added files

.gitignore

@@ -1 +1 @@
-.idea/
+.idea/*

README.md

@@ -1,5 +1,5 @@
 # DialoKG: Knowledge-Structure Aware Task-Oriented Dialogue Generation
-PyTorch code for NAACL 2022 paper: DialoKG: Knowledge-Structure Aware Task-Oriented Dialogue Generation [[PDF]](https://aclanthology.org/2022.findings-naacl.195.pdf).
+This repository contains PyTorch code for NAACL 2022 paper: DialoKG: Knowledge-Structure Aware Task-Oriented Dialogue Generation [[PDF]](https://aclanthology.org/2022.findings-naacl.195.pdf).
 
 [![Python 3.7](https://img.shields.io/badge/python-3.7-blue.svg)](https://www.python.org/downloads/release/python-370/)
 [![PyTorch](https://img.shields.io/badge/PyTorch-%23EE4C2C.svg?style=flat&logo=PyTorch&logoColor=white)](https://pytorch.org/)
@@ -14,8 +14,18 @@ chmod +x setup.sh
 ./setup.sh
 ```
 
+### 🏋️ Training
+```shell
+python train.py --dataset <DATASET-NAME> --params_file config/gpt2/params.json --device cuda
+```
+Valid dataset names: **incar**, **camrest**, **woz2.1**
+
+### 🎯 Evaluation
+```shell
+
+```
 
-### Citation
+### 📝 Citation
 ```
 @inproceedings{rony-etal-2022-dialokg,
     title = "{D}ialo{KG}: Knowledge-Structure Aware Task-Oriented Dialogue Generation",
@@ -32,8 +42,8 @@ chmod +x setup.sh
     abstract = "Task-oriented dialogue generation is challenging since the underlying knowledge is often dynamic and effectively incorporating knowledge into the learning process is hard. It is particularly challenging to generate both human-like and informative responses in this setting. Recent research primarily focused on various knowledge distillation methods where the underlying relationship between the facts in a knowledge base is not effectively captured. In this paper, we go one step further and demonstrate how the structural information of a knowledge graph can improve the system{'}s inference capabilities. Specifically, we propose DialoKG, a novel task-oriented dialogue system that effectively incorporates knowledge into a language model. Our proposed system views relational knowledge as a knowledge graph and introduces (1) a structure-aware knowledge embedding technique, and (2) a knowledge graph-weighted attention masking strategy to facilitate the system selecting relevant information during the dialogue generation. An empirical evaluation demonstrates the effectiveness of DialoKG over state-of-the-art methods on several standard benchmark datasets.",
 }
 ```
-### License
+### 📜 License
 [MIT]()
 
-### Contact
+### 📪 Contact
 For further information, contact the corresponding author Md Rashad Al Hasan Rony ([email](mailto:[email protected])).

config/gpt2/generation_params.json

@@ -0,0 +1,8 @@
+{
+  "no_sample": false,
+  "min_length": 1,
+  "max_length": 120,
+  "temperature": 0.18,
+  "top_k": 10,
+  "top_p": 0.9
+}

config/gpt2/params.json

@@ -0,0 +1,19 @@
+{
+    "dataset_args": {
+      "history_max_utterances": 3,
+      "history_max_tokens": 128,
+      "knowledge_max_tokens": 128
+    },
+    "task": "generation",
+    "model_name_or_path": "gpt2",
+    "per_gpu_train_batch_size": 4,
+    "per_gpu_eval_batch_size": 4,
+    "gradient_accumulation_steps": 4,
+    "learning_rate": 6.25e-5,
+    "adam_epsilon": 1e-8,
+    "max_grad_norm": 1,
+    "num_train_epochs": 1,
+    "warmup_steps": 0,
+    "fp16": "",
+    "seed": 42
+}