opt-350m-finetuned-v4-seinfeld

This model is a fine-tuned version of facebook/opt-350m on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.4775

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: 0.0005
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 1000
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.8604	0.79	8	2.7539
2.8058	1.59	16	2.6860
2.7427	2.4	24	2.6338
2.6593	3.2	32	2.5925
2.5373	3.99	40	2.5601
2.4923	4.79	48	2.5370
2.4102	5.59	56	2.5216
2.3373	6.4	64	2.5049
2.2341	7.2	72	2.4963
2.1286	7.99	80	2.4862
2.0673	8.79	88	2.4908
1.9938	9.59	96	2.4881
1.9015	10.4	104	2.4854
1.8172	11.2	112	2.5058
1.7113	11.99	120	2.4950
1.6409	12.79	128	2.5082
1.5622	13.59	136	2.5172
1.4724	14.4	144	2.5464

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.0
• Tokenizers 0.13.2