Releases: LlmKira/Openaibot
Releases · LlmKira/Openaibot
app_4.0.0_patch
Fix Bug
- Redis Message Read
app_1.0.0 V4 refactor
Merge pull request #376 from LlmKira/dev Refactor to v4
Update Duckduckgo_search && Usability maintenance
Merge pull request #363 from LlmKira/dev Feature Pack
Llmkira Plugin OpenAPI
- 定义插件对系统的交互
- 修复插件的循环引用
- 修复重新排序错误
- 可以同时定义
需要虚拟环境变量
和静默启动
属性。 - 提供 Ffmpeg 工具类将任意语音转换为 Telegram Voice
Update Pydantic && Update Ci
Fix
- update pydantic
- fix some bug
- Fix Urllib exception Expose the endpoint
- Fix aps #342
- Support create shecma from pydantic
- Fix ci
Change
- Pin Docker to the last available version/固定 Docker 为最后可用版本
- remove
requirements.txt
Full Changelog: nolib0.27.0_5...lib0.27.1
Fix Docker Ci
nolib0.27.0_5 feat(docker):fixed version
Update Pydantic && Update Ci
:sparkles: feat: Add poetry==1.6.1 to Dockerfile - Update Dockerfile to install poetry==1.6.1 alongside pm2 and npm
Update Pydantic && Update Ci
nolib0.27.0_3 :construction: chore(Dockerfile): update poetry install command
Fix Docker Ci
:hammer: chore(ci): update Docker actions versions - Updated the version of `docker/setup-qemu-action` from v2 to v3 - Updated the version of `docker/setup-buildx-action` from v2 to v3 - Updated the version of `docker/login-action` from v2.1.0 to the latest version This commit updates the versions of the Docker actions used in the CI workflow to their latest versions for better compatibility and performance.
0.27.0 Support GPT-4 Turbo with vision
⚠️ Broken Change
This version passes basic testing and syntax checking, but unexpected things may happen.
此版本通过基本测试和语法检查,但可能有意料之外的情况发生。
📦 Change
- 大多数接口被重构,PluginOpenapi 版本升级到
20231111
- 支持新的函数并行处理
- 支持 GPT-4 Turbo with vision 图传
- 适配性增强
- 修复 Bug
- 优化了传递速度
- 定制函数规范
- Standardized plugin system
📦 Change
- Most interfaces have been refactored, and the PluginOpenapi version has been upgraded to
20231111
- Support new function parallel processing(tools call)
- Support GPT-4 Turbo with vision
- Adaptability enhancement
- Bug fixes
- Optimized delivery speed
- Custom function specification
- Standardized plugin system
📦 All models can customize Schema
Example from code
SCHEMA_GROUP.add_model(
models=[
SingleModel(
model_name="chatglm3",
token_limit=4096,
request=Openai,
response=OpenaiResult,
schema_type="openai",
func_executor="function_call",
exception=None
),
SingleModel(
model_name="chatglm3-16k",
token_limit=16384,
request=Openai,
response=OpenaiResult,
schema_type="openai",
func_executor="function_call",
exception=None
),
]
)
What's Changed
Full Changelog: nolib0.26.3_3...lib0.27.0