metadata

license: mit
base_model: xlm-roberta-base
tags:
  - generated_from_trainer
datasets:
  - tweet_sentiment_multilingual
metrics:
  - accuracy
  - f1
model-index:
  - name: >-
      scenario-NON-KD-PR-COPY-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: tweet_sentiment_multilingual
          type: tweet_sentiment_multilingual
          config: all
          split: validation
          args: all
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5767746913580247
          - name: F1
            type: f1
            value: 0.5751836259585372

scenario-NON-KD-PR-COPY-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_

This model is a fine-tuned version of xlm-roberta-base on the tweet_sentiment_multilingual dataset. It achieves the following results on the evaluation set:

Loss: 4.1676
Accuracy: 0.5768
F1: 0.5752

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: 5e-05
train_batch_size: 32
eval_batch_size: 32
seed: 1123
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.0487	1.09	500	0.9666	0.5305	0.5194
0.9092	2.17	1000	0.9220	0.5760	0.5733
0.76	3.26	1500	1.0464	0.5791	0.5681
0.6233	4.35	2000	1.1732	0.5864	0.5809
0.4852	5.43	2500	1.1695	0.5590	0.5578
0.374	6.52	3000	1.3903	0.5691	0.5669
0.2851	7.61	3500	1.5832	0.5760	0.5711
0.2203	8.7	4000	1.6098	0.5737	0.5739
0.1863	9.78	4500	1.9189	0.5656	0.5566
0.1437	10.87	5000	2.1445	0.5826	0.5783
0.1302	11.96	5500	1.9960	0.5791	0.5723
0.1075	13.04	6000	2.5978	0.5710	0.5663
0.0957	14.13	6500	2.9129	0.5675	0.5682
0.0898	15.22	7000	2.8487	0.5799	0.5780
0.082	16.3	7500	2.8461	0.5714	0.5621
0.0673	17.39	8000	2.8416	0.5849	0.5767
0.0647	18.48	8500	3.1083	0.5849	0.5810
0.0597	19.57	9000	2.9063	0.5772	0.5700
0.0508	20.65	9500	3.1069	0.5706	0.5663
0.0492	21.74	10000	3.1434	0.5841	0.5853
0.0485	22.83	10500	2.9341	0.5887	0.5816
0.0373	23.91	11000	3.2828	0.5810	0.5807
0.0352	25.0	11500	3.1742	0.5864	0.5802
0.0326	26.09	12000	3.2767	0.5733	0.5734
0.0269	27.17	12500	3.5101	0.5826	0.5797
0.0338	28.26	13000	3.2453	0.5725	0.5693
0.0289	29.35	13500	3.3957	0.5694	0.5703
0.0232	30.43	14000	3.4668	0.5710	0.5714
0.0215	31.52	14500	3.5250	0.5721	0.5660
0.0197	32.61	15000	3.5990	0.5787	0.5755
0.0138	33.7	15500	3.7731	0.5745	0.5682
0.0177	34.78	16000	3.6367	0.5698	0.5671
0.0145	35.87	16500	3.8987	0.5725	0.5705
0.013	36.96	17000	3.8459	0.5745	0.5737
0.0133	38.04	17500	3.7106	0.5733	0.5711
0.0095	39.13	18000	3.8834	0.5683	0.5688
0.0091	40.22	18500	3.9118	0.5733	0.5731
0.0107	41.3	19000	3.9038	0.5768	0.5733
0.0089	42.39	19500	3.8957	0.5826	0.5784
0.0042	43.48	20000	4.1050	0.5775	0.5761
0.0067	44.57	20500	4.0982	0.5756	0.5739
0.0042	45.65	21000	4.2051	0.5737	0.5733
0.0057	46.74	21500	4.1266	0.5764	0.5764
0.0056	47.83	22000	4.1318	0.5787	0.5765
0.0034	48.91	22500	4.1443	0.5791	0.5772
0.003	50.0	23000	4.1676	0.5768	0.5752

Framework versions

Transformers 4.33.3
Pytorch 2.1.1+cu121
Datasets 2.14.5
Tokenizers 0.13.3