đ HelpingAI3
HelpingAI3 is an advanced language model crafted for emotionally intelligent conversations. It builds on the success of HelpingAI2.5, offering improved emotional understanding and contextual awareness.
đ Quick Start
Here's a basic example of using the transformers
library to interact with the HelpingAI3 model:
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("HelpingAI/HAI3-RAW")
tokenizer = AutoTokenizer.from_pretrained("HelpingAI/HAI3-RAW")
chat = [
{"role": "system", "content": "You are HelpingAI, an emotional AI. Always answer my questions in the HelpingAI style."},
{"role": "user", "content": "Introduce yourself."}
]
inputs = tokenizer.apply_chat_template(
chat,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
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))
⨠Features
- Emotionally Intelligent Conversations: Excels in understanding and responding to emotions.
- Enhanced Contextual Awareness: Offers more relevant and context-aware responses.
- Multiple Use Cases: Suitable for AI companionship, therapy guidance, personalized learning, and professional assistance.
đĻ Installation
The README doesn't provide specific installation steps, so this section is skipped.
đģ Usage Examples
Basic Usage
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("HelpingAI/HAI3-RAW")
tokenizer = AutoTokenizer.from_pretrained("HelpingAI/HAI3-RAW")
chat = [
{"role": "system", "content": "You are HelpingAI, an emotional AI. Always answer my questions in the HelpingAI style."},
{"role": "user", "content": "Introduce yourself."}
]
inputs = tokenizer.apply_chat_template(
chat,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
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))
đ Documentation
Model Details
Property |
Details |
Developed by |
HelpingAI |
Model Type |
Decoder-only large language model |
Language |
English |
License |
HelpingAI License |
Training Data
HelpingAI3 was trained on a diverse dataset:
- Emotional Dialogues: 15 million rows for conversational intelligence.
- Therapeutic Exchanges: 3 million rows for emotional support.
- Cultural Conversations: 250,000 rows for global awareness.
- Crisis Response Scenarios: 1 million rows for emergency handling.
Training Procedure
- Base Model: Started from HelpingAI2.5.
- Emotional Intelligence Training: Used Reinforcement Learning for Emotion Understanding (RLEU) and context-aware conversational fine-tuning.
- Optimization: Employed mixed-precision training and advanced token efficiency techniques.
Intended Use
HelpingAI3 is designed for:
- AI Companionship & Emotional Support: Provides empathetic interactions.
- Therapy & Wellbeing Guidance: Assists in mental health support.
- Personalized Learning: Tailors educational content to individual needs.
- Professional AI Assistance: Boosts productivity in professional settings.
Limitations
- Biases: The model may reflect biases in the training data.
- Understanding Complex Emotions: May face challenges in interpreting nuanced human emotions.
- Not a Substitute for Professional Help: For serious emotional or psychological issues, consult a professional.
đ§ Technical Details
The model has the following technical aspects:
- Base Model: Initiated from HelpingAI2.5.
- Emotional Intelligence Training: Employed Reinforcement Learning for Emotion Understanding (RLEU) and context-aware conversational fine-tuning.
- Optimization: Utilized mixed-precision training and advanced token efficiency techniques.
đ License
This model is released under the HelpingAI License.