kugler/xlmr-large-AmDi-synset-classifier

xlmr_synset_classifier

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

• Loss: 0.5821
• Accuracy: 0.8300
• F1: 0.8189
• Precision: 0.8299
• Recall: 0.8300
• F1 Macro: 0.6291
• Precision Macro: 0.6111
• Recall Macro: 0.6637
• F1 Micro: 0.8300
• Precision Micro: 0.8300
• Recall Micro: 0.8300

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	F1 Macro	Precision Macro	Recall Macro	F1 Micro	Precision Micro	Recall Micro
3.6302	0.6221	100	2.3997	0.4612	0.3541	0.3218	0.4612	0.1136	0.1245	0.1308	0.4612	0.4612	0.4612
1.6212	1.2442	200	0.9750	0.7479	0.7052	0.7046	0.7479	0.4218	0.4214	0.4650	0.7479	0.7479	0.7479
0.9307	1.8663	300	0.7650	0.7936	0.7685	0.7863	0.7936	0.5217	0.5204	0.5619	0.7936	0.7936	0.7936
0.6977	2.4883	400	0.6956	0.8089	0.7935	0.8090	0.8089	0.5696	0.5599	0.6015	0.8089	0.8089	0.8089
0.6152	3.1104	500	0.6451	0.8188	0.8051	0.8224	0.8188	0.6021	0.5949	0.6321	0.8188	0.8188	0.8188
0.5171	3.7325	600	0.5960	0.8331	0.8209	0.8322	0.8331	0.6287	0.6304	0.6524	0.8331	0.8331	0.8331
0.4772	4.3546	700	0.5903	0.8286	0.8178	0.8291	0.8286	0.6305	0.6244	0.6587	0.8286	0.8286	0.8286
0.437	4.9767	800	0.5821	0.8300	0.8189	0.8299	0.8300	0.6291	0.6111	0.6637	0.8300	0.8300	0.8300

Framework versions

• Transformers 4.45.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.20.3