DialoGPT-large-faqs-block-size-128-bs-16-lr-1e-05-deepspeed-stage2

This model is a fine-tuned version of microsoft/DialoGPT-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.4123

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	40	4.2793
No log	2.0	80	3.5752
No log	3.0	120	3.1238
No log	4.0	160	2.8875
No log	5.0	200	2.7358
No log	6.0	240	2.6321
No log	7.0	280	2.5629
No log	8.0	320	2.5147
No log	9.0	360	2.4783
No log	10.0	400	2.4595
No log	11.0	440	2.4370
No log	12.0	480	2.4229
2.7646	13.0	520	2.4167
2.7646	14.0	560	2.4109
2.7646	15.0	600	2.4084
2.7646	16.0	640	2.4146
2.7646	17.0	680	2.4085
2.7646	18.0	720	2.4139
2.7646	19.0	760	2.4137
2.7646	20.0	800	2.4123

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4.dev0
• Tokenizers 0.13.3