flan-t5-base-da-multiwoz2.0_800-loss-ep100

This model is a fine-tuned version of google/flan-t5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3457
• Accuracy: 41.0835
• Num: 7358
• Gen Len: 15.8615

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: 2e-05
• train_batch_size: 60
• eval_batch_size: 400
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Num	Gen Len
1.3234	2.22	200	0.5277	27.6507	7358	15.1567
0.5832	4.44	400	0.4297	33.2156	7358	16.0043
0.4991	6.67	600	0.4029	35.309	7358	16.0595
0.4618	8.89	800	0.3845	36.2347	7358	15.9841
0.4372	11.11	1000	0.3750	36.6748	7358	15.6254
0.4207	13.33	1200	0.3664	37.7404	7358	16.0586
0.4036	15.56	1400	0.3632	38.5822	7358	15.8968
0.3929	17.78	1600	0.3588	38.4617	7358	15.1256
0.3809	20.0	1800	0.3554	39.6862	7358	15.7427
0.3718	22.22	2000	0.3528	40.097	7358	15.9545
0.3658	24.44	2200	0.3517	39.9142	7358	15.3572
0.3581	26.67	2400	0.3486	40.6066	7358	15.7185
0.3513	28.89	2600	0.3476	40.6688	7358	16.0911
0.3449	31.11	2800	0.3469	41.0736	7358	15.7513
0.3395	33.33	3000	0.3457	41.0835	7358	15.8615
0.3318	35.56	3200	0.3486	41.0794	7358	15.9333
0.3283	37.78	3400	0.3474	41.5949	7358	16.0735
0.3229	40.0	3600	0.3476	41.8376	7358	15.7744
0.3184	42.22	3800	0.3474	41.6082	7358	16.0443
0.313	44.44	4000	0.3472	41.9014	7358	16.0557

Framework versions

• Transformers 4.18.0
• Pytorch 1.10.0+cu111
• Datasets 2.5.1
• Tokenizers 0.12.1