Inserloft/ ParrotCode

4 Downloads Updated yesterday

A lightweight coding assistant by Inserloft Research, designed for fast Python code completion and local development.

ollama run Inserloft/ParrotCode

curl http://localhost:11434/api/chat \
  -d '{
    "model": "Inserloft/ParrotCode",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='Inserloft/ParrotCode',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'Inserloft/ParrotCode',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Models

Name

1 model

Size / Usage

Context

Input

ParrotCode:latest

517MB · 2K context window · Text · yesterday

ParrotCode:latest

517MB

2K

Text

Readme

Parrot Coder 1 🦜

Parrot Coder is a highly efficient, lightweight (~200M parameters) language model optimized for rapid code autocompletion and generation, running directly on your local machine via Ollama.

Developed by Inserloft Research.

🚀 Quick Start

Download the model:

ollama push Inserloft/ParrotCode

Run the model locally:

ollama run Inserloft/ParrotCode

💻 API Usage

You can seamlessly integrate Parrot Coder into your local IDEs or applications using the Ollama REST API:

curl http://localhost:11434/api/generate -d '{
  "model": "Inserloft/ParrotCode",
  "prompt": "def calculate_fibonacci(n):",
  "stream": false
}'

🧠 Model Details

Architecture: GPT-2 based with LoRA adapters (Merged)
Precision: FP16 (GGUF)
Primary Language: Python
License: Apache 2.0

📌 Use Cases

Code Autocompletion: Easily integrates with extensions like Continue.dev or Twinny in VSCode/JetBrains.
Offline Coding Assistant: Fast, privacy-focused code generation without requiring internet access.

For the full .safetensors weights or PyTorch inference, visit our Hugging Face repository.

![Parrot Coder.png](/assets/Inserloft/ParrotCode/6d3fc81a-edc0-42a1-a37a-83825fd51294)
# Parrot Coder 1 🦜

**Parrot Coder** is a highly efficient, lightweight (~200M parameters) language model optimized for rapid code autocompletion and generation, running directly on your local machine via Ollama.

Developed by [**Inserloft Research**](https://research.Inserloft.com).

## 🚀 Quick Start
Download the model:
```bash
ollama push Inserloft/ParrotCode
```

Run the model locally:
```bash
ollama run Inserloft/ParrotCode
```

## 💻 API Usage

You can seamlessly integrate Parrot Coder into your local IDEs or applications using the Ollama REST API:

```bash
curl http://localhost:11434/api/generate -d '{
  "model": "Inserloft/ParrotCode",
  "prompt": "def calculate_fibonacci(n):",
  "stream": false
}'
```

## 🧠 Model Details
- **Architecture:** GPT-2 based with LoRA adapters (Merged)
- **Precision:** FP16 (GGUF)
- **Primary Language:** Python
- **License:** Apache 2.0

## 📌 Use Cases
- **Code Autocompletion:** Easily integrates with extensions like Continue.dev or Twinny in VSCode/JetBrains.
- **Offline Coding Assistant:** Fast, privacy-focused code generation without requiring internet access.

For the full `.safetensors` weights or PyTorch inference, visit our [Hugging Face repository](https://huggingface.co/InserloftResearch/Parrot-Coder).

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)