File size: 1,831 Bytes
eb514ec 7fe2663 eb514ec 7fe2663 eb514ec 7fe2663 76de0cc 7fe2663 eb514ec 7fe2663 eb514ec 7fe2663 eb514ec 7fe2663 eb514ec 76de0cc eb514ec 7fe2663 eb514ec 7fe2663 eb514ec |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 |
---
language:
- en
license: apache-2.0
tags:
- t5-small
- text2text-generation
- natural language generation
- conversational system
- task-oriented dialog
datasets:
- ConvLab/tm1
- ConvLab/tm2
- ConvLab/tm3
metrics:
- Slot Error Rate
- sacrebleu
model-index:
- name: t5-small-nlg-tm1_tm2_tm3
results:
- task:
type: text2text-generation
name: natural language generation
dataset:
type: ConvLab/tm1, ConvLab/tm2, ConvLab/tm3
name: TM1+TM2+TM3
split: test
metrics:
- type: Slot Error Rate
value: 2.1
name: SER
- type: sacrebleu
value: 51.5
name: BLEU
widget:
- text: "tm1: [inform][pizza_ordering]([name.store][Domino's])\n\nsystem: "
- text: "tm2: [inform][restaurant-search]([name.restaurant][Via 313, the Violet Crown Social Club],[price_range][$8 per slice])\n\nsystem: "
- text: "tm3: [inform][movie]([name.movie][Star Wars],[name.movie][The Grudge])\n\nsystem: "
inference:
parameters:
max_length: 100
---
# t5-small-nlg-tm1_tm2_tm3
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on [Taskmaster-1](https://huggingface.co/datasets/ConvLab/tm1), [Taskmaster-2](https://huggingface.co/datasets/ConvLab/tm2), and [Taskmaster-3](https://huggingface.co/datasets/ConvLab/tm3).
Refer to [ConvLab-3](https://github.com/ConvLab/ConvLab-3) for model description and usage.
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Adafactor
- lr_scheduler_type: linear
- num_epochs: 10.0
### Framework versions
- Transformers 4.18.0
- Pytorch 1.10.2+cu102
- Datasets 1.18.3
- Tokenizers 0.11.0
|