Enterprise Plans
AI Application Development
AI Translator API
AI SEO/GEO Service
Web Scraping Service
Log in
const { OpenAI } = require('openai');
const api = new OpenAI({
baseURL: 'https://api.ai.cc/v1',
apiKey: '',
});
const main = async () => {
const result = await api.chat.completions.create({
model: 'databricks/dolly-v2-3b',
messages: [
{
role: 'system',
content: 'You are an AI assistant who knows everything.',
},
{
role: 'user',
content: 'Tell me, why is the sky blue?'
}
],
});
const message = result.choices[0].message.content;
console.log(`Assistant: ${message}`);
};
main();
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.ai.cc/v1",
api_key="",
)
response = client.chat.completions.create(
model="databricks/dolly-v2-3b",
messages=[
{
"role": "system",
"content": "You are an AI assistant who knows everything.",
},
{
"role": "user",
"content": "Tell me, why is the sky blue?"
},
],
)
message = response.choices[0].message.content
print(f"Assistant: {message}")
-
AI Playground
Test all API models in the sandbox environment before you integrate.
We provide more than 300 models to integrate into your app.
Product Detail
Dolly v2 (3B): A Powerful Instruction-Following LLM by Databricks
Dolly v2 (3B), an advanced instruction-following large language model, was developed by Databricks, Inc. and officially released on April 12, 2023. Designated as Dolly-v2-3b, this model is meticulously engineered to comprehend and execute diverse language tasks based on user instructions. Its foundation lies in the Pythia-2.8b model, significantly enhanced through fine-tuning on approximately 15,000 high-quality instruction/response pairs, ensuring highly relevant and accurate output generation.
✨Key Capabilities of Dolly v2 (3B)
- Precision Fine-tuning: Optimized with ~15k instruction/response pairs for superior instruction adherence and quality.
- Versatile Task Performance: Highly capable in diverse tasks such as brainstorming, text classification, closed-book and open-book question answering, content generation, information extraction, and summarization.
- Commercial-Ready License: Available under a permissive CC-BY-SA license, making it suitable for commercial applications.
- Scalable Options: Larger variants, dolly-v2-7b and dolly-v2-12b, are also available for more extensive requirements.
🎯Intended Applications and Language Support
Dolly v2 (3B) is specifically tailored for a broad spectrum of Natural Language Processing (NLP) tasks. These include creative brainstorming, categorization, answering questions (both with provided context and general knowledge), generating new text, extracting specific information, and condensing lengthy content. It serves as an excellent choice for applications that demand reliable and consistent instruction following, though users should note it is not positioned as a state-of-the-art model.
Language Focus: The model's primary language of support is English. While processing other languages might be possible, accuracy could be diminished due to the predominant English content in its training dataset.
🧠Technical Specifications
Architecture
Dolly v2 (3B) is underpinned by the robust Pythia-2.8b model, leveraging a powerful Transformer-based architecture renowned for its efficacy in understanding and generating complex language patterns.
Training Data: Dolly v2 (3B) Official Blog Post
The model's strong instruction-following capabilities are refined through training on the databricks-dolly-15k dataset. This unique dataset, comprising approximately 15,000 instruction/response pairs, was meticulously generated by Databricks employees. It encompasses a diverse array of domains such as brainstorming, classification, question answering, and summarization, aligning with principles discussed in the InstructGPT paper.
- Source: Public internet, including Wikipedia.
- Size: Approximately 15,000 instruction/response pairs.
- Knowledge Cutoff: The model's knowledge base is current up to April 2023.
- Diversity & Bias: The dataset may exhibit biases and specific interests inherent to Databricks employees and, being partly derived from public internet data, can carry existing societal biases.
📊Performance Insights
Dolly v2 (3B) demonstrates robust performance in instruction-following tasks, notably surpassing its foundational model, Pythia-2.8b. It offers competitive results when benchmarked against other models within its parameter range. However, it is not engineered to compete with much larger, state-of-the-art models like GPT-4 or LLaMA-3.
- Accuracy: While generally effective, the model may encounter challenges with highly complex syntactic prompts, intricate programming problems, precise mathematical operations, nuanced factual accuracy, and accurate handling of dates and times.
- Speed: Optimized for efficient inference on GPUs, its actual operational speed can vary significantly based on specific hardware configurations.
- Robustness: Capable of managing a wide array of instructions, though it might occasionally produce errors in exceptionally complex or ambiguous scenarios.
🛠️Usage Guidelines
Dolly v2 (3B) is engineered for seamless integration into various applications through conventional API calls or SDKs. Developers can leverage its instruction-following capabilities for a multitude of NLP tasks. For comprehensive code samples and detailed integration instructions, it is recommended to consult the official Databricks documentation or relevant open-source repositories.
⚠️Ethical Considerations
Databricks is deeply committed to advancing AI technologies that are helpful, honest, and harmless. Nevertheless, like all artificial intelligence models, Dolly v2 (3B) possesses inherent limitations. It may generate outputs that reflect biases present within its training data, potentially leading to content that is biased or inadvertently harmful. Users are strongly advised to implement robust moderation systems and ethical safeguards when deploying and utilizing the model in real-world applications.
⚖️Licensing Information
Dolly v2 (3B) is distributed under the CC-BY-SA license (Creative Commons Attribution-ShareAlike). This is a permissive open-source license that expressly permits commercial use, thereby facilitating wide adoption and seamless integration into various commercial products and services.
❓Frequently Asked Questions (FAQ)
Q1: What is Dolly v2 (3B)?
A1: Dolly v2 (3B) is a large, instruction-following language model developed by Databricks, engineered to perform various NLP tasks based on user-provided instructions.
Q2: Can Dolly v2 (3B) be used for commercial projects?
A2: Absolutely. Dolly v2 (3B) is released under the CC-BY-SA license, which explicitly permits its use in commercial applications and products.
Q3: What are the primary strengths of Dolly v2 (3B)?
A3: Its main strengths include robust instruction-following capabilities, versatility across a wide range of NLP tasks (like summarization and QA), and its open-source nature, allowing for broad adoption.
Q4: What are the known limitations of Dolly v2 (3B)?
A4: It may face difficulties with highly complex syntax, programming problems, intricate mathematical operations, precise factual recall, and can inherit biases from its training data. It is not considered a state-of-the-art model compared to larger alternatives.
Q5: What languages does Dolly v2 (3B) support?
A5: The model primarily supports English. While it might process other languages, the accuracy and fluency could be diminished due to the English-centric nature of its training data.
Learn how you can transformyour company with AICC APIs
