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-v2-7b',
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-v2-7b",
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 (7B): An Open-Source Instruction-Following LLM
Dolly v2 (7B) is an innovative instruction-following large language model developed by Databricks. Released on April 12, 2023, this model is built upon the Pythia-6.9b architecture and has been fine-tuned on a substantial dataset of approximately 15,000 instruction/response pairs. Notably, Dolly v2 (7B) is licensed for commercial use, making it a powerful and accessible tool for a wide range of natural language processing tasks.
💡 Key Information
- Model Name: Dolly v2 (7B)
- Developer/Creator: Databricks
- Release Date: April 12, 2023
- Version: 2.0
- Model Type: Instruction-following Large Language Model
✅ Core Capabilities & Features
This model stands out for several reasons:
- • Instruction-following: Excels at understanding and executing explicit instructions.
- • Open-source: Promotes transparency and community-driven development.
- • Commercially licensed: Permitted for use in commercial applications under the Apache 2.0 license.
- • High-quality training: Fine-tuned on a curated dataset for superior instruction adherence.
- • Compact size: With approximately 6.9 billion parameters, it offers a good balance of performance and resource efficiency.
🛠️ Intended Applications
Dolly v2 (7B) is versatile and can be applied to a variety of natural language processing tasks, including:
- ▪️ Brainstorming and idea generation
- ▪️ Text classification
- ▪️ Answering both closed and open-ended questions
- ▪️ Creative and factual text generation
- ▪️ Information extraction from documents
- ▪️ Summarization of lengthy texts
Language Support: The model primarily supports English language tasks.
⚙️ Technical Specifications
Architecture
Dolly v2 (7B) is built upon the robust Pythia-6.9b architecture, a well-regarded transformer-based model known for its scalability and performance.
Training Data
Data Source & Size: The model underwent fine-tuning using the proprietary databricks-dolly-15k dataset. This high-quality dataset comprises approximately 15,000 instruction/response pairs, meticulously generated by Databricks employees.
Knowledge Cutoff: The foundational knowledge of Dolly v2 (7B) is derived from the Pythia-6.9b pre-training phase, enhanced significantly by the instruction-following capabilities developed during the fine-tuning process.
Performance Metrics
While Dolly v2 (7B) is not designed to achieve state-of-the-art benchmark scores against larger, more complex models, it consistently demonstrates surprisingly high-quality instruction-following behavior. Below are some benchmark results:
| Benchmark | Score |
|---|---|
| ARC (25-shot) | 0.392 |
| HellaSwag (10-shot) | 0.633838 |
| MMLU (5-shot) | 0.406997 |
| TruthfulQA (0-shot) | 0.444444 |
Comparison to Other Models
While Dolly v2 (7B) may not outperform significantly larger models like GPT-3 (175B parameters) in raw benchmark scores, it offers a compelling balance between strong performance and manageable resource requirements, making it highly practical for many applications.
🚀 Usage Guidelines
Code Sample
<snippet data-name="open-ai.chat-completion" data-model="databricks/dolly-v2-7b"></snippet>⚖️ Ethical Considerations
The development of Dolly v2 (7B) incorporated strong ethical considerations. The training data utilized does not contain obscenity, protected intellectual property, or personally identifying information pertaining to non-public figures. However, it's important to acknowledge that the model may reflect biases present in the data generated by Databricks employees. Users should exercise caution and awareness regarding potential biases in model outputs.
📄 Licensing Information
Dolly v2 (7B) is released under the Apache 2.0 license. This permissive license grants users the flexibility for both research and commercial applications, fostering broad adoption and innovation. For more details on Dolly v2's origins, you can refer to the Databricks Dolly v2 Announcement.
⚠️ Limitations of Dolly v2 (7B)
While powerful, Dolly v2 (7B) has certain limitations to be aware of:
- • Not bleeding-edge: It is not designed to compete with the performance of more modern or significantly larger model architectures.
- • Complex prompts: May struggle with prompts that are syntactically highly complex.
- • Specialized tasks: Has limited capabilities in programming problems and advanced mathematical operations.
- • Factual accuracy: May produce factual errors or exhibit hallucination, requiring human oversight.
- • Data biases: Reflects potential biases inherent in the training data generated by Databricks employees.
❓ Frequently Asked Questions (FAQ)
Q1: What is Dolly v2 (7B)?
A1: Dolly v2 (7B) is an instruction-following large language model developed by Databricks, based on the Pythia-6.9b model and fine-tuned on 15,000 instruction/response pairs.
Q2: Can Dolly v2 (7B) be used for commercial projects?
A2: Yes, Dolly v2 (7B) is released under the Apache 2.0 license, which permits both research and commercial use.
Q3: What are the primary applications of Dolly v2 (7B)?
A3: It is designed for various NLP tasks including brainstorming, classification, question answering (closed and open), text generation, information extraction, and summarization.
Q4: How does Dolly v2 (7B) compare to larger models like GPT-3?
A4: While it may not outperform models like GPT-3 in raw benchmarks, Dolly v2 (7B) offers a strong balance of performance and significantly lower resource requirements, making it a practical choice for many applications.
Q5: Are there any ethical concerns with Dolly v2 (7B)?
A5: The model was developed with ethical considerations, excluding obscenity, IP, and non-public PII from its training data. However, it may still reflect biases from the Databricks employees who generated the fine-tuning data.
Learn how you can transformyour company with AICC APIs
