8957
0.2425	98.4	2460	0.2361	0.4695	0.7969	0.8489	nan	0.9175	0.6763	0.0	0.8423	0.5662	0.8526	0.8987
0.2366	99.2	2480	0.2291	0.4736	0.7995	0.8527	nan	0.9217	0.6774	0.0	0.8483	0.5726	0.8558	0.9016
0.2256	100.0	2500	0.2220	0.4787	0.8026	0.8579	nan	0.9308	0.6744	0.0	0.8552	0.5811	0.8605	0.9054

Framework versions

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