diff --git a/README.md b/README.md index a08bbe5..bbdc159 100644 --- a/README.md +++ b/README.md @@ -4,7 +4,7 @@ ## Description -[Obsidian](https://obsidian.md) exists to help you build a second brain. Gladdis exists to give that brain a pen (let it do the writing for once), ears (talking is easier than typing) and multiple split personalities (give your system prompts names). Further plans for v1.0 include: a brush (let it paint like a true artist), a voice (listening is easier than reading), an internal monologue (together we are more than our sum) and the ability to execute code (what could possibly go wrong ?). +[Obsidian](https://obsidian.md) exists to help you build a second brain. Gladdis exists to give that brain a pen (let it do the writing for once), ears (talking is easier than typing) and multiple split personalities (give your system prompts names). Further plans for v1.0 include: a brush (let it paint like an artist), a voice (listening is easier than reading), an internal monologue (together we are more than our sum) and the ability to execute code (what could possibly go wrong ?). ### Current Features @@ -12,6 +12,7 @@ - **Whisper Transcription**: audio files WikiLinks are auto-transcribed, use the Obsidian "audio recorder" core plugin. - **Document Loading**: text files WikiLinks are auto-loaded, just drag and drop files into the conversation window. - **PDF & Web Support**: load PDF WikiLinks just like text files, load web pages external links as well using ``. +- **Image Vision AIs**: send images to vision supporting LLMs using local WikiLinks or Markdown external web links. - **System Prompts**: name your system prompts, create personalities and use FrontMatter to choose whom to talk to. ## Documentation @@ -39,16 +40,16 @@ The defaults from the settings panel can be overridden using FrontMatter in any ```yaml --- gladdis: - label: Gladdis # name the AI in the config - config: Gladdis # switch the AI you talk to - model: gpt-3.5-turbo # switch the model at any time + config: Gladdis # select which AI you talk to + model: gpt-4o # choose the current AI model + server: http:// # use local or custom servers temperature: 42 # test different temperatures - top_p_param: 100 # or various probability mass + top_p_param: 88 # or various probability mass whisper: - config: Whisper # switch transcription language - model: whisper-1 # only one model at the moment + config: Whisper # switch transcription config + model: whisper-1 # change the transcription AI temperature: 24 # test different temperatures - echoOutput: true # output transcription callout + echoOutput: true # write transcription callout deleteFile: false # cleanup after transcription --- ``` @@ -82,16 +83,26 @@ The very last message is the prompt, the LLM response streams below. ### Available Models -Only the OpenAI models are supported at the moment, however more APIs and local models are coming. +The following [OpenAI models](https://platform.openai.com/docs/models) are supported via the paying API: -| Label | Token Limit | Notes | -| ----- | :-----------: | ----- | -| `gpt-4-turbo-preview` | 128 k | The most powerful and expensive. | -| `gpt-4-32k` | 32 k | The old GPT-4, with large context. | -| `gpt-4` | 8 k | The original GPT-4, from June 2023. | -| `gpt-3.5-turbo` | 16 k | The updated GPT-3.5, cheaper than 4. | -| `gpt-3.5-turbo-16k` | 16 k | The old GPT-3.5, with large context. | -| `gpt-3.5-turbo-0613` | 4 k | The original GPT-3.5, from June 2023. | +| Label | Tokens | Vision | Notes | +| ----- | :----: | :----: | ----- | +| `gpt-4o` | 128 k | ✅ | OpenAI's high-intelligence flagship model for complex, multi-step tasks. | +| `gpt-4o-mini` | 128 k | ✅ | OpenAI's affordable and intelligent small model for fast, lightweight tasks. | +| `chatgpt-4o-latest` | 128 k | ✅ | Dynamic model continuously updated to the current version of GPT-4o in ChatGPT. | +| `gpt-4-turbo` | 128 k | ✅ | The latest GPT-4 Turbo model with vision capabilities (GPT-4o is cheaper and faster). | +| `gpt-3.5-turbo` | 16 k | | The latest GPT-3.5 Turbo model with higher accuracy (GPT-4o mini is cheaper and more capable). | + +Open Source models can be used locally via [Ollama](https://ollama.com/) or [LocalAI](https://localai.io/): + +| Label | Tokens | Vision | Notes | +| ----- | :----: | :----: | ----- | +| `llama3.2` | 128 k | | Meta's new 1B and 3B models optimized for multilingual dialogue use cases. | +| `llama3.1` | 128 k | | Meta's 8B, 70B and 405B models with overall stronger reasoning capabilities. | +| `phi3.5` / `phi3` | 128 k | | Microsoft's new Phi 3.5 "Mini" (3.8B) and Phi 3 "Medium" (14B) open models. | +| `qwen2.5` | 128 k | | Alibaba Cloud's Qwen multilingual Chinese models ranging from 0.5B to 72B. | +| `gemma2` | 128 k | | Google's Gemma 2B, 9B and 27B models featuring a brand new architecture. | +| `llava` | 4 k | ✅ | Family of Large Language-and-Vision Assistant models between 7B and 34B. | ## Future Development @@ -110,7 +121,7 @@ Only the OpenAI models are supported at the moment, however more APIs and local ### With the BRAT plugin 1. Install the [Beta Reviewers Auto-update Tool plugin](https://github.com/TfTHacker/obsidian42-brat). -2. Follow the instructions for ["Adding a beta plugin"](https://tfthacker.com/Obsidian+Plugins+by+TfTHacker/BRAT+-+Beta+Reviewer's+Auto-update+Tool/Quick+guide+for+using+BRAT#Adding+a+beta+plugin). +2. Follow the instructions for ["Adding a beta plugin"](https://tfthacker.com/brat-quick-guide#Adding+a+beta+plugin). ### Manually from GitHub