📚 docs: Separate LiteLLM and Ollama Documentation (#1948)

* Separate LiteLLM and Ollama Documentation

* Clarify Ollama Setup

* Fix litellm config

docker-compose.override.yml.example

@@ -92,3 +92,69 @@ version: '3.4'
 #   meilisearch:
 #     ports:
 #       - 7700:7700
+
+# # ADD OLLAMA
+#  ollama:
+#    image: ollama/ollama:latest
+#    deploy:
+#      resources:
+#        reservations:
+#          devices:
+#            - driver: nvidia
+#              capabilities: [compute, utility]
+#    ports:
+#      - "11434:11434"
+#    volumes:
+#      - ./ollama:/root/.ollama
+
+# # ADD LITELLM BASIC - NEED TO CONFIGURE litellm-config.yaml, ONLY NEED ENV TO ENABLE REDIS FOR CACHING OR LANGFUSE FOR MONITORING
+#  litellm:
+#    image: ghcr.io/berriai/litellm:main-latest
+#    volumes:
+#      - ./litellm/litellm-config.yaml:/app/config.yaml
+#    command: [ "--config", "/app/config.yaml", "--port", "8000", "--num_workers", "8" ]
+#    environment:
+#      REDIS_HOST: redis
+#      REDIS_PORT: 6379
+#      REDIS_PASSWORD: RedisChangeMe
+#      LANGFUSE_PUBLIC_KEY: pk-lf-RandomStringFromLangfuseWebInterface
+#      LANGFUSE_SECRET_KEY: sk-lf-RandomStringFromLangfuseWebInterface
+#      LANGFUSE_HOST: http://langfuse-server:3000
+
+# # ADD LITELLM CACHING
+#  redis:
+#    image: redis:7-alpine
+#    command:
+#    - sh
+#    - -c # this is to evaluate the $REDIS_PASSWORD from the env
+#    - redis-server --appendonly yes --requirepass $$REDIS_PASSWORD ## $$ because of docker-compose
+#    environment:
+#      REDIS_PASSWORD: RedisChangeMe
+#    volumes:
+#    - ./redis:/data
+  
+# # ADD LITELLM MONITORING
+#  langfuse-server:
+#    image: ghcr.io/langfuse/langfuse:latest
+#    depends_on:
+#      - db
+#    ports:
+#      - "3000:3000"
+#    environment:
+#      - NODE_ENV=production
+#      - DATABASE_URL=postgresql://postgres:PostgresChangeMe@db:5432/postgres
+#      - NEXTAUTH_SECRET=ChangeMe
+#      - SALT=ChangeMe
+#      - NEXTAUTH_URL=http://localhost:3000
+#      - TELEMETRY_ENABLED=${TELEMETRY_ENABLED:-true}
+#      - NEXT_PUBLIC_SIGN_UP_DISABLED=${NEXT_PUBLIC_SIGN_UP_DISABLED:-false}
+#      - LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES=${LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES:-false}
+#  db:
+#    image: postgres
+#    restart: always
+#    environment:
+#      - POSTGRES_USER=postgres
+#      - POSTGRES_PASSWORD=PostgresChangeMe
+#      - POSTGRES_DB=postgres
+#    volumes:
+#      - ./postgres:/var/lib/postgresql/data

docs/install/configuration/ai_endpoints.md

@@ -259,5 +259,55 @@ Some of the endpoints are marked as **Known,** which means they might have speci
       forcePrompt: false
       modelDisplayLabel: "together.ai"
 ```
+## LiteLLM
+> LiteLLM API key: master_key value [LiteLLM](./litellm.md)
 
-![image](https://github.com/danny-avila/LibreChat/assets/32828263/fe3eae7a-d157-4f21-bb98-00688f261967)
+**Notes:**
+
+- Reference [LiteLLM](./litellm.md) for configuration.
+
+```yaml
+    - name: "LiteLLM"
+      apiKey: "sk-from-config-file"
+      baseURL: "http://localhost:8000/v1"
+      models:
+        default: ["gpt-3.5-turbo"]
+        fetch: true
+      titleConvo: true
+      titleModel: "gpt-3.5-turbo"
+      summarize: false
+      summaryModel: "gpt-3.5-turbo"
+      forcePrompt: false
+      modelDisplayLabel: "LiteLLM"
+```
+
+## Ollama
+> Ollama API key: Required but ignored - [Ollama OpenAI Compatibility](https://github.com/ollama/ollama/blob/main/docs/openai.md)
+
+**Notes:**
+
+- **Known:** icon provided.
+- **Known issue:** fetching list of models is not supported. See [Pull Request 2728](https://github.com/ollama/ollama/pull/2728).
+- Download models with ollama run command. See [Ollama Library](https://ollama.com/library)
+- The example includes a top 5 popular model list from the Ollama Library, which was last updated on March 1, 2024, for your convenience.
+
+```yaml
+    - name: "Ollama"
+      apiKey: "ollama"
+      baseURL: "http://localhost:11434/v1/"
+      models:
+        default: [
+          "llama2",
+          "mistral",
+          "codellama",
+          "dolphin-mixtral",
+          "mistral-openorca"
+          ]
+        fetch: false # fetching list of models is not supported
+      titleConvo: true
+      titleModel: "llama2"
+      summarize: false
+      summaryModel: "llama2"
+      forcePrompt: false
+      modelDisplayLabel: "Ollama"
+```

docs/install/configuration/litellm.md

@@ -1,5 +1,5 @@
 ---
-title: 🚅 LiteLLM and Ollama
+title: 🚅 LiteLLM
 description: Using LibreChat with LiteLLM Proxy 
 weight: -7
 ---
@@ -12,12 +12,18 @@ Use **[LiteLLM Proxy](https://docs.litellm.ai/docs/simple_proxy)** for:
 * Authentication & Spend Tracking Virtual Keys
 
 ## Start LiteLLM Proxy Server 
-### Pip install litellm 
+### 1. Uncomment desired sections in docker-compose.override.yml
-```shell
+The override file contains sections for the below LiteLLM features
-pip install litellm
-```
 
-### Create a config.yaml for litellm proxy 
+#### Caching with Redis
+Litellm supports in-memory, redis, and s3 caching. Note: Caching currently only works with exact matching.
+
+#### Performance Monitoring with Langfuse
+Litellm supports various logging and observability options.  The settings below will enable Langfuse which will provide a cache_hit tag showing which conversations used cache.
+
+### 2. Create a config.yaml for LiteLLM proxy 
+LiteLLM requires a configuration file in addition to the override file. The file 
+below has the options to enable llm proxy to various providers, load balancing, Redis caching, and Langfuse monitoring. Review documentation for other configuration options.
 More information on LiteLLM configurations here: **[docs.litellm.ai/docs/simple_proxy](https://docs.litellm.ai/docs/simple_proxy)**
 
 ```yaml
@@ -40,42 +46,29 @@ model_list:
       api_base: https://openai-france-1234.openai.azure.com/
       api_key: 
       rpm: 1440
+  - model_name: mixtral
+    litellm_params:
+      model: ollama/mixtral:8x7b-instruct-v0.1-q5_K_M
+      api_base: http://ollama:11434
+      stream: True
+  - model_name: mistral
+    litellm_params:
+      model: ollama/mistral
+      api_base: http://ollama:11434
+      stream: True
+litellm_settings:
+  success_callback: ["langfuse"]
+  cache: True
+  cache_params:
+    type: "redis"
+    supported_call_types: ["acompletion", "completion", "embedding", "aembedding"]
+general_settings:
+  master_key: sk_live_SetToRandomValue
 ```
 
-### Start the proxy
+### 3. Configure LibreChat
-```shell
-litellm --config /path/to/config.yaml
 
-#INFO: Proxy running on http://0.0.0.0:8000
+Use `librechat.yaml` [Configuration file (guide here)](./ai_endpoints.md) to add Reverse Proxies as separate endpoints.
-```
-
-## Use LiteLLM Proxy Server with LibreChat
-
-
-#### 1. Clone the repo
-```shell
-git clone https://github.com/danny-avila/LibreChat.git
-```
-
-
-#### 2. Modify Librechat's `docker-compose.yml`
-```yaml
-OPENAI_REVERSE_PROXY=http://host.docker.internal:8000/v1/chat/completions
-```
-
-**Important**: As of v0.6.6, it's recommend you use the `librechat.yaml` [Configuration file (guide here)](./custom_config.md) to add Reverse Proxies as separate endpoints.
-
-#### 3. Save fake OpenAI key in Librechat's `.env` 
-
-Copy Librechat's `.env.example` to `.env` and overwrite the default OPENAI_API_KEY (by default it requires the user to pass a key).
-```env
-OPENAI_API_KEY=sk-1234
-```
-
-#### 4. Run LibreChat: 
-```shell
-docker compose up
-```
 
 ---
 
@@ -102,162 +95,4 @@ Key components and features include:
 - **Deployment and Performance**: Information on deploying LiteLLM Proxy and its performance metrics.
 - **Proxy CLI Arguments**: A wide range of command-line arguments for customization.
 
-Overall, LiteLLM Server offers a comprehensive suite of tools for managing, deploying, and interacting with a variety of LLMs, making it a versatile choice for large-scale AI applications.
+Overall, LiteLLM Server offers a comprehensive suite of tools for managing, deploying, and interacting with a variety of LLMs, making it a versatile choice for large-scale AI applications.
-
-## Ollama
-Use [Ollama](https://ollama.ai/) for
-
-* Run large language models on local hardware
-* Host multiple models
-* Dynamically load the model upon request
-
-### GPU Acceleration
-
-- **Linux**: Requires a Linux distrubution support by official Nvidia drivers. [Nvidia CUDA Toolkit](https://developer.nvidia.com/cuda-downloads?target_os=Linux)
-- **Windows**: Requires Windows Subsytem for Linux. Follow Nvidia instructions at [Nvidia WSL User Guide](https://docs.nvidia.com/cuda/wsl-user-guide/index.html)
-- **macOS**: [macOS Ollama Download](https://ollama.ai/download/mac)
-
-### docker-compose.override.yml with GPU
-```yaml
-version: "3.8"
-services:
-  litellm:
-    image: ghcr.io/berriai/litellm:main-latest
-    volumes:
-      - ./litellm/litellm-config.yaml:/app/config.yaml
-    command: [ "--config", "/app/config.yaml", "--port", "8000", "--num_workers", "8" ]
-  ollama:
-    image: ollama/ollama
-    deploy:
-      resources:
-        reservations:
-          devices:
-            - driver: nvidia
-              capabilities: [compute, utility]
-    ports:
-      - "11434:11434"
-    volumes:
-      - ./ollama:/root/.ollama
-
-```
-
-### Loading Models in Ollama
-1. Browse the available models at [Ollama Library](https://ollama.ai/library)
-2. Run ```docker exec -it ollama /bin/bash```
-3. Copy the text from the Tags tab from the library website. It should begin with 'ollama run'
-4. Check model size. Models that can run in GPU memory perform the best.
-5. Use /bye to exit the terminal
-
-### Litellm Ollama Configuration
-Add the below lines to the config to access the Ollama models
-```yaml
-  - model_name: mixtral
-    litellm_params:
-      model: ollama/mixtral:8x7b-instruct-v0.1-q5_K_M
-      api_base: http://ollama:11434
-      stream: True
-  - model_name: mistral
-    litellm_params:
-      model: ollama/mistral
-      api_base: http://ollama:11434
-      stream: True
-```
-
-## Caching with Redis
-Litellm supports in-memory, redis, and s3 caching. Note: Caching currently only works with exact matching.
-
-### Update docker-compose.override.yml to enable Redis
-Add the below service to your docker-compose.override.yml
-```yaml
-  redis:
-    image: redis:7-alpine
-    command:
-    - sh
-    - -c # this is to evaluate the $REDIS_PASSWORD from the env
-    - redis-server --appendonly yes --requirepass $$REDIS_PASSWORD ## $$ because of docker-compose
-    environment:
-      REDIS_PASSWORD: RedisChangeMe
-    volumes:
-    - ./redis:/data
-```
-
-Add the following to the environment variables in the litellm service inside the docker-compose.override.yml
-```yaml
-  litellm:
-    image: ghcr.io/berriai/litellm:main-latest
-    volumes:
-      - ./litellm/litellm-config.yaml:/app/config.yaml
-    command: [ "--config", "/app/config.yaml", "--port", "8000", "--num_workers", "8" ]
-    environment:
-      REDIS_HOST: redis
-      REDIS_PORT: 6379
-      REDIS_PASSWORD: RedisChangeMe
-```
-
-### Update Litellm Config File
-Add the below options to the litellm config file
-```yaml
-litellm_settings: # module level litellm settings - https://github.com/BerriAI/litellm/blob/main/litellm/__init__.py
-  cache: True          # set cache responses to True, litellm defaults to using a redis cache
-  cache_params:         # cache_params are optional
-    type: "redis"  # The type of cache to initialize. Can be "local" or "redis". Defaults to "local".
-
-    # Optional configurations
-    supported_call_types: ["acompletion", "completion", "embedding", "aembedding"] # defaults to all litellm call types
-```
-
-
-## Performance Monitoring with Langfuse
-Litellm supports various logging and observability options.  The settings below will enable Langfuse which will provide a cache_hit tag showing which conversations used cache.
-
-### Update docker-compose.override.yml to enable Langfuse
-Langfuse requires a postgres database, so add both postgres and langfuse services to the docker-compose.override.yml
-```yaml
-  langfuse-server:
-    image: ghcr.io/langfuse/langfuse:latest
-    depends_on:
-      - db
-    ports:
-      - "3000:3000"
-    environment:
-      - NODE_ENV=production
-      - DATABASE_URL=postgresql://postgres:PostgresChangeMe@db:5432/postgres
-      - NEXTAUTH_SECRET=ChangeMe
-      - SALT=ChangeMe
-      - NEXTAUTH_URL=http://localhost:3000
-      - TELEMETRY_ENABLED=${TELEMETRY_ENABLED:-true}
-      - NEXT_PUBLIC_SIGN_UP_DISABLED=${NEXT_PUBLIC_SIGN_UP_DISABLED:-false}
-      - LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES=${LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES:-false}
-
-  db:
-    image: postgres
-    restart: always
-    environment:
-      - POSTGRES_USER=postgres
-      - POSTGRES_PASSWORD=PostgresChangeMe
-      - POSTGRES_DB=postgres
-    volumes:
-      - ./postgres:/var/lib/postgresql/data
-```
-
-Once Langfuse is running, create an account by accessing the web interface on port 3000. Create a new project to obtain the needed public and private key used by the litellm config
-Add environement variable within the litellm service within docker-compose.override.yml
-```yaml
-  litellm:
-    image: ghcr.io/berriai/litellm:main-latest
-    ports:
-      - "8000:8000"
-    volumes:
-      - /srv/litellm/config/litellm-config.yaml:/app/config.yaml
-    command: [ "--config", "/app/config.yaml", "--port", "8000", "--num_workers", "8" ]
-    environment:
-      LANGFUSE_PUBLIC_KEY: pk-lf-RandomStringFromLangfuseWebInterface
-      LANGFUSE_SECRET_KEY: sk-lf-RandomStringFromLangfuseWebInterface
-      LANGFUSE_HOST: http://langfuse-server:3000
-```
-
-### Update litellm config file
-```yaml
-litellm_settings:
-  success_callback: ["langfuse"]
-```

docs/install/configuration/ollama.md

@@ -0,0 +1,29 @@
+---
+title: 🚅 Ollama
+description: Using LibreChat with Ollama 
+weight: -6
+---
+## Ollama
+Use [Ollama](https://ollama.ai/) for
+
+* Running large language models on local hardware
+* Hosting multiple models
+* Dynamically loading the model upon request
+
+### 1. Install Ollama
+#### Mac, Linux, Windows Install
+Ollama supports GPU acceleration on Nvidia, AMD, and Apple Metal. Follow Instructions at [Ollama Download](https://ollama.com/download) 
+
+#### Docker Install
+Reference docker-compose.override.yml.example for configuration of Ollama in a Docker environment.
+
+Run ```docker exec -it ollama /bin/bash``` to access the Ollama command within the container.
+
+### 2. Load Models in Ollama
+1. Browse the available models at [Ollama Library](https://ollama.ai/library)
+2. Copy the text from the Tags tab from the library website and paste it into the terminal. It should begin with 'ollama run'
+3. Check model size. Models that can run in GPU memory perform the best.
+4. Use /bye to exit the terminal
+
+### 3. Configure LibreChat
+Use `librechat.yaml` [Configuration file (guide here)](./ai_endpoints.md) to add Ollama as a separate endpoint.