scenario-kd-from-post-finetune-gold-silver-div-2-data-tweet_eval-sentiment-model

This model is a fine-tuned version of haryoaw/teacher_tweet_eval_sentiment_xlmr_base on the tweet_eval dataset. It achieves the following results on the evaluation set:

• Loss: 1.3344
• Accuracy: 0.734
• F1: 0.7061

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: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6969

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.838	0.7	1000	1.8178	0.71	0.6779
1.3953	1.4	2000	1.5675	0.72	0.6988
1.2818	2.1	3000	1.6086	0.7205	0.6892
1.1321	2.81	4000	1.6146	0.7235	0.7013
0.9348	3.51	5000	1.6493	0.719	0.6902
0.9139	4.21	6000	1.5044	0.719	0.7026
0.8836	4.91	7000	1.4467	0.728	0.7068
0.7857	5.61	8000	1.4894	0.727	0.6971
0.7366	6.31	9000	1.4895	0.7265	0.7052
0.7471	7.01	10000	1.4055	0.73	0.7089
0.7067	7.71	11000	1.4671	0.729	0.7055
0.6668	8.42	12000	1.4966	0.729	0.7011
0.6608	9.12	13000	1.4326	0.722	0.7006
0.6371	9.82	14000	1.4281	0.724	0.7005
0.6047	10.52	15000	1.4260	0.7225	0.6954
0.5929	11.22	16000	1.4013	0.7235	0.6974
0.5963	11.92	17000	1.3812	0.727	0.7053
0.5777	12.62	18000	1.3694	0.727	0.7029
0.5587	13.32	19000	1.4384	0.724	0.6951
0.5631	14.03	20000	1.3488	0.732	0.7097
0.538	14.73	21000	1.3758	0.7285	0.7042
0.5365	15.43	22000	1.3314	0.7305	0.7059
0.5339	16.13	23000	1.3576	0.729	0.7046
0.5232	16.83	24000	1.3652	0.732	0.7065
0.5082	17.53	25000	1.4092	0.729	0.7015
0.5015	18.23	26000	1.3548	0.726	0.7019
0.5035	18.93	27000	1.3637	0.7335	0.7076
0.4932	19.64	28000	1.3489	0.7205	0.6968
0.4867	20.34	29000	1.3753	0.7385	0.7135
0.4819	21.04	30000	1.3652	0.727	0.7050
0.4782	21.74	31000	1.3339	0.743	0.7194
0.4712	22.44	32000	1.3229	0.728	0.7059
0.4677	23.14	33000	1.3203	0.7365	0.7114
0.4736	23.84	34000	1.3605	0.7345	0.7114
0.4551	24.54	35000	1.3491	0.724	0.7016
0.458	25.25	36000	1.3663	0.739	0.7063
0.4545	25.95	37000	1.3502	0.734	0.7087
0.4471	26.65	38000	1.3435	0.7375	0.7100
0.441	27.35	39000	1.3069	0.7435	0.7155
0.4495	28.05	40000	1.3175	0.7285	0.7037
0.4411	28.75	41000	1.3344	0.734	0.7061

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1
• Datasets 2.14.5
• Tokenizers 0.13.3