Skip to content

vs4vijay/AI-Playground

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

36 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI-Playground


AI Playground for trying out LLM Models, Embeddings, Vector Stores, Semantic Search, RAG, Azure OpenAI, LLaMa, Mistral

Installation

pip install -U ai-playground

Local Installation

Pre-requisites:

  • Python 3.10+ and pip
# Start virtual environment
source ./activate

# Install requirements
pip install -r requirements.txt

Running the full playground

  • Copy .env.example to .env and fill in the values

  • Run the following command to start the server

python ai_playground.py

Models

wget -c https://huggingface.co/TheBloke/Mistral-7B-v0.1-GGUF/resolve/main/mistral-7b-v0.1.Q8_0.gguf wget -c https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF/resolve/main/mistral-7b-instruct-v0.1.Q8_0.gguf wget -c https://huggingface.co/TheBloke/Mistral-7B-OpenOrca-GGUF/resolve/main/mistral-7b-openorca.Q8_0.gguf

wget -c https://huggingface.co/TheBloke/Wizard-Vicuna-7B-Uncensored-GGUF/resolve/main/Wizard-Vicuna-7B-Uncensored.Q8_0.gguf

wget -c https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF/resolve/main/codellama-7b-instruct.Q8_0.gguf


Run Local LLM

Try with LLaMA.cpp

  • Extract LLaMA.cpp zip to bin/ directory
./bin/main.exe -m models/llama-2-7b-chat.Q8_0.gguf

Try with ollama

echo "FROM ./models/llama-2-13b-chat.Q5_K_M.gguf" > llama-2-13b-chat.Modelfile

ollama create llama2-13b-chat -f ./llama-2-13b-chat.Modelfile

ollama run llama2-13b-chat
# OR
ollama run llama3.1:7b
ollama serve llama3.1:7b

ollama ps

curl -X POST http://localhost:11434/api/generate -d '{
    "model": "wizardlm-uncensored",
    "prompt":"Who made Rose promise that she would never let go?"
}'

ollama run starcoder2:3b
ollama pull nomic-embed-text

Try with vLLM

pip install -U vllm

python -u -m vllm.entrypoints.openai.api_server --host 0.0.0.0 --model mistralai/Mistral-7B-v0.1

Try with FastChat

pip install -U fastchat

python -m fastchat.serve.openai_api_server --host localhost --port 8000

Try with LeptonAI

pip install -U leptonai

Use StarCoder

pip install transformers pip install torch torchvision pip install accelerate bitsandbytes pip install accelerate[torch]

Edit:

  • load_in_8bit=True

python starcoder.py

  • will download ~60 GB of model

Specs

RAM Required:

Model Size RAM Required
3B 8 GB
7B 16 GB
13B 32 GB

Chat UIs


Code Assistants / Copilots


Agents


Other Tools


Development Notes

pip install pyautogen

pip install openplayground
openplayground run

ollama run mistral

pip install -U jina

Ray Serve
pip install "ray[serve]"
https://github.com/ray-project/ray-llm

txtai

MLC AI - https://mlc.ai/package/
pip install --pre --force-reinstall mlc-ai-nightly mlc-chat-nightly -f https://mlc.ai/wheels
python -m mlc_chat.rest 

OpenLLM


https://github.com/FlowiseAI/Flowise


wget https://gpt4all.io/models/ggml-gpt4all-j.bin -O models/ggml-gpt4all-j

https://github.com/go-skynet/LocalAI
docker pull quay.io/go-skynet/local-ai:latest

nlpcloud

curl "https://api.nlpcloud.io/v1/<model_name>/entities" \
  -H "Authorization: Token <token>" \
  -H "Content-Type: application/json" \
  -X POST \
  -d '{"text":"John Doe has been working for Microsoft in Seattle since 1999."}'


https://github.com/microsoft/semantic-kernel
https://github.com/microsoft/guidance


https://skypilot.readthedocs.io/

Later:
https://github.com/Arize-ai/phoenix
https://github.com/explodinggradients/ragas
https://github.com/trypromptly/LLMStack


Q5_K_M



poetry export -f requirements.txt --output requirements.txt
poetry publish --build


lazypredict

mito

pip install langchain-serve

LangServe

pip install -U "langserve[all]"
pip install -U langchain-cli


langflow run


flowise

promptflow
pip install promptflow promptflow-tools


PyGWalker

# DSPy
pip install dspy-ai





https://github.com/ShreyaR/guardrails
https://github.com/guardrails-ai/guardrails



guidance
https://github.com/1rgs/jsonformer

LangChain
https://github.com/jina-ai/langchain-serve

LangFlow / Flowise / LangSmith
ChainLit

promptflow


LMQI
https://github.com/eth-sri/lmql

https://github.com/zilliztech/GPTCache

https://github.com/argilla-io/argilla

https://github.com/vllm-project/vllm

https://github.com/TransformerOptimus/SuperAGI

accelerate
  - accelerate config
  - accelerate env
bitsandbytes
wand
https://github.com/huggingface/text-generation-inference


ctransformers

spacy
spacy-llm
gorilla-cli
https://github.com/langgenius/dify
gptcache

https://github.com/wandb/openui