diff --git a/convlab/base_models/t5/README.md b/convlab/base_models/t5/README.md index 754a1db40474b485c78d5a7857ffef70b4658c24..4f915dc04098e54b715b2626ec20bdfc39400d50 100644 --- a/convlab/base_models/t5/README.md +++ b/convlab/base_models/t5/README.md @@ -43,10 +43,12 @@ Trained models and their performance are available in [Hugging Face Hub](https:/ | ------------------------------------------------------------ | ------------- | ---------------------------- | | [t5-small-goal2dialogue-multiwoz21](https://huggingface.co/ConvLab/t5-small-goal2dialogue-multiwoz21) | Goal2Dialogue | MultiWOZ 2.1 | | [t5-small-nlu-multiwoz21](https://huggingface.co/ConvLab/t5-small-nlu-multiwoz21) | NLU | MultiWOZ 2.1 | +| [t5-small-nlu-all-multiwoz21](https://huggingface.co/ConvLab/t5-small-nlu-all-multiwoz21) | NLU | MultiWOZ 2.1 all utterances | | [t5-small-nlu-sgd](https://huggingface.co/ConvLab/t5-small-nlu-sgd) | NLU | SGD | | [t5-small-nlu-tm1_tm2_tm3](https://huggingface.co/ConvLab/t5-small-nlu-tm1_tm2_tm3) | NLU | TM1+TM2+TM3 | | [t5-small-nlu-multiwoz21_sgd_tm1_tm2_tm3](https://huggingface.co/ConvLab/t5-small-nlu-multiwoz21_sgd_tm1_tm2_tm3) | NLU | MultiWOZ 2.1+SGD+TM1+TM2+TM3 | | [t5-small-nlu-multiwoz21-context3](https://huggingface.co/ConvLab/t5-small-nlu-multiwoz21-context3) | NLU (context=3) | MultiWOZ 2.1 | +| [t5-small-nlu-all-multiwoz21-context3](https://huggingface.co/ConvLab/t5-small-nlu-all-multiwoz21-context3) | NLU (context=3) | MultiWOZ 2.1 all utterances | | [t5-small-nlu-tm1-context3](https://huggingface.co/ConvLab/t5-small-nlu-tm1-context3) | NLU (context=3) | TM1 | | [t5-small-nlu-tm2-context3](https://huggingface.co/ConvLab/t5-small-nlu-tm2-context3) | NLU (context=3) | TM2 | | [t5-small-nlu-tm3-context3](https://huggingface.co/ConvLab/t5-small-nlu-tm3-context3) | NLU (context=3) | TM3 | @@ -55,6 +57,8 @@ Trained models and their performance are available in [Hugging Face Hub](https:/ | [t5-small-dst-tm1_tm2_tm3](https://huggingface.co/ConvLab/t5-small-dst-tm1_tm2_tm3) | DST | TM1+TM2+TM3 | | [t5-small-dst-multiwoz21_sgd_tm1_tm2_tm3](https://huggingface.co/ConvLab/t5-small-dst-multiwoz21_sgd_tm1_tm2_tm3) | DST | MultiWOZ 2.1+SGD+TM1+TM2+TM3 | | [t5-small-nlg-multiwoz21](https://huggingface.co/ConvLab/t5-small-nlg-multiwoz21) | NLG | MultiWOZ 2.1 | +| [t5-small-nlg-user-multiwoz21](https://huggingface.co/ConvLab/t5-small-nlg-user-multiwoz21) | NLG | MultiWOZ 2.1 user utterances | +| [t5-small-nlg-all-multiwoz21](https://huggingface.co/ConvLab/t5-small-nlg-all-multiwoz21) | NLG | MultiWOZ 2.1 all utterances | | [t5-small-nlg-sgd](https://huggingface.co/ConvLab/t5-small-nlg-sgd) | NLG | SGD | | [t5-small-nlg-tm1_tm2_tm3](https://huggingface.co/ConvLab/t5-small-nlg-tm1_tm2_tm3) | NLG | TM1+TM2+TM3 | | [t5-small-nlg-multiwoz21_sgd_tm1_tm2_tm3](https://huggingface.co/ConvLab/t5-small-nlg-multiwoz21_sgd_tm1_tm2_tm3) | NLG | MultiWOZ 2.1+SGD+TM1+TM2+TM3 | diff --git a/convlab/nlu/jointBERT/README.md b/convlab/nlu/jointBERT/README.md index 647007548f18ad5c3adf97565acb7897ab025c5f..80f18a837e1056a7326b9b49d486f9c2e676e55b 100755 --- a/convlab/nlu/jointBERT/README.md +++ b/convlab/nlu/jointBERT/README.md @@ -40,6 +40,7 @@ To illustrate that it is easy to use the model for any dataset that in our unifi <tr> <th></th> <th colspan=2>MultiWOZ 2.1</th> + <th colspan=2>MultiWOZ 2.1 all utterances</th> <th colspan=2>Taskmaster-1</th> <th colspan=2>Taskmaster-2</th> <th colspan=2>Taskmaster-3</th> @@ -52,12 +53,14 @@ To illustrate that it is easy to use the model for any dataset that in our unifi <th>Acc</th><th>F1</th> <th>Acc</th><th>F1</th> <th>Acc</th><th>F1</th> + <th>Acc</th><th>F1</th> </tr> </thead> <tbody> <tr> <td>BERTNLU</td> <td>74.5</td><td>85.9</td> + <td>59.5</td><td>80.0</td> <td>72.8</td><td>50.6</td> <td>79.2</td><td>70.6</td> <td>86.1</td><td>81.9</td> @@ -65,6 +68,7 @@ To illustrate that it is easy to use the model for any dataset that in our unifi <tr> <td>BERTNLU (context=3)</td> <td>80.6</td><td>90.3</td> + <td>58.1</td><td>79.6</td> <td>74.2</td><td>52.7</td> <td>80.9</td><td>73.3</td> <td>87.8</td><td>83.8</td> diff --git a/convlab/nlu/jointBERT/unified_datasets/configs/multiwoz21_all.json b/convlab/nlu/jointBERT/unified_datasets/configs/multiwoz21_all.json new file mode 100644 index 0000000000000000000000000000000000000000..b996324cf5d6195290b7aa59cd7745ad6851abb8 --- /dev/null +++ b/convlab/nlu/jointBERT/unified_datasets/configs/multiwoz21_all.json @@ -0,0 +1,27 @@ +{ + "dataset_name": "multiwoz21", + "data_dir": "unified_datasets/data/multiwoz21/all/context_window_size_0", + "output_dir": "unified_datasets/output/multiwoz21/all/context_window_size_0", + "zipped_model_path": "unified_datasets/output/multiwoz21/all/context_window_size_0/bertnlu_unified_multiwoz21_all_context0.zip", + "log_dir": "unified_datasets/output/multiwoz21/all/context_window_size_0/log", + "DEVICE": "cuda:0", + "seed": 2019, + "cut_sen_len": 40, + "use_bert_tokenizer": true, + "context_window_size": 0, + "model": { + "finetune": true, + "context": false, + "context_grad": false, + "pretrained_weights": "bert-base-uncased", + "check_step": 1000, + "max_step": 10000, + "batch_size": 128, + "learning_rate": 1e-4, + "adam_epsilon": 1e-8, + "warmup_steps": 0, + "weight_decay": 0.0, + "dropout": 0.1, + "hidden_units": 768 + } + } \ No newline at end of file diff --git a/convlab/nlu/jointBERT/unified_datasets/configs/multiwoz21_all_context3.json b/convlab/nlu/jointBERT/unified_datasets/configs/multiwoz21_all_context3.json new file mode 100644 index 0000000000000000000000000000000000000000..b23ef9d63900caf39b9dea548754e310f7577a93 --- /dev/null +++ b/convlab/nlu/jointBERT/unified_datasets/configs/multiwoz21_all_context3.json @@ -0,0 +1,27 @@ +{ + "dataset_name": "multiwoz21", + "data_dir": "unified_datasets/data/multiwoz21/all/context_window_size_3", + "output_dir": "unified_datasets/output/multiwoz21/all/context_window_size_3", + "zipped_model_path": "unified_datasets/output/multiwoz21/all/context_window_size_3/bertnlu_unified_multiwoz21_all_context3.zip", + "log_dir": "unified_datasets/output/multiwoz21/all/context_window_size_3/log", + "DEVICE": "cuda:0", + "seed": 2019, + "cut_sen_len": 40, + "use_bert_tokenizer": true, + "context_window_size": 3, + "model": { + "finetune": true, + "context": true, + "context_grad": true, + "pretrained_weights": "bert-base-uncased", + "check_step": 1000, + "max_step": 10000, + "batch_size": 128, + "learning_rate": 1e-4, + "adam_epsilon": 1e-8, + "warmup_steps": 0, + "weight_decay": 0.0, + "dropout": 0.1, + "hidden_units": 1536 + } + } \ No newline at end of file