DILAB-HYU
/

DialRet

@@ -1,199 +1,153 @@
 ---
 library_name: transformers
-tags: []
 ---
-# Model Card for Model ID
-<!-- Provide a quick summary of what the model is/does. -->
-## Model Details
-### Model Description
-<!-- Provide a longer summary of what this model is. -->
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-### Model Sources [optional]
-<!-- Provide the basic links for the model. -->
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
-## Uses
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-### Direct Use
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-[More Information Needed]
-### Downstream Use [optional]
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-[More Information Needed]
-### Out-of-Scope Use
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-[More Information Needed]
-## Bias, Risks, and Limitations
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-[More Information Needed]
-### Recommendations
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
-## How to Get Started with the Model
-Use the code below to get started with the model.
-[More Information Needed]
-## Training Details
-### Training Data
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-[More Information Needed]
-### Training Procedure
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-#### Preprocessing [optional]
-[More Information Needed]
-#### Training Hyperparameters
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-#### Speeds, Sizes, Times [optional]
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-[More Information Needed]
-## Evaluation
-<!-- This section describes the evaluation protocols and provides the results. -->
-### Testing Data, Factors & Metrics
-#### Testing Data
-<!-- This should link to a Dataset Card if possible. -->
-[More Information Needed]
-#### Factors
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
-[More Information Needed]
-#### Metrics
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-[More Information Needed]
-### Results
-[More Information Needed]
-#### Summary
-## Model Examination [optional]
-<!-- Relevant interpretability work for the model goes here -->
-[More Information Needed]
-## Environmental Impact
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
-## Technical Specifications [optional]
-### Model Architecture and Objective
-[More Information Needed]
-### Compute Infrastructure
-[More Information Needed]
-#### Hardware
-[More Information Needed]
-#### Software
-[More Information Needed]
-## Citation [optional]
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
-**BibTeX:**
-[More Information Needed]
-**APA:**
-[More Information Needed]
-## Glossary [optional]
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
-[More Information Needed]
-## More Information [optional]
-[More Information Needed]
-## Model Card Authors [optional]
-[More Information Needed]
-## Model Card Contact
-[More Information Needed]

 ---
+language:
+- en
 library_name: transformers
+pipeline_tag: text-generation
+license: mit
 ---
+# DialRet
+<p align="center">
+<picture>
+    <img src="./main_fig.png" width="100%" style="margin: 0px auto;">
+</picture>
+Proceedings of PAKDD 2025 paper "DialRet: Enhancing Dialogue Retention for Multi-Session Conversations"
+Yohan Na*, Dahye Kim*, and Dong-Kyu chae.
+<p align="center"> 🤗 <a href="https://huggingface.co/collections/DILAB-HYU/">Models</a> &nbsp | &nbsp 📜 <a href="https://">Paper</a> | &nbsp 💻 <a href="https://github.com/"> Github </a>
+> [!Note]
+> The paper is written from a multi-session dialogue perspective, which is far from the instruction performance targeted by recent models.
+## Table of Contents
+- [Introduction](#introduction)
+- [Model Performance](#performance)
+- [Quickstart](#quickstart)
+- [License](#license)
+- [Citation](#citation)
+- [Contributors](#contributors)
+- [Contact](#contact)
+<br>
+## Introduction
+DialRet is a dialogue-specific language model designed for multi-session conversations.
+Instead of using memory modules, it leverages long-context LMs and instruction-tuning across eight tasks (e.g., dialogue generation, summarization, speaker relation extraction).
+It improves understanding and retention of past dialogues.
+The paper also introduces MSC-Bench, a benchmark evaluating dialogue models on memorability, specificity, engagement, and humanness.
+Experiments show DialRet outperforms existing models in multi-session dialogue quality and retention.
+### Model Performance
+Below are partial report on the performance of the `DialRet`. Please refer to the [Paper](https://) for the full results.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/RNvRrtb8UdZRP3xwikyD-.png)
+## Quickstart
+#### Example Usage for `DialRet`
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+model_name = "DILAB-HYU/DialRet-L1"
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype=torch.bfloat16,
+    trust_remote_code=True,
+).to("cuda")
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+session_num = 3
+session_role_1 = "Neighbors A"
+session_role_2 = "Neighbors B"
+session_system_prompt = f"You will be shown a {session_num} session dialogues between {session_role_1} and {session_role_2}. Please read and understand given multiple Dialogue Session, then complete the task under the guidance of Task Introduction.\n\n"
+session_input = """```
+Dialogue Session #1:
+Neighbors A:Hi there! I saw your cat in my backyard earlier. She's quite cute. What's her name?
+Neighbors B:Oh, thanks! Her name is Luna. She's a rescue cat.
+Neighbors A:That's really cool. How old is she?
+Neighbors B:She's about 2 years old.
+Neighbors A:Does she like being outside?
+Neighbors B:Not really. She's pretty much an indoor cat. She likes to snuggle up and sleep all day.
+Neighbors A:That's adorable. My kids would love her!
+Neighbors B:You're welcome to come over and visit her anytime.
+Neighbors A:Thanks, I'd love to! By the way, did you get your fence fixed?
+Neighbors B:Yes, we had it repaired last weekend. It was a relief to finally get it fixed.
+Neighbors A:I'm glad to hear that. Did you have to call in a professional?
+Neighbors B:Yeah, we had to call a fencing company to come and take care of it. They did a great job though, so we're happy with the results.
+Neighbors A:Good to know! I may have to call them too if I ever need fence repairs.
+Neighbors B:Absolutely, I can give you their contact information if you'd like.
+Neighbors A:Thanks, I appreciate it. Anyway, I won't keep you too long. Thanks for telling me about Luna!
+Neighbors B:No problem, happy to talk about her. See you later!
+```
+```
+Dialogue Session #2:
+Neighbors A:Can you believe it? A tree just fell on my car!
+Neighbors B:Oh no! Are you okay?
+Neighbors A:Yeah, luckily I wasn't in it at the time. But my car is completely totaled.
+Neighbors B:That's terrible. Did you call your insurance company?
+Neighbors A:Not yet, I'm still in shock. Plus, I was in the middle of reading a really interesting book about philosophy.
+Neighbors B:Oh, what book are you reading?
+Neighbors A:It's called "The Republic" by Plato. It's all about the concept of justice and government.
+Neighbors B:That sounds really fascinating. I've always been interested in philosophy, but I never know where to start.
+Neighbors A:Well, "The Republic" is a classic. But if you're just starting out, I'd recommend "Meditations" by Marcus Aurelius. It's a great introduction to Stoicism.
+Neighbors B:Thanks for the recommendation. I'll definitely check it out. But in the meantime, let's get your car situation sorted out. Do you need any help with anything?
+Neighbors A:That would be great, actually. Do you have any experience dealing with insurance companies?
+```
+```
+Dialogue Session #3:
+Neighbors A:Hey, Neighbors B. I have a bit of a problem and was hoping you could help me out.
+Neighbors B:Sure thing! What's going on?
+Neighbors A:Well, I'm having some trouble with my computer. It's just not working the way it should be, and I don't know what to do.
+Neighbors B:Ah, I see. What kind of issues are you having?
+Neighbors A:The screen keeps freezing up, and I can't seem to get anything done. I'm really getting frustrated because I have some important work that needs to be done.
+Neighbors B:Hmm, that sounds really frustrating. I think I might be able to help, though. Have you tried restarting your computer?
+Neighbors A: ###
+```"""
+session_task = """Task Introduction:
+After reading the entire Dialogue Sessions, please create an appropriate response.
+```
+Task Result:"""
+input_text= session_system_prompt + session_input + session_task
+input_ids = tokenizer(input_text, return_tensors="pt").input_ids.to("cuda")
+outputs = model.generate(inputs, max_new_tokens=4096, do_sample=False) # Finetuned with length 8192
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+# Output:
+# Neighbors A:Yeah, I've been trying that but it doesn't seem to be helping.
+```
+<br>
+## License
+The `DialRet` models are licensed under [MIT](https://opensource.org/license/mit).
+<br>
+## Citation
+```
+@article{2025dialret,
+      title={DialRet: Enhancing Dialogue Retention forMulti-session Conversations},
+      author={Yohan Na, Dahye Kim, Dong-kyu Chae},
+      year={2025},
+      url={},
+}
+```
+<br>
+## Thanks to
+- Yoo Hyun Jeong, Jongsoo Lee, Seonggyeom Kim, Myeongsoo Han, Byeongtae Park, Eunseon Seong, Harim Lee.
+<br>