convnext-Mistral-UCM-without-captioning

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

• Loss: 0.9360
• Accuracy: 74.37
• Bleu-1: 0.8400
• Bleu-2: 0.7835
• Bleu-3: 0.7338
• Bleu-4: 0.6880
• Meteor: 0.8321
• Rouge-l: 0.8115
• Cider: 3.4751

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: 64
• eval_batch_size: 64
• seed: 50
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 128
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Bleu-1	Bleu-2	Bleu-3	Bleu-4	Meteor	Rouge-l	Cider
No log	1.0	148	3.7308	67.17	0.2828	0.1477	0.0605	0.0110	0.1410	0.2399	0.1455
No log	2.0	296	0.7857	71.2	0.4427	0.3414	0.2744	0.2349	0.3804	0.4072	1.0593
No log	3.0	444	0.7139	72.56	0.7159	0.6454	0.5937	0.5494	0.6784	0.6810	2.7803
No log	4.0	592	0.6261	73.43	0.8397	0.7773	0.7273	0.6835	0.7990	0.7981	3.3691
No log	5.0	740	0.6593	74.62	0.8210	0.7559	0.7032	0.6555	0.7955	0.7866	3.3208
No log	6.0	888	0.6611	73.98	0.8251	0.7684	0.7196	0.6746	0.7854	0.7822	3.2551
1.2841	7.0	1036	0.6809	73.58	0.7926	0.7311	0.6776	0.6305	0.7668	0.7663	3.0771
1.2841	8.0	1184	0.7132	73.6	0.8198	0.7561	0.7018	0.6517	0.7963	0.7896	3.0844
1.2841	9.0	1332	0.7024	74.59	0.8405	0.7817	0.7306	0.6843	0.8177	0.8122	3.3312
1.2841	10.0	1480	0.7562	74.24	0.8147	0.7453	0.6906	0.6439	0.7701	0.7661	3.1819
1.2841	11.0	1628	0.7607	74.98	0.8471	0.7915	0.7433	0.6984	0.8193	0.8080	3.4831
1.2841	12.0	1776	0.7484	74.84	0.8430	0.7850	0.7381	0.6959	0.7950	0.7892	3.3640
1.2841	13.0	1924	0.7635	73.59	0.8225	0.7649	0.7149	0.6682	0.7842	0.7793	3.2893
0.3441	14.0	2072	0.7843	74.43	0.8383	0.7829	0.7317	0.6845	0.8164	0.7968	3.4086
0.3441	15.0	2220	0.8008	75.45	0.8517	0.7954	0.7478	0.7037	0.8134	0.8073	3.5558
0.3441	16.0	2368	0.8012	74.65	0.8514	0.7977	0.7521	0.7104	0.8309	0.8175	3.5537
0.3441	17.0	2516	0.8180	74.71	0.8452	0.7905	0.7441	0.7023	0.8251	0.8145	3.4183
0.3441	18.0	2664	0.8264	73.75	0.8313	0.7754	0.7287	0.6854	0.8167	0.7972	3.3675
0.3441	19.0	2812	0.8142	74.59	0.8554	0.7962	0.7439	0.6961	0.8260	0.8195	3.4548
0.3441	20.0	2960	0.8162	74.57	0.8278	0.7720	0.7226	0.6768	0.8161	0.8014	3.3411
0.2844	21.0	3108	0.8308	75.14	0.8475	0.7907	0.7446	0.7025	0.8148	0.8100	3.4721
0.2844	22.0	3256	0.8372	74.47	0.8238	0.7693	0.7216	0.6770	0.8134	0.7944	3.4387
0.2844	23.0	3404	0.8560	74.22	0.8450	0.7940	0.7485	0.7066	0.8261	0.8121	3.4413
0.2844	24.0	3552	0.8674	74.94	0.8389	0.7857	0.7406	0.7002	0.8100	0.7977	3.4734
0.2844	25.0	3700	0.9104	75.38	0.8633	0.8121	0.7673	0.7261	0.8327	0.8237	3.5823
0.2844	26.0	3848	0.9113	74.18	0.8157	0.7531	0.7024	0.6581	0.7860	0.7735	3.2752
0.2844	27.0	3996	0.8788	74.72	0.8489	0.7904	0.7422	0.6976	0.8193	0.8086	3.5050
0.2659	28.0	4144	0.9124	74.57	0.8167	0.7512	0.6996	0.6537	0.7932	0.7802	3.3738
0.2659	29.0	4292	0.9147	74.78	0.8355	0.7803	0.7343	0.6912	0.8299	0.8106	3.5364
0.2659	30.0	4440	0.8889	74.26	0.8434	0.7821	0.7271	0.6755	0.8201	0.7946	3.3550
0.2659	31.0	4588	0.9030	74.2	0.8248	0.7662	0.7172	0.6720	0.8015	0.7843	3.3138
0.2659	32.0	4736	0.9195	74.02	0.8264	0.7666	0.7163	0.6720	0.7912	0.7861	3.3463
0.2659	33.0	4884	0.8961	74.67	0.8408	0.7814	0.7312	0.6867	0.8244	0.8037	3.4254
0.2659	34.0	5032	0.8996	74.84	0.8527	0.7963	0.7498	0.7061	0.8313	0.8125	3.5791
0.2593	35.0	5180	0.9360	74.37	0.8400	0.7835	0.7338	0.6880	0.8321	0.8115	3.4751

Framework versions

• Transformers 5.12.1
• Pytorch 2.12.1+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2