README.md

metadata

library_name: transformers
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: swin-resnet-mistral-SYDNEY-with-all-captioning
    results: []

swin-resnet-mistral-SYDNEY-with-all-captioning

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

• Loss: 1.3232
• Accuracy: 65.7
• Bleu-1: 0.5949
• Bleu-2: 0.5339
• Bleu-3: 0.4914
• Bleu-4: 0.4557
• Meteor: 0.5436
• Rouge-l: 0.6098
• Cider: 1.6017

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: 1024
• 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	44	1.2803	40.4	0.3525	0.2768	0.2257	0.1881	0.5016	0.4202	0.5907
No log	2.0	88	1.0514	61.77	0.4774	0.3355	0.2522	0.1907	0.4166	0.4002	0.4287
No log	3.0	132	0.9772	63.06	0.4879	0.3441	0.2625	0.2029	0.4153	0.4023	0.4082
No log	4.0	176	0.9050	63.2	0.4597	0.3120	0.2307	0.1715	0.3604	0.3639	0.3364
No log	5.0	220	0.8187	63.37	0.5164	0.3760	0.2936	0.2326	0.4161	0.4183	0.5877
No log	6.0	264	0.7221	64.99	0.4858	0.3419	0.2627	0.2076	0.3554	0.3831	0.6604
No log	7.0	308	0.6234	65.16	0.5295	0.3903	0.3082	0.2448	0.4427	0.4342	0.7510
No log	8.0	352	0.5437	65.47	0.5267	0.3811	0.2972	0.2365	0.4509	0.4387	0.7940
No log	9.0	396	0.5210	66.25	0.5267	0.4112	0.3427	0.2956	0.4322	0.4488	1.1293
No log	10.0	440	0.5277	66.31	0.6364	0.5407	0.4771	0.4297	0.5541	0.5557	1.8319
No log	11.0	484	0.5085	65.06	0.6104	0.5088	0.4397	0.3882	0.5494	0.5520	1.7389
No log	12.0	528	0.5123	66.97	0.6496	0.5495	0.4797	0.4301	0.5773	0.5768	1.7341
No log	13.0	572	0.5340	66.23	0.4950	0.3817	0.3181	0.2718	0.4101	0.4507	1.2149
No log	14.0	616	0.5329	65.39	0.6253	0.5224	0.4452	0.3868	0.5576	0.5502	1.5926
No log	15.0	660	0.5461	65.92	0.6656	0.5754	0.5075	0.4546	0.5780	0.5894	1.8762
No log	16.0	704	0.5435	64.68	0.6365	0.5344	0.4565	0.3999	0.5685	0.5655	1.8068
No log	17.0	748	0.5619	66.19	0.6833	0.5911	0.5134	0.4465	0.5917	0.6082	1.7530
No log	18.0	792	0.5653	67.21	0.6432	0.5915	0.5493	0.5167	0.6103	0.6437	2.0025
No log	19.0	836	0.5855	63.68	0.6954	0.5975	0.5215	0.4622	0.6169	0.6120	1.9900
No log	20.0	880	0.6408	65.66	0.6691	0.5775	0.5106	0.4595	0.6005	0.6201	1.6515
No log	21.0	924	0.6872	67.74	0.6715	0.5886	0.5357	0.4988	0.5834	0.6151	1.9363
No log	22.0	968	0.6886	67.71	0.6965	0.6232	0.5719	0.5328	0.6193	0.6512	1.9162
No log	23.0	1012	0.7542	68.1	0.6502	0.5819	0.5336	0.4944	0.5734	0.6080	1.7041
0.6311	24.0	1056	0.8377	68.45	0.6886	0.6151	0.5662	0.5214	0.5968	0.6452	1.8615
0.6311	25.0	1100	1.1727	66.68	0.6665	0.5867	0.5296	0.4833	0.5548	0.6124	1.4923
0.6311	26.0	1144	1.2276	65.85	0.6264	0.5559	0.5134	0.4719	0.5668	0.6141	1.6275
0.6311	27.0	1188	1.3551	66.24	0.5980	0.5307	0.4856	0.4470	0.5345	0.6031	1.4574
0.6311	28.0	1232	1.2643	67.33	0.6410	0.5789	0.5327	0.4950	0.5766	0.6416	1.6530
0.6311	29.0	1276	1.4213	65.98	0.4811	0.3962	0.3426	0.2991	0.4590	0.5586	0.9854
0.6311	30.0	1320	1.3364	65.73	0.5691	0.4999	0.4555	0.4207	0.5231	0.5991	1.3969
0.6311	31.0	1364	1.3737	65.49	0.5799	0.5158	0.4759	0.4416	0.5276	0.6097	1.4832
0.6311	32.0	1408	1.3232	65.7	0.5949	0.5339	0.4914	0.4557	0.5436	0.6098	1.6017

Framework versions

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