hbertv1-tiny-wt-frz-48-Massive-intent-emb-comp

This model is a fine-tuned version of gokuls/model_v1_complete_training_wt_init_48_tiny_emb_comp_frz on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.8947
• Accuracy: 0.7787

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: 64
• eval_batch_size: 64
• seed: 33
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.6262	1.0	180	3.0736	0.2892
2.6684	2.0	360	2.2654	0.4206
2.0129	3.0	540	1.7714	0.5396
1.5846	4.0	720	1.4366	0.6522
1.2957	5.0	900	1.2535	0.6950
1.0996	6.0	1080	1.1380	0.7098
0.9632	7.0	1260	1.0479	0.7334
0.861	8.0	1440	1.0077	0.7570
0.7925	9.0	1620	0.9793	0.7664
0.7277	10.0	1800	0.9500	0.7664
0.6832	11.0	1980	0.9333	0.7683
0.6476	12.0	2160	0.9121	0.7737
0.6167	13.0	2340	0.9049	0.7767
0.5925	14.0	2520	0.9038	0.7767
0.5797	15.0	2700	0.8947	0.7787

Framework versions

• Transformers 4.31.0
• Pytorch 1.14.0a0+410ce96
• Datasets 2.13.1
• Tokenizers 0.13.3