Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,58 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
language:
|
| 5 |
+
- en
|
| 6 |
+
tags:
|
| 7 |
+
- semiconductor
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<div style="text-align:center">
|
| 11 |
+
<p><strong><span style="font-size: 24px;">ChipExpert-8B-Instruct</span></strong></p>
|
| 12 |
+
</div>
|
| 13 |
+
|
| 14 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 15 |
+
__ChipExpert__ is an open-source, instructional LLM dedicated to the semiconductor industry, covering knowledge across multiple sub-domains, including analog circuits, digital circuits, radio frequency (RF), semiconductor devices, electronic design automation (EDA), system-on-chip (SoC), in-memory computing, and more. This model aims to assist students in the IC field to learn fundamental knowledge, engineers to inquire about technical details, and researchers to investigate cutting-edge papers and research topics. The ultimate goal of this model is to help the integrated circuit industry reduce the learning barrier for students and lower the training costs for engineers.
|
| 16 |
+
|
| 17 |
+
You can directly experience the capabilities and performance of ChipExpert by accessing <font color="blue">the provided link</font>(http://27.18.114.16:23023). The UI is shown in the figure below.
|
| 18 |
+
|
| 19 |
+
<div align="center">
|
| 20 |
+
<img src="images/demo.png" alt="" width="1000" height="1000">
|
| 21 |
+
</div>
|
| 22 |
+
|
| 23 |
+
## Model Details
|
| 24 |
+
- The base model of ChipExpert was built upon the Llama-3 8B Base Model.
|
| 25 |
+
- The training process can be divided into three main phases: Continue Pretraining -> Supervised Fine-Tuning -> Reinforcement Learning with Human Feedback.
|
| 26 |
+
- The continue pretraining data comes from specialized corpora in ten sub-domains of IC design.
|
| 27 |
+
- The majority of the data in the instruction dataset is synthetic data, generated through an innovative multi-agent QA system, ensuring the quantity, quality, and diversity of the data.
|
| 28 |
+
- The professionalism of the responses is enhanced by utilizing innovative AI alignment methods.
|
| 29 |
+
- ChipExpert got benchmarked on the ChipBench dataset which will be released for the community soon!
|
| 30 |
+
- ChipExpert achieves superior performance in both foundational and cutting-edge knowledge in the field of semiconductors compared to mainstream closed-source models.
|
| 31 |
+
|
| 32 |
+
## Model Description
|
| 33 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 34 |
+
|
| 35 |
+
- Developed by: NCTIEDA (National Center of Technology Innovation for EDA)
|
| 36 |
+
<div align="center">
|
| 37 |
+
<img src="images/logo.png" alt="" width="150" height="150">
|
| 38 |
+
</div>
|
| 39 |
+
- Model type: Instruction Model
|
| 40 |
+
- Language(s): English
|
| 41 |
+
- License: Apache License 2.0
|
| 42 |
+
- Finetuned from model: Llama 3
|
| 43 |
+
|
| 44 |
+
## Out-of-Scope Use
|
| 45 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 46 |
+
This is the first version of ChipExpert, more capable versions with enhanced abilities will be released soon.
|
| 47 |
+
|
| 48 |
+
## Model Card Authors
|
| 49 |
+
|
| 50 |
+
- Chao Zhang: zc13815813901@gmail.com
|
| 51 |
+
- Wensuo Wang: 1136823452@qq.com
|
| 52 |
+
- Frenkie: zhaoyangzh@seu.edu.cn
|
| 53 |
+
|
| 54 |
+
## Model Card Contact
|
| 55 |
+
|
| 56 |
+
- Chao Zhang: zc13815813901@gmail.com
|
| 57 |
+
- Wensuo Wang: 1136823452@qq.com
|
| 58 |
+
- Frenkie: zhaoyangzh@seu.edu.cn
|