docs: refactor import model documentation
This commit is contained in:
Ho Duc Hieu
parent
f323447664
commit
1622b4e686
@ -24,16 +24,29 @@ This is currently under development.
|
||||
import Tabs from "@theme/Tabs";
|
||||
import TabItem from "@theme/TabItem";
|
||||
|
||||
Jan is compatible with all GGUF models.
|
||||
In this guide, we will walk you through how to import models manually. In Jan, you can use a local model directly on your computer or connect to a remote server.
|
||||
|
||||
If you can not find the model you want in the Hub or have a custom model you want to use, you can import it manually.
|
||||
- Local Model: Jan is compatible with all GGUF models. If you can not find the model you want in the Hub or have a custom model you want to use, you can import it manually by following the [Steps to Manually Import a Local Model](#steps-to-manually-import-a-local-model) section.
|
||||
|
||||
- Remote Model: Jan also supports integration with remote models. To establish a connection with these remote model, you can configure the client connection to a remote/ local server by following the [OpenAI Platform Configuration](#openai-platform-configuration) or [Engines with OAI Compatible Configuration](#engines-with-oai-compatible-configuration) section. Please note that at the moment, you can only connect to one OpenAI compatible at a time (e.g OpenAI Platform, Azure OpenAI, LM Studio, etc).
|
||||
|
||||
```mermaid
|
||||
graph TB
|
||||
Model --> LocalModel[Local model]
|
||||
Model --> RemoteModel[Remote model]
|
||||
LocalModel[Local Model] --> NitroEngine[Nitro Engine]
|
||||
RemoteModel[Remote Model] --> OpenAICompatible[OpenAI Compatible]
|
||||
|
||||
OpenAICompatible --> OpenAIPlatform[OpenAI Platform]
|
||||
OpenAICompatible --> OAIEngines[Engines with OAI Compatible: Jan server, Azure OpenAI, LM Studio, vLLM, etc]
|
||||
```
|
||||
|
||||
## Steps to Manually Import a Local Model
|
||||
|
||||
In this guide, we will show you how to import a GGUF model from [HuggingFace](https://huggingface.co/), using our latest model, [Trinity](https://huggingface.co/janhq/trinity-v1-GGUF), as an example.
|
||||
|
||||
> We are fast shipping a UI to make this easier, but it's a bit manual for now. Apologies.
|
||||
|
||||
## Steps to Manually Import a Model
|
||||
|
||||
### 1. Create a Model Folder
|
||||
|
||||
Navigate to the `~/jan/models` folder. You can find this folder by going to `App Settings` > `Advanced` > `Open App Directory`.
|
||||
@ -168,30 +181,13 @@ Restart Jan and navigate to the Hub. Locate your model and click the `Download`
|
||||
|
||||
Your model is now ready to use in Jan.
|
||||
|
||||
## Configuring Client Connection to Remote/Local Server
|
||||
## OpenAI Platform Configuration
|
||||
|
||||
In this guide, we will show you how to configure a client connection to a remote/local server, using LM Studio as an example.
|
||||
In this guide, we will show you how to configure with OpenAI Platform, using the OpenAI GPT 3.5 Turbo 16k model as an example.
|
||||
|
||||
At the moment, you can only connect to one compatible server at a time (e.g OpenAI Platform, Azure OpenAI, LM Studio, etc).
|
||||
### 1. Create a Model JSON
|
||||
|
||||
### 1. Configure Local Server in Engine
|
||||
|
||||
Navigate to the `~/jan/engines` folder and modify the `openai.json` file. Please note that at the moment the code supports any openai compatible endpoint only read `engine/openai.json` file, thus, it will not search any other files in this directory.
|
||||
|
||||
Configure `full_url` properties with the endpoint server that you want to connect. For example, if you want to connect to LM Studio, you can configure as follows:
|
||||
|
||||
```js
|
||||
{
|
||||
// highlight-next-line
|
||||
"full_url": "http://<REMOTE_LMSTUDIO_IP>:<REMOTE_LMSTUDIO_PORT>/v1/chat/completions",
|
||||
// Skip api_key if your local server does not require authentication
|
||||
// "api_key": "sk-<your key here>"
|
||||
}
|
||||
```
|
||||
|
||||
### 2. Create a Model JSON
|
||||
|
||||
Navigate to the `~/jan/models` folder. Create a folder named `remote-lmstudio` and create a `model.json` file inside the folder including the following configurations:
|
||||
Navigate to the `~/jan/models` folder. Create a folder named `gpt-3.5-turbo-16k` and create a `model.json` file inside the folder including the following configurations:
|
||||
|
||||
- Ensure the filename must be `model.json`.
|
||||
- Ensure the `id` property matches the folder name you created.
|
||||
@ -201,19 +197,90 @@ Navigate to the `~/jan/models` folder. Create a folder named `remote-lmstudio` a
|
||||
|
||||
```js
|
||||
{
|
||||
"source_url": "https://lmstudio.ai",
|
||||
"source_url": "https://openai.com",
|
||||
// highlight-next-line
|
||||
"id": "remote-lmstudio",
|
||||
"id": "gpt-3.5-turbo-16k",
|
||||
"object": "model",
|
||||
"name": "remote lmstudio",
|
||||
"name": "OpenAI GPT 3.5 Turbo 16k",
|
||||
"version": "1.0",
|
||||
"description": "Jan integration with remote LMstudio server",
|
||||
"description": "OpenAI GPT 3.5 Turbo 16k model is extremely good",
|
||||
// highlight-start
|
||||
"format": "api",
|
||||
"settings": {},
|
||||
"parameters": {},
|
||||
"metadata": {
|
||||
"author": "OpenAI",
|
||||
"tags": ["General", "Big Context Length"]
|
||||
},
|
||||
"engine": "openai",
|
||||
"state": "ready"
|
||||
// highlight-end
|
||||
}
|
||||
```
|
||||
|
||||
### 2. Configure OpenAI API Keys
|
||||
|
||||
You can find your API keys in the [OpenAI Platform](https://platform.openai.com/api-keys) and set the OpenAI API keys in `~/jan/engines/openai.json` file.
|
||||
|
||||
```js
|
||||
{
|
||||
"full_url": "https://api.openai.com/v1/chat/completions",
|
||||
// highlight-next-line
|
||||
"api_key": "sk-<your key here>"
|
||||
}
|
||||
```
|
||||
|
||||
### 3. Start the Model
|
||||
|
||||
Restart Jan and navigate to the Hub. Then, select your configured model and start the model.
|
||||
|
||||
|
||||
|
||||
## Engines with OAI Compatible Configuration
|
||||
|
||||
In this guide, we will show you how to configure a client connection to a remote/local server, using Jan API local server as an example.
|
||||
|
||||
### 1. Configure Local Server in Engine
|
||||
|
||||
Navigate to the `~/jan/engines` folder and modify the `openai.json` file. Please note that at the moment the code supports any openai compatible endpoint only read `engine/openai.json` file, thus, it will not search any other files in this directory.
|
||||
|
||||
Configure `full_url` properties with the endpoint server that you want to connect. For example, if you want to connect to Jan API local server, you can configure as follows:
|
||||
|
||||
```js
|
||||
{
|
||||
// highlight-next-line
|
||||
"full_url": "http://localhost:1337/v1/chat/completions",
|
||||
// Skip api_key if your local server does not require authentication
|
||||
// "api_key": "sk-<your key here>"
|
||||
}
|
||||
```
|
||||
|
||||
### 2. Create a Model JSON
|
||||
|
||||
Navigate to the `~/jan/models` folder. Create a folder named `remote-jan` and create a `model.json` file inside the folder including the following configurations:
|
||||
|
||||
- Ensure the filename must be `model.json`.
|
||||
- Ensure the `id` property matches the folder name you created.
|
||||
- Ensure the `format` property is set to `api`.
|
||||
- Ensure the `engine` property is set to `openai`.
|
||||
- Ensure the `state` property is set to `ready`.
|
||||
|
||||
```js
|
||||
{
|
||||
"source_url": "https://jan.ai",
|
||||
// highlight-next-line
|
||||
"id": "remote-jan",
|
||||
"object": "model",
|
||||
"name": "remote jan",
|
||||
"model": "tinyllama-1.1b",
|
||||
"version": "1.0",
|
||||
"description": "Jan integration with remote Jan API server",
|
||||
// highlight-next-line
|
||||
"format": "api",
|
||||
"settings": {},
|
||||
"parameters": {},
|
||||
"metadata": {
|
||||
"author": "LMstudio",
|
||||
"author": "Jan",
|
||||
"tags": [
|
||||
"remote",
|
||||
"awesome"
|
||||
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 372 KiB |
Loading…
x
Reference in New Issue
Block a user