V0.1 (#112)

v0.1 consists of 7 large language models and 2 inference libraries to run rubra models. v0.0 is no longer supported as a majority of that work has moved to https://github.com/gptscript-ai/gptscript --------- Co-authored-by: Yingbei <[email protected]>
rubra-ai · Jul 1, 2024 · dfd9a66 · dfd9a66
1 parent abf872e
commit dfd9a66
Show file tree

Hide file tree

Showing 124 changed files with 1,707 additions and 17,037 deletions.
diff --git a/.gitignore b/.gitignore
@@ -228,22 +228,6 @@ zipalign*
 **/node_modules
 **/docs/yarn.lock
 
-#Tauri
-tauri/src-tauri/target
-logs
-*.log
-npm-debug.log*
-yarn-debug.log*
-yarn-error.log*
-pnpm-debug.log*
-lerna-debug.log*
-
-node_modules
-dist
-dist-ssr
-*.local
-**/.xwin-cache
-
 # Editor directories and files
 .vscode/*
 !.vscode/extensions.json

diff --git a/.gitmodules b/.gitmodules
@@ -0,0 +1,9 @@
+[submodule "tools.cpp"]
+	path = tools.cpp
+	url = https://github.com/rubra-ai/tools.cpp.git
+[submodule "vllm"]
+	path = vllm
+	url = https://github.com/rubra-ai/vllm.git
+[submodule "rubra-tools"]
+	path = rubra-tools
+	url = https://github.com/rubra-ai/rubra-tools.git
diff --git a/Makefile b/Makefile
@@ -67,4 +67,4 @@ validate-docs:
 		echo "Encountered dirty repo!"; \
 		git diff; \
 		exit 1 \
-	;fi
+	;fi
diff --git a/README.md b/README.md
@@ -1,59 +1,42 @@
-# Rubra
-
-Rubra is an open-source ChatGPT. It's designed for users who want:
-
-- **Multi-Model Support:** Rubra integrates with a variety of LLMs, including a local model optimized for Rubra, as well as models from OpenAI and Anthropic. More providers will be added in the future.
-- **Assistant Tools:** Create powerful assistants using tools for web search, knowledge retrieval, and more, all designed to augment your LLMs with the information they need to be truly helpful.
-- **OpenAI API Compatibility:** Use Rubra's OpenAI-compatible Assistants API, allowing you to use OpenAI's Python and JavaScript libraries to create and manage Assistants.
-- **Self-Hosting:** Keep your data private and secure by running Rubra on your own hardware.
-
-## Getting Started
+<p align="left">
+    <a href="README_CN.md">中文</a>&nbsp ｜ &nbspEnglish&nbsp </a>
+</p>
+<br><br>
 
-### Prerequisites
+# Rubra
 
-- M-series Mac or Linux with GPU
-  - On MacOS you need to have Xcode Command Line Tools installed: `xcode-select --install`
-- At least 16 GB RAM
-- At least 10 GB of available disk space
-- Docker and Docker Compose (>= v2.17.0) installed
+#### Rubra is a collection of open-weight, tool-calling LLMs.
 
-### Installation
+Rubra enhances the top open-weight large language models with tool-calling capability. The ability to call user-defined external tools in a deterministic manner while reasoning and chatting makes Rubra models ideal for agentic use cases.
 
-Rubra offers a simple one-command installation:
+All models are enhanced from the top open-source LLMs with further post-training and methods that effectively teach instruct-tuned models new skills while mitigating catastrophic forgetting. For easy use, we extend popular inferencing projects, allowing you to run Rubra models easily.
 
-```bash
-curl -sfL https://get.rubra.ai | sh -s -- start
-```
+## Enhanced Models
 
-After installation, access the Rubra UI at `http://localhost:8501` and start exploring the capabilities of your new ChatGPT-like assistant.
+| Enhanced Model                                                        | Context Length | Size | Parent Model Publisher |
+|-----------------------------------------------------------------------|----------------|------|------------------------|
+| [rubra-ai/Meta-Llama-3-8B-Instruct](https://huggingface.co/rubra-ai/Meta-Llama-3-8B-Instruct)   | 8,000          | 8B   | Meta             |
+| [rubra-ai/Meta-Llama-3-70B-Instruct](https://huggingface.co/rubra-ai/Meta-Llama-3-70B-Instruct) | 8,000          | 70B  | Meta             |
+| [rubra-ai/gemma-1.1-2b-it](https://huggingface.co/rubra-ai/gemma-1.1-2b-it)                     | 8,192          | 2B   | Google                 |
+| [rubra-ai/Mistral-7B-Instruct-v0.3](https://huggingface.co/rubra-ai/Mistral-7B-Instruct-v0.3)   | 32,000         | 7B   | Mistral              |
+| [rubra-ai/Mistral-7B-Instruct-v0.2](https://huggingface.co/rubra-ai/Mistral-7B-Instruct-v0.2)    | 32,000         | 7B   | Mistral              |
+| [rubra-ai/Phi-3-vision-128k-instruct](https://huggingface.co/rubra-ai/Phi-3-vision-128k-instruct)| 128,000        | 3B   | Microsoft              |
+| [rubra-ai/Qwen2-7B-Instruct](https://huggingface.co/rubra-ai/Qwen2-7B-Instruct)                 | 131,072        | 7B   | Qwen                   |
 
-## Usage
+## Demo
 
-Here's a quick example of how to create an assistant using Rubra's API, compatible with OpenAI's libraries:
+Try out the models immediately without downloading anything in Our [Huggingface Spaces](https://huggingface.co/spaces/sanjay920/rubra-v0.1-dev)! It's free and requires no login.
 
-```python
-from openai import OpenAI
+## Run Rubra Models Locally
 
-client = OpenAI(
-    base_url="http://localhost:8000",  # Rubra backend
-    api_key=""
-)
+We extend the following inferencing tools to run Rubra models in an OpenAI-compatible tool-calling format for local use:
 
-assistant = client.beta.assistants.create(
-  instructions="You are a customer support chatbot. Use your knowledge base to best respond to customer queries.",
-  model="rubra_local",
-  tools=[{"type": "retrieval"}],
-  file_ids=[client.files.create(file=open("knowledge.txt", "rb"), purpose='assistants').id]
-)
-```
+- [llama.cpp](https://github.com/ggerganov/llama.cpp)
+- [vllm](https://github.com/vllm-project/vllm)
 
 ## Contributing
 
-We welcome contributions from the developer community! Whether it's adding new features, fixing bugs, or improving documentation, your help is invaluable. Check out our [contributing guidelines](CONTRIBUTING.md) for more information on how to get involved.
-
-## Support
-
-If you encounter any issues or have questions, please file an issue on GitHub. For more detailed guidance and discussions, join our community on [Discord](https://discord.gg/swvAH2DXZH) or [Slack](https://slack.acorn.io) or start a [Github discussion](https://github.com/rubra-ai/rubra/discussions).
+Contributions to Rubra are welcome! We'd love to improve tool-calling capability in the models based on your feedback. Please open an issue if your tool doesn't work.
 
 ---
 

diff --git a/README_CN.md b/README_CN.md
@@ -0,0 +1,52 @@
+<p align="left">
+    中文</a>&nbsp ｜ &nbsp<a href="README.md">English</a>&nbsp</a>
+</p>
+<br><br>
+
+
+# Rubra
+
+#### Rubra 是一系列开放权重、聚焦于工具调用的大模型（LLM）。
+
+Rubra 增强了当前最流行的一系列开放权重大模型（LLM）的工具调用能力。以能够在和用户对话时以稳定的方式调用用户定义的外部工具，使 Rubra 大模型非常适用于Agent相关的场景。
+
+所有模型均基于流行的Instruct模型，通过进一步的微调，有效地教授或增强模型调用工具的能力，同时尽可能减少模型在基础能力和知识上的流失。为了便于用户使用，我们扩展了流行的llm本地部署项目，让您可以轻松运行 Rubra 模型。
+
+## Rubra模型系列
+
+| 模型 | 最大上下文长度 | 大小 | 基础模型发布者 |
+|---------------------------------------------------------------|----------------|------|----------------------|
+| [rubra-ai/Meta-Llama-3-8B-Instruct](https://huggingface.co/rubra-ai/Meta-Llama-3-8B-Instruct)   | 8,000          | 8B   | Meta             |
+| [rubra-ai/Meta-Llama-3-70B-Instruct](https://huggingface.co/rubra-ai/Meta-Llama-3-70B-Instruct) | 8,000          | 70B  | Meta             |
+| [rubra-ai/gemma-1.1-2b-it](https://huggingface.co/rubra-ai/gemma-1.1-2b-it)                     | 8,192          | 2B   | Google                 |
+| [rubra-ai/Mistral-7B-Instruct-v0.3](https://huggingface.co/rubra-ai/Mistral-7B-Instruct-v0.3)   | 32,000         | 7B   | Mistral              |
+| [rubra-ai/Mistral-7B-Instruct-v0.2](https://huggingface.co/rubra-ai/Mistral-7B-Instruct-v0.2)   | 32,000         | 7B   | Mistral              |
+| [rubra-ai/Phi-3-vision-128k-instruct](https://huggingface.co/rubra-ai/Phi-3-vision-128k-instruct)| 128,000        | 3B   | Microsoft              |
+| [rubra-ai/Qwen2-7B-Instruct](https://huggingface.co/rubra-ai/Qwen2-7B-Instruct)                 | 131,072        | 7B   | Qwen                   |
+
+## Demo
+
+在我们的 [Huggingface Spaces](https://huggingface.co/spaces/sanjay920/rubra-v0.1-dev) 上可以免费试用以上的大模型，不需要登录！
+
+## 本地部署运行 Rubra 模型
+
+我们扩展了以下部署工具，以在 OpenAI 风格的API格式下本地运行 Rubra 模型：
+
+- [llama.cpp](https://github.com/ggerganov/llama.cpp)
+- [vllm](https://github.com/vllm-project/vllm)
+
+## 贡献
+
+欢迎您参与对 Rubra 的进一步开发！我们希望根据您的反馈改进模型的工具调用能力。如果您的工具调用不起作用或出错，请创建一个Issue并分享您遇到的问题。
+
+---
+
+## 许可证
+
+版权所有 (c) 2024 Acorn Labs, Inc.
+
+根据 Apache License, Version 2.0（“许可证”）授权；您不得在不符合许可证的情况下使用此文件。您可以在以下网址获取许可证副本：
+
+<http://www.apache.org/licenses/LICENSE-2.0>
+
+除非适用法律要求或书面同意，按“原样”分发的软件不附带任何明示或暗示的保证或条件。请参阅许可证以了解特定语言的管理权限和限制。