English | [**中文**](README_ZH.md)
# AgentScope
Start building LLM-empowered multi-agent applications in an easier way.
[](https://arxiv.org/abs/2402.14034)
[](https://pypi.org/project/agentscope/)
[](https://pypi.org/project/agentscope/)
[](https://modelscope.github.io/agentscope/#welcome-to-agentscope-tutorial-hub)
[](https://modelscope.github.io/agentscope/)
[](https://modelscope.cn/studios?name=agentscope&page=1&sort=latest)
[](./LICENSE)
[](https://modelscope.github.io/agentscope/tutorial/contribute.html)
If you find our work helpful, please kindly
cite [our paper](https://arxiv.org/abs/2402.14034).
Welcome to join our community on
| [Discord](https://discord.gg/eYMpfnkG8h) | DingTalk |
|----------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------|
| <img src="https://gw.alicdn.com/imgextra/i1/O1CN01hhD1mu1Dd3BWVUvxN_!!6000000000238-2-tps-400-400.png" width="100" height="100"> | <img src="https://img.alicdn.com/imgextra/i2/O1CN01tuJ5971OmAqNg9cOw_!!6000000001747-0-tps-444-460.jpg" width="100" height="100"> |
----
## News
- <img src="https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png" alt="new" width="30" height="30"/>**[2024-04-30]** We release **AgentScope** v0.0.4 now!
- <img src="https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png" alt="new" width="30" height="30"/>**[2024-04-27]** [AgentScope Workstation](https://agentscope.aliyun.com/) is now online! You are welcome to try building your multi-agent application simply with our *drag-and-drop platform* and ask our *copilot* questions about AgentScope!
- <img src="https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png" alt="new" width="30" height="30"/>**[2024-04-19]** AgentScope supports Llama3 now! We provide [scripts](./examples/model_llama3) and example [model configuration](./examples/model_llama3) for quick set-up. Feel free to try llama3 in our examples!
- <img src="https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png" alt="new" width="30" height="30"/>**[2024-04-06]** We release **AgentScope** v0.0.3 now!
- <img src="https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png" alt="new" width="30" height="30"/>**[2024-04-06]** New examples [Gomoku](./examples/game_gomoku), [Conversation with ReAct Agent](./examples/conversation_with_react_agent), [Conversation with RAG Agent](./examples/conversation_with_RAG_agents) and [Distributed Parallel Search](./examples/distributed_search) are available now!
- **[2024-03-19]** We release **AgentScope** v0.0.2 now! In this new version,
AgentScope supports [ollama](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#supported-models)(A local CPU inference engine), [DashScope](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#supported-models) and Google [Gemini](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#supported-models) APIs.
- **[2024-03-19]** New examples ["Autonomous Conversation with Mentions"](./examples/conversation_with_mentions) and ["Basic Conversation with LangChain library"](./examples/conversation_with_langchain) are available now!
- **[2024-03-19]** The [Chinese tutorial](https://modelscope.github.io/agentscope/zh_CN/index.html) of AgentScope is online now!
- **[2024-02-27]** We release **AgentScope v0.0.1** now, which is also
available in [PyPI](https://pypi.org/project/agentscope/)!
- **[2024-02-14]** We release our paper "AgentScope: A Flexible yet Robust
Multi-Agent Platform" in [arXiv](https://arxiv.org/abs/2402.14034) now!
---
## What's AgentScope?
AgentScope is an innovative multi-agent platform designed to empower developers
to build multi-agent applications with large-scale models.
It features three high-level capabilities:
- 🤝 **Easy-to-Use**: Designed for developers, with [fruitful components](https://modelscope.github.io/agentscope/en/tutorial/204-service.html#),
[comprehensive documentation](https://modelscope.github.io/agentscope/en/index.html), and broad compatibility. Besides, [AgentScope Workstation](https://agentscope.aliyun.com/) provides a *drag-and-drop programming platform* and a *copilot* for beginners of AgentScope!
- ✅ **High Robustness**: Supporting customized fault-tolerance controls and
retry mechanisms to enhance application stability.
- 🚀 **Actor-Based Distribution**: Building distributed multi-agent
applications in a centralized programming manner for streamlined development.
**Supported Model Libraries**
AgentScope provides a list of `ModelWrapper` to support both local model
services and third-party model APIs.
| API | Task | Model Wrapper | Configuration | Some Supported Models |
|------------------------|-----------------|---------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------|-----------------------------------------------|
| OpenAI API | Chat | [`OpenAIChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/openai_model.py) |[guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#openai-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/openai_chat_template.json) | gpt-4, gpt-3.5-turbo, ... |
| | Embedding | [`OpenAIEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/openai_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#openai-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/openai_embedding_template.json) | text-embedding-ada-002, ... |
| | DALL·E | [`OpenAIDALLEWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/openai_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#openai-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/openai_dall_e_template.json) | dall-e-2, dall-e-3 |
| DashScope API | Chat | [`DashScopeChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_chat_template.json) | qwen-plus, qwen-max, ... |
| | Image Synthesis | [`DashScopeImageSynthesisWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br>[template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_image_synthesis_template.json) | wanx-v1 |
| | Text Embedding | [`DashScopeTextEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_text_embedding_template.json) | text-embedding-v1, text-embedding-v2, ... |
| | Multimodal | [`DashScopeMultiModalWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_multimodal_template.json) | qwen-vl-max, qwen-vl-chat-v1, qwen-audio-chat |
| Gemini API | Chat | [`GeminiChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/gemini_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#gemini-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/gemini_chat_template.json) | gemini-pro, ... |
| | Embedding | [`GeminiEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/gemini_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#gemini-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/gemini_embedding_template.json) | models/embedding-001, ... |
| ollama | Chat | [`OllamaChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/ollama_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#ollama-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/ollama_chat_template.json) | llama3, llama2, Mistral, ... |
| | Embedding | [`OllamaEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/ollama_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#ollama-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/ollama_embedding_template.json) | llama2, Mistral, ... |
| | Generation | [`OllamaGenerationWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/ollama_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#ollama-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/ollama_generate_template.json) | llama2, Mistral, ... |
| Post Request based API | - | [`PostAPIModelWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/post_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#post-request-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/postapi_model_config_template.json) | - |
**Supported Local Model Deployment**
AgentScope enables developers to rapidly deploy local model services using
the following libraries.
- [ollama (CPU inference)](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#ollama)
- [Flask + Transformers](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#with-transformers-library)
- [Flask + ModelScope](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#with-modelscope-library)
- [FastChat](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#fastchat)
- [vllm](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#vllm)
**Supported Services**
- Web Search
- Data Query
- Retrieval
- Code Execution
- File Operation
- Text Processing
**Example Applications**
- Model
- <img src="https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png" alt="new" width="30" height="30"/>[Using Llama3 in AgentScope](./examples/model_llama3)
- Conversation
- [Basic Conversation](./examples/conversation_basic)
- [Autonomous Conversation with Mentions](./examples/conversation_with_mentions)
- [Self-Organizing Conversation](./examples/conversation_self_organizing)
- [Basic Conversation with LangChain library](./examples/conversation_with_langchain)
- [Conversation with ReAct Agent](./examples/conversation_with_react_agent)
- [Conversation with RAG Agent](./examples/conversation_with_RAG_agents)
- Game
- [Gomoku](./examples/game_gomoku)
- [Werewolf](./examples/game_werewolf)
- Distribution
- [Distributed Conversation](./examples/distributed_basic)
- [Distributed Debate](./examples/distributed_debate)
- [Distributed Parallel Search](./examples/distributed_search)
- [Distributed Large Scale Simulation](./examples/distributed_simulation)
More models, services and examples are coming soon!
## Installation
AgentScope requires **Python 3.9** or higher.
***Note: This project is currently in active development, it's recommended to
install AgentScope from source.***
### From source
- Install AgentScope in editable mode:
```bash
# Pull the source code from GitHub
git clone https://github.com/modelscope/agentscope.git
# Install the package in editable mode
cd agentscope
pip install -e .
```
- To build distributed multi-agent applications:
```bash
# On windows
pip install -e .[distribute]
# On mac
pip install -e .\[distribute\]
```
### Using pip
- Install AgentScope from pip:
```bash
pip install agentscope
```
## Quick Start
### Configuration
In AgentScope, the model deployment and invocation are decoupled by
`ModelWrapper`.
To use these model wrappers, you need to prepare a model config file as
follows.
```python
model_config = {
# The identifies of your config and used model wrapper
"config_name": "{your_config_name}", # The name to identify the config
"model_type": "{model_type}", # The type to identify the model wrapper
# Detailed parameters into initialize the model wrapper
# ...
}
```
Taking OpenAI Chat API as an example, the model configuration is as follows:
```python
openai_model_config = {
"config_name": "my_openai_config", # The name to identify the config
"model_type": "openai_chat", # The type to identify the model wrapper
# Detailed parameters into initialize the model wrapper
"model_name": "gpt-4", # The used model in openai API, e.g. gpt-4, gpt-3.5-turbo, etc.
"api_key": "xxx", # The API key for OpenAI API. If not set, env
# variable OPENAI_API_KEY will be used.
"organization": "xxx", # The organization for OpenAI API. If not set, env
# variable OPENAI_ORGANIZATION will be used.
}
```
More details about how to set up local model services and prepare model
configurations is in our
[tutorial](https://modelscope.github.io/agentscope/index.html#welcome-to-agentscope-tutorial-hub).
### Create Agents
Create built-in user and assistant agents as follows.
```python
from agentscope.agents import DialogAgent, UserAgent
import agentscope
# Load model configs
agentscope.init(model_configs="./model_configs.json")
# Create a dialog agent and a user agent
dialog_agent = DialogAgent(name="assistant",
model_config_name="my_openai_config")
user_agent = UserAgent()
```
### Construct Conversation
In AgentScope, **message** is the bridge among agents, which is a
**dict** that contains two necessary fields `name` and `content` and an
optional field `url` to local files (image, video or audio) or website.
```python
from agentscope.message import Msg
x = Msg(name="Alice", content="Hi!")
x = Msg("Bob", "What about this picture I took?", url="/path/to/picture.jpg")
```
Start a conversation between two agents (e.g. dialog_agent and user_agent)
with the following code:
```python
x = None
while True:
x = dialog_agent(x)
x = user_agent(x)
if x.content == "exit": # user input "exit" to exit the conversation_basic
break
```
### AgentScope Studio
AgentScope provides an easy-to-use runtime user interface capable of
displaying multimodal output on the front end, including text, images,
audio and video. To start a studio, you should install the `full` version
of AgentScope.
```
# On windows
pip install -e .[full]
# On mac
pip install -e .\[full\]
```
Once installed, you can just run
```
as_studio path/to/your/script.py
```
Then the studio will be launched at `localhost:xxxx`, and you can see the UI similar to the
following:
To be able to use the `as_studio` functionality, please implement the `main` function in your code. More detail can be found in [src/agentscope/web/README.md](src/agentscope/web/README.md).
## Tutorial
- [Getting Started](https://modelscope.github.io/agentscope/en/tutorial/quick_start.html)
- [About AgentScope](https://modelscope.github.io/agentscope/en/tutorial/101-agentscope.html)
- [Installation](https://modelscope.github.io/agentscope/en/tutorial/102-installation.html)
- [Quick Start](https://modelscope.github.io/agentscope/en/tutorial/103-example.html)
- [Crafting Your First Application](https://modelscope.github.io/agentscope/en/tutorial/104-usecase.html)
- [Logging and WebUI](https://modelscope.github.io/agentscope/en/tutorial/105-logging.html#)
- [Advanced Exploration](https://modelscope.github.io/agentscope/en/tutorial/advance.html)
- [Customize Your Own Agent](https://modelscope.github.io/agentscope/en/tutorial/201-agent.html)
- [Pipeline and MsgHub](https://modelscope.github.io/agentscope/en/tutorial/202-pipeline.html)
- [Model](https://modelscope.github.io/agentscope/en/tutorial/203-model.html)
- [Service](https://modelscope.github.io/agentscope/en/tutorial/204-service.html)
- [Memory](https://modelscope.github.io/agentscope/en/tutorial/205-memory.html)
- [Prompt Engineering](https://modelscope.github.io/agentscope/en/tutorial/206-prompt.html)
- [Monitor](https://modelscope.github.io/agentscope/en/tutorial/207-monitor.html)
- [Distribution](https://modelscope.github.io/agentscope/en/tutorial/208-distribute.html)
- [Get Involved](https://modelscope.github.io/agentscope/en/tutorial/contribute.html)
- [Join AgentScope Community](https://modelscope.github.io/agentscope/en/tutorial/301-community.html)
- [Contribute to AgentScope](https://modelscope.github.io/agentscope/en/tutorial/302-contribute.html)
## License
AgentScope is released under Apache License 2.0.
## Contributing
Contributions are always welcomed!
We provide a developer version with additional pre-commit hooks to perform
checks compared to the official version:
```bash
# For windows
pip install -e .[dev]
# For mac
pip install -e .\[dev\]
# Install pre-commit hooks
pre-commit install
```
Please refer to our [Contribution Guide](https://modelscope.github.io/agentscope/en/tutorial/302-contribute.html) for more details.
## References
If you find our work helpful for your research or application, please
cite [our paper](https://arxiv.org/abs/2402.14034):
```
@article{agentscope,
author = {Dawei Gao and
Zitao Li and
Weirui Kuang and
Xuchen Pan and
Daoyuan Chen and
Zhijian Ma and
Bingchen Qian and
Liuyi Yao and
Lin Zhu and
Chen Cheng and
Hongzhu Shi and
Yaliang Li and
Bolin Ding and
Jingren Zhou},
title = {AgentScope: A Flexible yet Robust Multi-Agent Platform},
journal = {CoRR},
volume = {abs/2402.14034},
year = {2024},
}
```
Raw data
{
"_id": null,
"home_page": "https://github.com/modelscope/agentscope",
"name": "agentscope",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "deep-learning, multi agents, agents",
"author": "SysML team of Alibaba Tongyi Lab",
"author_email": "gaodawei.gdw@alibaba-inc.com",
"download_url": "https://files.pythonhosted.org/packages/3e/a6/441bb8020cd7b83fc272e0d569a2a155ec77d407667b504268c6f760ffd3/agentscope-0.0.4.tar.gz",
"platform": null,
"description": "English | [**\u4e2d\u6587**](README_ZH.md)\n\n# AgentScope\n\nStart building LLM-empowered multi-agent applications in an easier way.\n\n[](https://arxiv.org/abs/2402.14034)\n[](https://pypi.org/project/agentscope/)\n[](https://pypi.org/project/agentscope/)\n[](https://modelscope.github.io/agentscope/#welcome-to-agentscope-tutorial-hub)\n[](https://modelscope.github.io/agentscope/)\n[](https://modelscope.cn/studios?name=agentscope&page=1&sort=latest)\n\n[](./LICENSE)\n[](https://modelscope.github.io/agentscope/tutorial/contribute.html)\n\nIf you find our work helpful, please kindly\ncite [our paper](https://arxiv.org/abs/2402.14034).\n\nWelcome to join our community on\n\n| [Discord](https://discord.gg/eYMpfnkG8h) | DingTalk |\n|----------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------|\n| <img src=\"https://gw.alicdn.com/imgextra/i1/O1CN01hhD1mu1Dd3BWVUvxN_!!6000000000238-2-tps-400-400.png\" width=\"100\" height=\"100\"> | <img src=\"https://img.alicdn.com/imgextra/i2/O1CN01tuJ5971OmAqNg9cOw_!!6000000001747-0-tps-444-460.jpg\" width=\"100\" height=\"100\"> |\n\n----\n\n## News\n\n- <img src=\"https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png\" alt=\"new\" width=\"30\" height=\"30\"/>**[2024-04-30]** We release **AgentScope** v0.0.4 now!\n\n- <img src=\"https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png\" alt=\"new\" width=\"30\" height=\"30\"/>**[2024-04-27]** [AgentScope Workstation](https://agentscope.aliyun.com/) is now online! You are welcome to try building your multi-agent application simply with our *drag-and-drop platform* and ask our *copilot* questions about AgentScope!\n\n- <img src=\"https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png\" alt=\"new\" width=\"30\" height=\"30\"/>**[2024-04-19]** AgentScope supports Llama3 now! We provide [scripts](./examples/model_llama3) and example [model configuration](./examples/model_llama3) for quick set-up. Feel free to try llama3 in our examples!\n\n- <img src=\"https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png\" alt=\"new\" width=\"30\" height=\"30\"/>**[2024-04-06]** We release **AgentScope** v0.0.3 now!\n\n- <img src=\"https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png\" alt=\"new\" width=\"30\" height=\"30\"/>**[2024-04-06]** New examples [Gomoku](./examples/game_gomoku), [Conversation with ReAct Agent](./examples/conversation_with_react_agent), [Conversation with RAG Agent](./examples/conversation_with_RAG_agents) and [Distributed Parallel Search](./examples/distributed_search) are available now!\n\n- **[2024-03-19]** We release **AgentScope** v0.0.2 now! In this new version,\nAgentScope supports [ollama](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#supported-models)(A local CPU inference engine), [DashScope](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#supported-models) and Google [Gemini](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#supported-models) APIs.\n\n- **[2024-03-19]** New examples [\"Autonomous Conversation with Mentions\"](./examples/conversation_with_mentions) and [\"Basic Conversation with LangChain library\"](./examples/conversation_with_langchain) are available now!\n\n- **[2024-03-19]** The [Chinese tutorial](https://modelscope.github.io/agentscope/zh_CN/index.html) of AgentScope is online now!\n\n- **[2024-02-27]** We release **AgentScope v0.0.1** now, which is also\navailable in [PyPI](https://pypi.org/project/agentscope/)!\n- **[2024-02-14]** We release our paper \"AgentScope: A Flexible yet Robust\nMulti-Agent Platform\" in [arXiv](https://arxiv.org/abs/2402.14034) now!\n\n---\n\n## What's AgentScope?\n\nAgentScope is an innovative multi-agent platform designed to empower developers\nto build multi-agent applications with large-scale models.\nIt features three high-level capabilities:\n\n- \ud83e\udd1d **Easy-to-Use**: Designed for developers, with [fruitful components](https://modelscope.github.io/agentscope/en/tutorial/204-service.html#),\n[comprehensive documentation](https://modelscope.github.io/agentscope/en/index.html), and broad compatibility. Besides, [AgentScope Workstation](https://agentscope.aliyun.com/) provides a *drag-and-drop programming platform* and a *copilot* for beginners of AgentScope!\n\n- \u2705 **High Robustness**: Supporting customized fault-tolerance controls and\nretry mechanisms to enhance application stability.\n\n- \ud83d\ude80 **Actor-Based Distribution**: Building distributed multi-agent\napplications in a centralized programming manner for streamlined development.\n\n**Supported Model Libraries**\n\nAgentScope provides a list of `ModelWrapper` to support both local model\nservices and third-party model APIs.\n\n| API | Task | Model Wrapper | Configuration | Some Supported Models |\n|------------------------|-----------------|---------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------|-----------------------------------------------|\n| OpenAI API | Chat | [`OpenAIChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/openai_model.py) |[guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#openai-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/openai_chat_template.json) | gpt-4, gpt-3.5-turbo, ... |\n| | Embedding | [`OpenAIEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/openai_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#openai-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/openai_embedding_template.json) | text-embedding-ada-002, ... |\n| | DALL\u00b7E | [`OpenAIDALLEWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/openai_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#openai-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/openai_dall_e_template.json) | dall-e-2, dall-e-3 |\n| DashScope API | Chat | [`DashScopeChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_chat_template.json) | qwen-plus, qwen-max, ... |\n| | Image Synthesis | [`DashScopeImageSynthesisWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br>[template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_image_synthesis_template.json) | wanx-v1 |\n| | Text Embedding | [`DashScopeTextEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_text_embedding_template.json) | text-embedding-v1, text-embedding-v2, ... |\n| | Multimodal | [`DashScopeMultiModalWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/dashscope_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#dashscope-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/dashscope_multimodal_template.json) | qwen-vl-max, qwen-vl-chat-v1, qwen-audio-chat |\n| Gemini API | Chat | [`GeminiChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/gemini_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#gemini-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/gemini_chat_template.json) | gemini-pro, ... |\n| | Embedding | [`GeminiEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/gemini_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#gemini-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/gemini_embedding_template.json) | models/embedding-001, ... |\n| ollama | Chat | [`OllamaChatWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/ollama_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#ollama-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/ollama_chat_template.json) | llama3, llama2, Mistral, ... |\n| | Embedding | [`OllamaEmbeddingWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/ollama_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#ollama-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/ollama_embedding_template.json) | llama2, Mistral, ... |\n| | Generation | [`OllamaGenerationWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/ollama_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#ollama-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/ollama_generate_template.json) | llama2, Mistral, ... |\n| Post Request based API | - | [`PostAPIModelWrapper`](https://github.com/modelscope/agentscope/blob/main/src/agentscope/models/post_model.py) | [guidance](https://modelscope.github.io/agentscope/en/tutorial/203-model.html#post-request-api) <br> [template](https://github.com/modelscope/agentscope/blob/main/examples/model_configs_template/postapi_model_config_template.json) | - |\n\n**Supported Local Model Deployment**\n\nAgentScope enables developers to rapidly deploy local model services using\nthe following libraries.\n\n- [ollama (CPU inference)](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#ollama)\n- [Flask + Transformers](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#with-transformers-library)\n- [Flask + ModelScope](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#with-modelscope-library)\n- [FastChat](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#fastchat)\n- [vllm](https://github.com/modelscope/agentscope/blob/main/scripts/README.md#vllm)\n\n**Supported Services**\n\n- Web Search\n- Data Query\n- Retrieval\n- Code Execution\n- File Operation\n- Text Processing\n\n**Example Applications**\n\n- Model\n - <img src=\"https://img.alicdn.com/imgextra/i3/O1CN01SFL0Gu26nrQBFKXFR_!!6000000007707-2-tps-500-500.png\" alt=\"new\" width=\"30\" height=\"30\"/>[Using Llama3 in AgentScope](./examples/model_llama3)\n\n- Conversation\n - [Basic Conversation](./examples/conversation_basic)\n - [Autonomous Conversation with Mentions](./examples/conversation_with_mentions)\n - [Self-Organizing Conversation](./examples/conversation_self_organizing)\n - [Basic Conversation with LangChain library](./examples/conversation_with_langchain)\n - [Conversation with ReAct Agent](./examples/conversation_with_react_agent)\n - [Conversation with RAG Agent](./examples/conversation_with_RAG_agents)\n\n- Game\n - [Gomoku](./examples/game_gomoku)\n - [Werewolf](./examples/game_werewolf)\n\n- Distribution\n - [Distributed Conversation](./examples/distributed_basic)\n - [Distributed Debate](./examples/distributed_debate)\n - [Distributed Parallel Search](./examples/distributed_search)\n - [Distributed Large Scale Simulation](./examples/distributed_simulation)\n\nMore models, services and examples are coming soon!\n\n## Installation\n\nAgentScope requires **Python 3.9** or higher.\n\n***Note: This project is currently in active development, it's recommended to\ninstall AgentScope from source.***\n\n### From source\n\n- Install AgentScope in editable mode:\n\n```bash\n# Pull the source code from GitHub\ngit clone https://github.com/modelscope/agentscope.git\n\n# Install the package in editable mode\ncd agentscope\npip install -e .\n```\n\n- To build distributed multi-agent applications:\n\n```bash\n# On windows\npip install -e .[distribute]\n# On mac\npip install -e .\\[distribute\\]\n```\n\n### Using pip\n\n- Install AgentScope from pip:\n\n```bash\npip install agentscope\n```\n\n## Quick Start\n\n### Configuration\n\nIn AgentScope, the model deployment and invocation are decoupled by\n`ModelWrapper`.\n\nTo use these model wrappers, you need to prepare a model config file as\nfollows.\n\n```python\nmodel_config = {\n # The identifies of your config and used model wrapper\n \"config_name\": \"{your_config_name}\", # The name to identify the config\n \"model_type\": \"{model_type}\", # The type to identify the model wrapper\n\n # Detailed parameters into initialize the model wrapper\n # ...\n}\n```\n\nTaking OpenAI Chat API as an example, the model configuration is as follows:\n\n```python\nopenai_model_config = {\n \"config_name\": \"my_openai_config\", # The name to identify the config\n \"model_type\": \"openai_chat\", # The type to identify the model wrapper\n\n # Detailed parameters into initialize the model wrapper\n \"model_name\": \"gpt-4\", # The used model in openai API, e.g. gpt-4, gpt-3.5-turbo, etc.\n \"api_key\": \"xxx\", # The API key for OpenAI API. If not set, env\n # variable OPENAI_API_KEY will be used.\n \"organization\": \"xxx\", # The organization for OpenAI API. If not set, env\n # variable OPENAI_ORGANIZATION will be used.\n}\n```\n\nMore details about how to set up local model services and prepare model\nconfigurations is in our\n[tutorial](https://modelscope.github.io/agentscope/index.html#welcome-to-agentscope-tutorial-hub).\n\n### Create Agents\n\nCreate built-in user and assistant agents as follows.\n\n```python\nfrom agentscope.agents import DialogAgent, UserAgent\nimport agentscope\n\n# Load model configs\nagentscope.init(model_configs=\"./model_configs.json\")\n\n# Create a dialog agent and a user agent\ndialog_agent = DialogAgent(name=\"assistant\",\n model_config_name=\"my_openai_config\")\nuser_agent = UserAgent()\n```\n\n### Construct Conversation\n\nIn AgentScope, **message** is the bridge among agents, which is a\n**dict** that contains two necessary fields `name` and `content` and an\noptional field `url` to local files (image, video or audio) or website.\n\n```python\nfrom agentscope.message import Msg\n\nx = Msg(name=\"Alice\", content=\"Hi!\")\nx = Msg(\"Bob\", \"What about this picture I took?\", url=\"/path/to/picture.jpg\")\n```\n\nStart a conversation between two agents (e.g. dialog_agent and user_agent)\nwith the following code:\n\n```python\nx = None\nwhile True:\n x = dialog_agent(x)\n x = user_agent(x)\n if x.content == \"exit\": # user input \"exit\" to exit the conversation_basic\n break\n```\n\n### AgentScope Studio\n\nAgentScope provides an easy-to-use runtime user interface capable of\ndisplaying multimodal output on the front end, including text, images,\naudio and video. To start a studio, you should install the `full` version\nof AgentScope.\n\n```\n# On windows\npip install -e .[full]\n# On mac\npip install -e .\\[full\\]\n```\n\nOnce installed, you can just run\n\n```\nas_studio path/to/your/script.py\n```\n\nThen the studio will be launched at `localhost:xxxx`, and you can see the UI similar to the\nfollowing:\n\nTo be able to use the `as_studio` functionality, please implement the `main` function in your code. More detail can be found in [src/agentscope/web/README.md](src/agentscope/web/README.md).\n\n## Tutorial\n\n- [Getting Started](https://modelscope.github.io/agentscope/en/tutorial/quick_start.html)\n - [About AgentScope](https://modelscope.github.io/agentscope/en/tutorial/101-agentscope.html)\n - [Installation](https://modelscope.github.io/agentscope/en/tutorial/102-installation.html)\n - [Quick Start](https://modelscope.github.io/agentscope/en/tutorial/103-example.html)\n - [Crafting Your First Application](https://modelscope.github.io/agentscope/en/tutorial/104-usecase.html)\n - [Logging and WebUI](https://modelscope.github.io/agentscope/en/tutorial/105-logging.html#)\n- [Advanced Exploration](https://modelscope.github.io/agentscope/en/tutorial/advance.html)\n - [Customize Your Own Agent](https://modelscope.github.io/agentscope/en/tutorial/201-agent.html)\n - [Pipeline and MsgHub](https://modelscope.github.io/agentscope/en/tutorial/202-pipeline.html)\n - [Model](https://modelscope.github.io/agentscope/en/tutorial/203-model.html)\n - [Service](https://modelscope.github.io/agentscope/en/tutorial/204-service.html)\n - [Memory](https://modelscope.github.io/agentscope/en/tutorial/205-memory.html)\n - [Prompt Engineering](https://modelscope.github.io/agentscope/en/tutorial/206-prompt.html)\n - [Monitor](https://modelscope.github.io/agentscope/en/tutorial/207-monitor.html)\n - [Distribution](https://modelscope.github.io/agentscope/en/tutorial/208-distribute.html)\n- [Get Involved](https://modelscope.github.io/agentscope/en/tutorial/contribute.html)\n - [Join AgentScope Community](https://modelscope.github.io/agentscope/en/tutorial/301-community.html)\n - [Contribute to AgentScope](https://modelscope.github.io/agentscope/en/tutorial/302-contribute.html)\n\n## License\n\nAgentScope is released under Apache License 2.0.\n\n## Contributing\n\nContributions are always welcomed!\n\nWe provide a developer version with additional pre-commit hooks to perform\nchecks compared to the official version:\n\n```bash\n# For windows\npip install -e .[dev]\n# For mac\npip install -e .\\[dev\\]\n\n# Install pre-commit hooks\npre-commit install\n```\n\nPlease refer to our [Contribution Guide](https://modelscope.github.io/agentscope/en/tutorial/302-contribute.html) for more details.\n\n## References\n\nIf you find our work helpful for your research or application, please\ncite [our paper](https://arxiv.org/abs/2402.14034):\n\n```\n@article{agentscope,\n author = {Dawei Gao and\n Zitao Li and\n Weirui Kuang and\n Xuchen Pan and\n Daoyuan Chen and\n Zhijian Ma and\n Bingchen Qian and\n Liuyi Yao and\n Lin Zhu and\n Chen Cheng and\n Hongzhu Shi and\n Yaliang Li and\n Bolin Ding and\n Jingren Zhou},\n title = {AgentScope: A Flexible yet Robust Multi-Agent Platform},\n journal = {CoRR},\n volume = {abs/2402.14034},\n year = {2024},\n}\n```\n",
"bugtrack_url": null,
"license": "Apache License 2.0",
"summary": "AgentScope: A Flexible yet Robust Multi-Agent Platform.",
"version": "0.0.4",
"project_urls": {
"Download": "https://github.com/modelscope/agentscope/archive/v0.0.4.tar.gz",
"Homepage": "https://github.com/modelscope/agentscope"
},
"split_keywords": [
"deep-learning",
" multi agents",
" agents"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "304a10a97f373382d7937809038cfef26c1b3072040eebd5b90e399f150e3c66",
"md5": "dcb2d25eda06fad2779a409a0a777b73",
"sha256": "f1109cf0b8f0cf8367417b0691da9f87d6f004435847750ea6251ca8801cdc07"
},
"downloads": -1,
"filename": "agentscope-0.0.4-py3-none-any.whl",
"has_sig": false,
"md5_digest": "dcb2d25eda06fad2779a409a0a777b73",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 432657,
"upload_time": "2024-04-30T09:49:25",
"upload_time_iso_8601": "2024-04-30T09:49:25.886393Z",
"url": "https://files.pythonhosted.org/packages/30/4a/10a97f373382d7937809038cfef26c1b3072040eebd5b90e399f150e3c66/agentscope-0.0.4-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "3ea6441bb8020cd7b83fc272e0d569a2a155ec77d407667b504268c6f760ffd3",
"md5": "dd5e8b31ea05cc49eed260f321caef8d",
"sha256": "505b195e47046906810021866f2f5dddf493b6ef4ea2851acbc022f0e4b6b8c3"
},
"downloads": -1,
"filename": "agentscope-0.0.4.tar.gz",
"has_sig": false,
"md5_digest": "dd5e8b31ea05cc49eed260f321caef8d",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 393503,
"upload_time": "2024-04-30T09:49:28",
"upload_time_iso_8601": "2024-04-30T09:49:28.353916Z",
"url": "https://files.pythonhosted.org/packages/3e/a6/441bb8020cd7b83fc272e0d569a2a155ec77d407667b504268c6f760ffd3/agentscope-0.0.4.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-04-30 09:49:28",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "modelscope",
"github_project": "agentscope",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "agentscope"
}
JSON
