nutrition-extractor

This model is a fine-tuned version of microsoft/layoutlmv3-large on the v5 of the nutrient extraction dataset. It achieves the following results on the evaluation set:

• Loss: 0.0773
• Precision: 0.9403
• Recall: 0.9464
• F1: 0.9433
• Accuracy: 0.9888

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.9661	0.1896	15	1.1509	0.0	0.0	0.0	0.8105
0.9772	0.3791	30	0.8344	0.0452	0.0079	0.0134	0.8130
0.8483	0.5687	45	0.6955	0.1203	0.1271	0.1236	0.8443
0.6646	0.7583	60	0.6043	0.1366	0.1890	0.1586	0.8586
0.6051	0.9479	75	0.5304	0.2540	0.3138	0.2808	0.8792
0.5079	1.1374	90	0.4677	0.3022	0.3683	0.3320	0.8869
0.4924	1.3270	105	0.4053	0.3902	0.4566	0.4208	0.9011
0.4113	1.5166	120	0.3613	0.4406	0.5037	0.4700	0.9081
0.3866	1.7062	135	0.3171	0.4975	0.5582	0.5261	0.9216
0.3464	1.8957	150	0.2863	0.5246	0.6109	0.5645	0.9303
0.3171	2.0853	165	0.2557	0.5986	0.6918	0.6418	0.9458
0.2938	2.2749	180	0.2342	0.6400	0.7149	0.6754	0.9492
0.2678	2.4645	195	0.2214	0.6571	0.7518	0.7013	0.9536
0.2362	2.6540	210	0.2066	0.6873	0.7588	0.7213	0.9566
0.2175	2.8436	225	0.1944	0.7137	0.7810	0.7458	0.9593
0.2128	3.0332	240	0.1820	0.7432	0.8184	0.7790	0.9646
0.1968	3.2227	255	0.1728	0.7592	0.8202	0.7885	0.9662
0.185	3.4123	270	0.1656	0.7667	0.8276	0.7960	0.9675
0.1702	3.6019	285	0.1640	0.7512	0.8161	0.7823	0.9655
0.1686	3.7915	300	0.1541	0.7831	0.8341	0.8078	0.9675
0.1671	3.9810	315	0.1527	0.7696	0.8383	0.8025	0.9679
0.1532	4.1706	330	0.1421	0.8114	0.8567	0.8334	0.9710
0.1505	4.3602	345	0.1301	0.8185	0.8646	0.8409	0.9727
0.126	4.5498	360	0.1274	0.8330	0.8688	0.8505	0.9746
0.128	4.7393	375	0.1166	0.8395	0.8771	0.8579	0.9757
0.1313	4.9289	390	0.1158	0.8583	0.8872	0.8725	0.9767
0.1112	5.1185	405	0.1140	0.8447	0.8794	0.8617	0.9752
0.0968	5.3081	420	0.1002	0.8646	0.8970	0.8805	0.9787
0.1129	5.4976	435	0.0993	0.8605	0.8919	0.8759	0.9786
0.1109	5.6872	450	0.0974	0.8517	0.8812	0.8662	0.9779
0.0868	5.8768	465	0.0957	0.8741	0.9043	0.8889	0.9801
0.1022	6.0664	480	0.0922	0.8778	0.9099	0.8936	0.9808
0.0814	6.2559	495	0.0890	0.8821	0.9025	0.8922	0.9806
0.0843	6.4455	510	0.0946	0.8847	0.9039	0.8942	0.9809
0.0803	6.6351	525	0.0846	0.8951	0.9145	0.9047	0.9826
0.0936	6.8246	540	0.0856	0.8893	0.9168	0.9028	0.9823
0.08	7.0142	555	0.0806	0.8990	0.9168	0.9078	0.9831
0.0672	7.2038	570	0.0839	0.8859	0.9117	0.8987	0.9822
0.0675	7.3934	585	0.0836	0.8903	0.9150	0.9025	0.9827
0.0757	7.5829	600	0.0837	0.8927	0.9191	0.9057	0.9830
0.0719	7.7725	615	0.0803	0.8998	0.9214	0.9105	0.9833
0.0671	7.9621	630	0.0831	0.8985	0.9201	0.9091	0.9831
0.0661	8.1517	645	0.0760	0.9061	0.9228	0.9144	0.9840
0.051	8.3412	660	0.0780	0.9121	0.9302	0.9211	0.9848
0.0524	8.5308	675	0.0805	0.9071	0.9247	0.9158	0.9851
0.067	8.7204	690	0.0768	0.9050	0.9242	0.9145	0.9846
0.0647	8.9100	705	0.0802	0.9135	0.9274	0.9204	0.9844
0.0581	9.0995	720	0.0721	0.9064	0.9214	0.9138	0.9846
0.0461	9.2891	735	0.0762	0.9044	0.9307	0.9173	0.9850
0.0496	9.4787	750	0.0748	0.9120	0.9288	0.9203	0.9851
0.0495	9.6682	765	0.0799	0.9126	0.9311	0.9218	0.9854
0.0523	9.8578	780	0.0772	0.9103	0.9335	0.9217	0.9856
0.0517	10.0474	795	0.0809	0.9120	0.9238	0.9178	0.9847
0.0411	10.2370	810	0.0792	0.9118	0.9316	0.9216	0.9860
0.0435	10.4265	825	0.0735	0.9157	0.9339	0.9247	0.9855
0.0393	10.6161	840	0.0729	0.9095	0.9293	0.9193	0.9853
0.0445	10.8057	855	0.0761	0.9069	0.9316	0.9191	0.9856
0.0432	10.9953	870	0.0739	0.9056	0.9307	0.9180	0.9851
0.0387	11.1848	885	0.0699	0.9146	0.9348	0.9246	0.9869
0.042	11.3744	900	0.0711	0.9187	0.9298	0.9242	0.9863
0.0374	11.5640	915	0.0697	0.9210	0.9325	0.9268	0.9864
0.0328	11.7536	930	0.0746	0.9111	0.9325	0.9217	0.9861
0.0403	11.9431	945	0.0749	0.9131	0.9325	0.9227	0.9857
0.0327	12.1327	960	0.0731	0.9191	0.9395	0.9292	0.9868
0.0284	12.3223	975	0.0738	0.9172	0.9372	0.9271	0.9862
0.0314	12.5118	990	0.0711	0.9253	0.9330	0.9291	0.9867
0.0342	12.7014	1005	0.0775	0.9254	0.9339	0.9296	0.9862
0.038	12.8910	1020	0.0725	0.9271	0.9348	0.9310	0.9864
0.0262	13.0806	1035	0.0756	0.9313	0.9330	0.9321	0.9869
0.0282	13.2701	1050	0.0734	0.9215	0.9385	0.9299	0.9869
0.0251	13.4597	1065	0.0701	0.9254	0.9395	0.9324	0.9872
0.0304	13.6493	1080	0.0712	0.9296	0.9395	0.9345	0.9873
0.0327	13.8389	1095	0.0746	0.9209	0.9362	0.9285	0.9867
0.0318	14.0284	1110	0.0721	0.9265	0.9325	0.9295	0.9870
0.0293	14.2180	1125	0.0739	0.9290	0.9376	0.9333	0.9869
0.0251	14.4076	1140	0.0753	0.9281	0.9372	0.9326	0.9869
0.0269	14.5972	1155	0.0769	0.9205	0.9362	0.9283	0.9862
0.0263	14.7867	1170	0.0745	0.9209	0.9358	0.9283	0.9866
0.0239	14.9763	1185	0.0754	0.9298	0.9372	0.9335	0.9876
0.0214	15.1659	1200	0.0756	0.9298	0.9367	0.9332	0.9873
0.0218	15.3555	1215	0.0759	0.9294	0.9372	0.9333	0.9875
0.0212	15.5450	1230	0.0744	0.9275	0.9344	0.9309	0.9871
0.0243	15.7346	1245	0.0716	0.9224	0.9395	0.9309	0.9875
0.0268	15.9242	1260	0.0730	0.9220	0.9399	0.9309	0.9869
0.0229	16.1137	1275	0.0755	0.9170	0.9395	0.9281	0.9867
0.0201	16.3033	1290	0.0714	0.9300	0.9390	0.9345	0.9879
0.0224	16.4929	1305	0.0693	0.9297	0.9413	0.9355	0.9878
0.0243	16.6825	1320	0.0699	0.9253	0.9450	0.9351	0.9878
0.022	16.8720	1335	0.0733	0.9192	0.9404	0.9296	0.9868
0.0207	17.0616	1350	0.0685	0.9329	0.9376	0.9352	0.9884
0.0211	17.2512	1365	0.0749	0.9314	0.9418	0.9366	0.9877
0.0188	17.4408	1380	0.0731	0.9323	0.9418	0.9370	0.9876
0.0187	17.6303	1395	0.0734	0.9330	0.9390	0.9360	0.9878
0.0203	17.8199	1410	0.0732	0.9357	0.9409	0.9382	0.9879
0.0204	18.0095	1425	0.0729	0.9306	0.9418	0.9362	0.9878
0.0151	18.1991	1440	0.0718	0.9285	0.9418	0.9351	0.9879
0.016	18.3886	1455	0.0730	0.9309	0.9404	0.9356	0.9879
0.0225	18.5782	1470	0.0743	0.9326	0.9404	0.9365	0.9877
0.0158	18.7678	1485	0.0759	0.9298	0.9427	0.9362	0.9875
0.0169	18.9573	1500	0.0780	0.9348	0.9409	0.9378	0.9883
0.0167	19.1469	1515	0.0790	0.9347	0.9399	0.9373	0.9877
0.0163	19.3365	1530	0.0743	0.9336	0.9422	0.9379	0.9875
0.0158	19.5261	1545	0.0737	0.9370	0.9413	0.9391	0.9883
0.0183	19.7156	1560	0.0766	0.9247	0.9418	0.9332	0.9870
0.0154	19.9052	1575	0.0770	0.9322	0.9409	0.9365	0.9875
0.018	20.0948	1590	0.0758	0.9298	0.9432	0.9365	0.9873
0.0142	20.2844	1605	0.0744	0.9255	0.9413	0.9333	0.9877
0.0151	20.4739	1620	0.0730	0.9323	0.9418	0.9370	0.9883
0.018	20.6635	1635	0.0727	0.9321	0.9455	0.9387	0.9886
0.0137	20.8531	1650	0.0708	0.9375	0.9427	0.9401	0.9887
0.0139	21.0427	1665	0.0714	0.9330	0.9459	0.9394	0.9885
0.0134	21.2322	1680	0.0699	0.9322	0.9404	0.9363	0.9883
0.0155	21.4218	1695	0.0713	0.9330	0.9455	0.9392	0.9886
0.0142	21.6114	1710	0.0711	0.9365	0.9473	0.9419	0.9887
0.0147	21.8009	1725	0.0701	0.9335	0.9464	0.9399	0.9887
0.0129	21.9905	1740	0.0700	0.9364	0.9459	0.9411	0.9886
0.0131	22.1801	1755	0.0737	0.9399	0.9399	0.9399	0.9880
0.0121	22.3697	1770	0.0726	0.9312	0.9441	0.9376	0.9883
0.0133	22.5592	1785	0.0733	0.9249	0.9450	0.9349	0.9878
0.0126	22.7488	1800	0.0753	0.9331	0.9409	0.9370	0.9878
0.0133	22.9384	1815	0.0738	0.9286	0.9436	0.9361	0.9882
0.0113	23.1280	1830	0.0733	0.9327	0.9413	0.9370	0.9882
0.012	23.3175	1845	0.0774	0.9413	0.9413	0.9413	0.9884
0.0109	23.5071	1860	0.0772	0.9414	0.9427	0.9420	0.9887
0.0135	23.6967	1875	0.0782	0.9386	0.9390	0.9388	0.9879
0.0122	23.8863	1890	0.0767	0.9368	0.9385	0.9377	0.9880
0.0109	24.0758	1905	0.0763	0.9406	0.9445	0.9426	0.9887
0.0117	24.2654	1920	0.0763	0.9430	0.9404	0.9417	0.9884
0.013	24.4550	1935	0.0741	0.9415	0.9441	0.9428	0.9885
0.0092	24.6445	1950	0.0768	0.9326	0.9469	0.9397	0.9885
0.0101	24.8341	1965	0.0737	0.9328	0.9422	0.9375	0.9884
0.0092	25.0237	1980	0.0732	0.9385	0.9455	0.9420	0.9887
0.0084	25.2133	1995	0.0738	0.9331	0.9478	0.9404	0.9886
0.0088	25.4028	2010	0.0769	0.9365	0.9399	0.9382	0.9881
0.0116	25.5924	2025	0.0739	0.9402	0.9445	0.9424	0.9886
0.0101	25.7820	2040	0.0733	0.9399	0.9473	0.9436	0.9890
0.0122	25.9716	2055	0.0759	0.9375	0.9432	0.9403	0.9883
0.0098	26.1611	2070	0.0762	0.9415	0.9445	0.9430	0.9884
0.0088	26.3507	2085	0.0776	0.9393	0.9432	0.9412	0.9880
0.01	26.5403	2100	0.0769	0.9354	0.9427	0.9390	0.9878
0.0112	26.7299	2115	0.0750	0.9375	0.9422	0.9398	0.9881
0.0087	26.9194	2130	0.0745	0.9389	0.9441	0.9415	0.9883
0.0076	27.1090	2145	0.0728	0.9372	0.9445	0.9409	0.9885
0.0091	27.2986	2160	0.0749	0.9401	0.9436	0.9419	0.9886
0.01	27.4882	2175	0.0766	0.9414	0.9436	0.9425	0.9886
0.008	27.6777	2190	0.0774	0.9401	0.9432	0.9416	0.9885
0.0095	27.8673	2205	0.0784	0.9419	0.9436	0.9428	0.9886
0.0109	28.0569	2220	0.0776	0.9329	0.9450	0.9389	0.9883
0.0088	28.2464	2235	0.0794	0.9357	0.9413	0.9385	0.9879
0.0068	28.4360	2250	0.0800	0.9389	0.9441	0.9415	0.9886
0.0097	28.6256	2265	0.0817	0.9400	0.9418	0.9409	0.9880
0.0075	28.8152	2280	0.0804	0.9403	0.9395	0.9399	0.9878
0.0103	29.0047	2295	0.0750	0.9338	0.9450	0.9394	0.9887
0.01	29.1943	2310	0.0749	0.9333	0.9441	0.9387	0.9887
0.0087	29.3839	2325	0.0754	0.9368	0.9450	0.9409	0.9886
0.0092	29.5735	2340	0.0758	0.9374	0.9418	0.9396	0.9881
0.008	29.7630	2355	0.0776	0.9368	0.9445	0.9406	0.9884
0.0065	29.9526	2370	0.0781	0.9384	0.9427	0.9405	0.9883
0.0076	30.1422	2385	0.0776	0.9372	0.9445	0.9409	0.9886
0.0087	30.3318	2400	0.0768	0.9340	0.9418	0.9379	0.9883
0.0064	30.5213	2415	0.0766	0.9368	0.9445	0.9406	0.9883
0.0083	30.7109	2430	0.0779	0.9355	0.9385	0.9370	0.9881
0.0086	30.9005	2445	0.0756	0.9348	0.9409	0.9378	0.9879
0.0073	31.0900	2460	0.0753	0.9351	0.9450	0.9400	0.9884
0.0082	31.2796	2475	0.0763	0.9402	0.9441	0.9421	0.9884
0.0058	31.4692	2490	0.0759	0.9411	0.9455	0.9433	0.9890
0.0071	31.6588	2505	0.0766	0.9388	0.9432	0.9410	0.9886
0.008	31.8483	2520	0.0774	0.9366	0.9418	0.9392	0.9882
0.0084	32.0379	2535	0.0771	0.9390	0.9464	0.9427	0.9890
0.0058	32.2275	2550	0.0766	0.9400	0.9478	0.9439	0.9892
0.0079	32.4171	2565	0.0777	0.9354	0.9432	0.9393	0.9884
0.0066	32.6066	2580	0.0770	0.9394	0.9459	0.9427	0.9889
0.0068	32.7962	2595	0.0755	0.9364	0.9455	0.9409	0.9889
0.0071	32.9858	2610	0.0742	0.9358	0.9436	0.9397	0.9887
0.0105	33.1754	2625	0.0748	0.9341	0.9427	0.9384	0.9885
0.006	33.3649	2640	0.0753	0.9386	0.9464	0.9425	0.9889
0.0075	33.5545	2655	0.0748	0.9390	0.9455	0.9422	0.9889
0.0068	33.7441	2670	0.0757	0.9381	0.9455	0.9418	0.9889
0.0061	33.9336	2685	0.0757	0.9360	0.9455	0.9407	0.9888
0.0074	34.1232	2700	0.0753	0.9364	0.9450	0.9407	0.9885
0.006	34.3128	2715	0.0760	0.9362	0.9432	0.9397	0.9884
0.0064	34.5024	2730	0.0759	0.9364	0.9459	0.9411	0.9888
0.0055	34.6919	2745	0.0760	0.9368	0.9445	0.9406	0.9889
0.0063	34.8815	2760	0.0767	0.9375	0.9432	0.9403	0.9886
0.0065	35.0711	2775	0.0769	0.9385	0.9445	0.9415	0.9888
0.0075	35.2607	2790	0.0767	0.9407	0.9464	0.9436	0.9889
0.0067	35.4502	2805	0.0767	0.9399	0.9459	0.9429	0.9888
0.0058	35.6398	2820	0.0771	0.9403	0.9455	0.9429	0.9886
0.0044	35.8294	2835	0.0775	0.9406	0.9445	0.9426	0.9884
0.0062	36.0190	2850	0.0776	0.9398	0.9445	0.9422	0.9885
0.0067	36.2085	2865	0.0778	0.9411	0.9450	0.9430	0.9886
0.0073	36.3981	2880	0.0773	0.9408	0.9469	0.9438	0.9888
0.0057	36.5877	2895	0.0773	0.9412	0.9464	0.9438	0.9887
0.0062	36.7773	2910	0.0773	0.9398	0.9455	0.9426	0.9887
0.0056	36.9668	2925	0.0773	0.9390	0.9459	0.9424	0.9887
0.006	37.1564	2940	0.0772	0.9390	0.9464	0.9427	0.9889
0.007	37.3460	2955	0.0773	0.9394	0.9464	0.9429	0.9888
0.0066	37.5355	2970	0.0773	0.9399	0.9464	0.9431	0.9889
0.0066	37.7251	2985	0.0773	0.9399	0.9464	0.9431	0.9889
0.0067	37.9147	3000	0.0773	0.9403	0.9464	0.9433	0.9888

Framework versions

• Transformers 4.40.2
• Pytorch 2.4.1
• Datasets 2.19.0
• Tokenizers 0.19.1