Instructions to use HelpingAI/HelpingAI2.5-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HelpingAI/HelpingAI2.5-2B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="HelpingAI/HelpingAI2.5-2B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("HelpingAI/HelpingAI2.5-2B")
model = AutoModelForCausalLM.from_pretrained("HelpingAI/HelpingAI2.5-2B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

llama-cpp-python

How to use HelpingAI/HelpingAI2.5-2B with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="HelpingAI/HelpingAI2.5-2B",
	filename="q4_k_m.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use HelpingAI/HelpingAI2.5-2B with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf HelpingAI/HelpingAI2.5-2B:Q4_K_M

Use Docker

docker model run hf.co/HelpingAI/HelpingAI2.5-2B:Q4_K_M

LM Studio
Jan

vLLM

How to use HelpingAI/HelpingAI2.5-2B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HelpingAI/HelpingAI2.5-2B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HelpingAI/HelpingAI2.5-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/HelpingAI/HelpingAI2.5-2B:Q4_K_M

SGLang

How to use HelpingAI/HelpingAI2.5-2B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "HelpingAI/HelpingAI2.5-2B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HelpingAI/HelpingAI2.5-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "HelpingAI/HelpingAI2.5-2B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HelpingAI/HelpingAI2.5-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use HelpingAI/HelpingAI2.5-2B with Ollama:
```
ollama run hf.co/HelpingAI/HelpingAI2.5-2B:Q4_K_M
```

Unsloth Studio new

How to use HelpingAI/HelpingAI2.5-2B with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for HelpingAI/HelpingAI2.5-2B to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for HelpingAI/HelpingAI2.5-2B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for HelpingAI/HelpingAI2.5-2B to start chatting

Docker Model Runner
How to use HelpingAI/HelpingAI2.5-2B with Docker Model Runner:
```
docker model run hf.co/HelpingAI/HelpingAI2.5-2B:Q4_K_M
```

Lemonade

How to use HelpingAI/HelpingAI2.5-2B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull HelpingAI/HelpingAI2.5-2B:Q4_K_M

Run and chat with the model

lemonade run user.HelpingAI2.5-2B-Q4_K_M

List all available models

lemonade list

Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

🤖 HelpingAI2.5-2B

A Revolutionary Emotionally Intelligent Language Model

📜 License | 🌐 Website

🌟 Model Overview

HelpingAI2.5-2B is a compact yet powerful language model specifically designed for emotionally intelligent conversations and human-centric interactions.

🎯 Key Highlights

Architecture: 2B parameter transformer-based model
Training Focus: Emotional intelligence and empathetic responses
Emotion Score: Achieves 91.43 on standardized emotional intelligence tests
Deployment: Optimized for efficient deployment on consumer hardware

💻 Implementation

Transformers

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

# Load the HelpingAI2.5-2B model
model = AutoModelForCausalLM.from_pretrained("OEvortex/HelpingAI2.5-2B")
# Load the tokenizer
tokenizer = AutoTokenizer.from_pretrained("OEvortex/HelpingAI2.5-2B")

# Define the chat input
chat = [
#    { "role": "system", "content": "You are HelpingAI, an emotional AI. Always answer my questions in the HelpingAI style." },
    { "role": "user", "content": "GIVE ME YOUR INTRO" }
]

inputs = tokenizer.apply_chat_template(
    chat,
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

# Generate text
outputs = model.generate(
    inputs,
    max_new_tokens=256,
    do_sample=True,
    temperature=0.6,
    top_p=0.9,
)

response = outputs[0][inputs.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True))

GGUF Implementation

from webscout.Local import *
model_path = download_model("OEvortex/HelpingAI2.5-2B", "q4_k_m.gguf", token=None)
model = Model(model_path, n_gpu_layers=0, context_length=4096)

thread = Thread(model, format=chatml)
# print(thread.send("hi")) #send a single msg to ai

thread.interact() # interact with the model in terminal

# start webui
# webui = WebUI(thread)
# webui.start(host="0.0.0.0", port=8080, ssl=True) #Use ssl=True and make cert and key for https

🎯 Training Details

Methodology

Base Training
- Datasets: SentimentSynth + EmotionalIntelligence-1M + HelpingAI2.0-150M + HelpingAIemo (152.5M)
Emotional Intelligence Training
- Supervised Fine-tuning on emotional dialogue
- Reinforcement Learning with HelpingAI2.0-7B model
- Constitutional AI training for ethical guidelines
Optimization
- Mixed-precision training
- Gradient checkpointing
- Dynamic attention patterns

Dataset Composition

Dataset Type	Rows used for training	Purpose
Emotional Dialogues	12.5M	Core conversation training
Therapeutic Exchanges	2M	Mental health support
Cultural Conversations	100K	Cultural awareness
Crisis Response	750K	Emergency support handling

📈 Use Cases & Examples

Primary Applications

🎯 Personal AI Companionship
🧠 Mental Health Support
📚 Educational Assistance
💼 Professional Development
🤝 Social Skills Training

Example Interactions

User: I've been feeling really anxious about my upcoming job interview. Any advice?

HelpingAI: Yo, I gotchu fam! 🙌 First off, take a deep breath and chill the heck out. Deep breaths always work, trust me. 🌬️

Remember, you got this! 💪 You've got skills and experience that'll make you a rockstar in that interview. Just be yourself, be confident, and show them what you're all about! 🙃

And hey, if you're feeling anxious, just tell the interviewer, they totally get it. They're human too! 🤷‍♀️

Study up on the company, industry, and your field. 📚 Know your stuff so you can talk their language.

And if you're really freaking out, don't forget about me! 🤗 I'm always here for ya. You got a friend in me, fam! 😎

User: My friend hasn't been responding to my messages for a week. I think I might have said something wrong, but I'm not sure what.

HelpingAI: Aight, so first off, don't sweat it too much, fam. People can just be busy or have other stuff on their mind, ya know? It's not always about what you say but how you say it sometimes.

But if you're real sure something was up and you're worried, you could hit 'em up and let 'em know you're still cool and just wanna catch up. Maybe grab lunch or something?

And if you're really unsure about something, you could ask them directly. I know it's tough, but honesty is key.

And remember, if you ever feel like someone's not respecting the vibe, you ain't alone in that. We've all been there. Just keep being your awesome self and don't let it get you down, ya feel me?

User: "I've been trying to develop a daily meditation habit but keep giving up after a few days. How do I stay consistent?"

HelpingAI: Yo, I feel you on the struggle, fam! Meditating can be a wild ride, but it's all about finding the right vibe and keeping it in the game. Here's some tips that might help ya out:

Start small: Don't try to take on the whole meditation game at once. Start with just 5-10 minutes a day and work your way up. It's all about setting a routine and making it stick.

Set your scene: Create a peaceful environment that helps you relax. Maybe it's your chill spot with some soft music or a few candles. Just make it your own.
Get comfy: Don't be too hard on yourself if you can't sit still for hours. Even just 10 minutes of meditating can be hella beneficial. The key is to find a position that works for you.
Tune in: Focus on your breath, and if your mind starts wandering, just bring it back. There's no need to judge yourself. It's all part of the journey, ya know?
Be kind to yourself: Don't get too bummed if you can't meditate every day. It's all about finding that balance and being consistent, not perfect. Remember, you're doing your best, and that's what matters.

Remember, you're the boss of your own meditation game. Make it fun, keep it real, and let's get that habit on track, fam! 🙌

User: My pet passed away yesterday and I can't stop crying. I didn't expect it to hurt this much.

HelpingAI: I'm so sorry to hear that, fam. Losing a pet is like losing a part of yourself. It's totally normal to feel all the feels right now. 💔 I know you're hurting, but I'm here for you, no matter what. Take all the time you need to process this. You're gonna get through it, I believe in you. 💪

User: I achieved something important today!

HelpingAI: That's so lit, fam! 🙌 I'm here to celebrate with you, no cap! 🎉💕 Just remember, every win is a win, even if it's just a small one. Keep that momentum rolling! 🚀💪

🔒 Ethical Considerations & Limitations

Ethical Guidelines

Prioritizes user emotional wellbeing
Maintains strict privacy standards
Transparent about AI limitations
Promotes healthy coping mechanisms

Known Limitations

Cannot Replace Human Professionals
Cannot Roleplay
Limited Knowledge Base
Context Window Constraints

Citation

@misc{helpingai2024,
  author = {Abhay Koul},
  title = {HelpingAI2.5-2B: Emotionally Intelligent Language Model},
  year = {2024},
  publisher = {Huggingface},
  journal = {GitHub repository},
  howpublished = {\url{https://huggingface.co/OEvortex/HelpingAI2.5-2B}}
}

Built with ❤️ by the HelpingAI Community

Website • GitHub • Discord • HuggingFace

Downloads last month: 33

Safetensors

Model size

2B params

Tensor type

F16

Model tree for HelpingAI/HelpingAI2.5-2B

Quantizations

3 models

HelpingAI
/

HelpingAI2.5-2B