scenario-kd-from-post-finetune-gold-silver-div-2-data-tyqiangz-multilingual-sent

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

• Loss: 3.7936
• Accuracy: 0.2685
• F1: 0.2380

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
3.0493	17.24	1000	3.7325	0.2809	0.2293
2.9173	34.48	2000	3.7722	0.2932	0.2521
2.8726	51.72	3000	3.7327	0.2840	0.2380
2.8515	68.97	4000	3.7081	0.2685	0.2381
2.8311	86.21	5000	3.7213	0.2716	0.2371
2.8245	103.45	6000	3.7201	0.2778	0.2390
2.8144	120.69	7000	3.7450	0.2778	0.2483
2.8063	137.93	8000	3.7183	0.2809	0.2386
2.804	155.17	9000	3.6985	0.2870	0.2517
2.7968	172.41	10000	3.6685	0.2840	0.2535
2.7939	189.66	11000	3.6682	0.2809	0.2443
2.793	206.9	12000	3.7012	0.2840	0.2439
2.7889	224.14	13000	3.7035	0.2654	0.2425
2.7832	241.38	14000	3.7338	0.2809	0.2515
2.7847	258.62	15000	3.7130	0.2840	0.2487
2.7789	275.86	16000	3.6999	0.2685	0.2353
2.7706	293.1	17000	3.6997	0.2809	0.2545
2.7781	310.34	18000	3.7201	0.2809	0.2514
2.7721	327.59	19000	3.6860	0.2778	0.2496
2.7742	344.83	20000	3.6770	0.2809	0.2487
2.771	362.07	21000	3.7196	0.2840	0.2571
2.7716	379.31	22000	3.7462	0.2654	0.2424
2.7616	396.55	23000	3.7363	0.2778	0.2467
2.7621	413.79	24000	3.7720	0.2716	0.2495
2.7635	431.03	25000	3.7128	0.2685	0.2409
2.77	448.28	26000	3.7065	0.2716	0.2380
2.7606	465.52	27000	3.7269	0.2840	0.2410
2.7613	482.76	28000	3.7081	0.2870	0.2489
2.762	500.0	29000	3.7697	0.2778	0.2321
2.7589	517.24	30000	3.7486	0.2778	0.2438
2.7556	534.48	31000	3.7936	0.2685	0.2380

Framework versions

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