mt0-base_lora_lr5e-05_bs4_epoch20_wd0.01

This model is a fine-tuned version of bigscience/mt0-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0002

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
0.5596	1.0	157	0.2422
0.2051	2.0	314	0.0247
0.0709	3.0	471	0.0035
0.0284	4.0	628	0.0020
0.023	5.0	785	0.0016
0.0157	6.0	942	0.0013
0.0187	7.0	1099	0.0008
0.0147	8.0	1256	0.0006
0.0138	9.0	1413	0.0005
0.0161	10.0	1570	0.0004
0.0074	11.0	1727	0.0004
0.0121	12.0	1884	0.0002
0.0063	13.0	2041	0.0002
0.0072	14.0	2198	0.0002
0.0069	15.0	2355	0.0002
0.0056	16.0	2512	0.0002
0.0112	17.0	2669	0.0002
0.008	18.0	2826	0.0002
0.0091	19.0	2983	0.0002
0.0065	20.0	3140	0.0002

Framework versions

• PEFT 0.7.1
• Transformers 4.36.2
• Pytorch 2.0.1
• Datasets 2.16.1
• Tokenizers 0.15.0